WordPress.com
School Autonomy in Denver: The Impact of Innovation SchoolsPhilip GigliottiRockefeller College of Public Affairs and PolicyUniversity at Albany, SUNYAbstract:Improving performance in struggling urban schools is one of the most persistent challenges in education. In Denver, Colorado, a program called Innovation Schools provided urban public schools with autonomy to implement comprehensive managerial and educational reform plans by waiving district policies. I evaluate the Innovation Schools reform using a difference-in-differences design and find that the program increased end-of-year standardized test scores by 0.1 to 0.3 standard deviations in Innovation Schools. However, exploration of how these effects developed over time suggests that the early impacts of the program faded out following year two of implementation. These findings suggest the program was able to rapidly turn around low-performing schools, but these schools struggled to sustain results. This suggests school turnaround may be possible, but schools need to sustain efforts over time to preserve results. JEL Codes: I22, C23.Acknowledgements: I would like to thank Erika Martin, Alan Wagner, Beth Schueler, Michael Kang, and Andre Kiesel for their comments on this draft.IntroductionLow performance in urban public schools is a persistent educational problem, driven by persistent achievement gaps between wealthy and poor students. Recent studies investigate whether it is possible to “turn around” these schools, by implementing comprehensive managerial interventions to rapidly improve performance PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IZWlzc2VsPC9BdXRob3I+PFllYXI+MjAxODwvWWVhcj48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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IZWlzc2VsPC9BdXRob3I+PFllYXI+MjAxODwvWWVhcj48
UmVjTnVtPjE1MzwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oSGVpc3NlbCBhbmQgTGFkZCAyMDE4LCBT
Y2h1ZWxlciwgR29vZG1hbiwgYW5kIERlbWluZyAyMDE3LCBaaW1tZXIsIEhlbnJ5LCBhbmQgS2hv
IDIwMTcsIFN0cnVuayBldCBhbC4gMjAxNik8L0Rpc3BsYXlUZXh0PjxyZWNvcmQ+PHJlYy1udW1i
ZXI+MTUzPC9yZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iZWQ1
YXZyeHBsc3R0ZmllcnhmMXB4cDB2Mjk5cDl2dHZkZWV0IiB0aW1lc3RhbXA9IjE1NTUwNzg4MDMi
PjE1Mzwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUi
PjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+SGVpc3NlbCwgSmVu
bmlmZXIgQS48L2F1dGhvcj48YXV0aG9yPkxhZGQsIEhlbGVuIEYuPC9hdXRob3I+PC9hdXRob3Jz
PjwvY29udHJpYnV0b3JzPjx0aXRsZXM+PHRpdGxlPlNjaG9vbCBUdXJuYXJvdW5kIGluIE5vcnRo
IENhcm9saW5hOiBBIFJlZ3Jlc3Npb24gRGlzY29udGludWl0eSBBbmFseXNpczwvdGl0bGU+PHNl
Y29uZGFyeS10aXRsZT5FY29ub21pY3Mgb2YgRWR1Y2F0aW9uIFJldmlldzwvc2Vjb25kYXJ5LXRp
dGxlPjwvdGl0bGVzPjxwZXJpb2RpY2FsPjxmdWxsLXRpdGxlPkVjb25vbWljcyBvZiBFZHVjYXRp
b24gUmV2aWV3PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48dm9sdW1lPjYyPC92b2x1bWU+PG51
bWJlcj4zMDItMzIwPC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTg8L3llYXI+PC9kYXRlcz48dXJs
cz48L3VybHM+PC9yZWNvcmQ+PC9DaXRlPjxDaXRlPjxBdXRob3I+U2NodWVsZXI8L0F1dGhvcj48
WWVhcj4yMDE3PC9ZZWFyPjxSZWNOdW0+MTUyPC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj4x
NTI8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4
cGxzdHRmaWVyeGYxcHhwMHYyOTlwOXZ0dmRlZXQiIHRpbWVzdGFtcD0iMTU1NTA3ODU0MSI+MTUy
PC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8
L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5TY2h1ZWxlciwgQmV0aCBF
LjwvYXV0aG9yPjxhdXRob3I+R29vZG1hbiwgSm9zaHVhIFMuPC9hdXRob3I+PGF1dGhvcj5EZW1p
bmcsIERhdmlkIEouPC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjx0aXRsZXM+PHRp
dGxlPkNhbiBTdGF0ZXMgVGFrZSBPdmVyIGFuZCBUdXJuIEFyb3VuZCBTY2hvb2wgRGlzdHJpY3Rz
PyBFdmlkZW5jZSBmcm9tIExhd3JlbmNlLCBNYXNzYWNodXNldHRzPC90aXRsZT48c2Vjb25kYXJ5
LXRpdGxlPkVkdWNhdGlvbmFsIEV2YWx1YXRpb24gYW5kIFBvbGljeSBBbmFseXNpczwvc2Vjb25k
YXJ5LXRpdGxlPjwvdGl0bGVzPjxwZXJpb2RpY2FsPjxmdWxsLXRpdGxlPkVkdWNhdGlvbmFsIEV2
YWx1YXRpb24gYW5kIFBvbGljeSBBbmFseXNpczwvZnVsbC10aXRsZT48L3BlcmlvZGljYWw+PHBh
Z2VzPjMxMS0zMzI8L3BhZ2VzPjx2b2x1bWU+Mzk8L3ZvbHVtZT48bnVtYmVyPjI8L251bWJlcj48
ZGF0ZXM+PHllYXI+MjAxNzwveWVhcj48L2RhdGVzPjx1cmxzPjwvdXJscz48L3JlY29yZD48L0Np
dGU+PENpdGU+PEF1dGhvcj5aaW1tZXI8L0F1dGhvcj48WWVhcj4yMDE3PC9ZZWFyPjxSZWNOdW0+
MTU4PC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj4xNTg8L3JlYy1udW1iZXI+PGZvcmVpZ24t
a2V5cz48a2V5IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYyOTlwOXZ0
dmRlZXQiIHRpbWVzdGFtcD0iMTU1NjExNDE1NyI+MTU4PC9rZXk+PC9mb3JlaWduLWtleXM+PHJl
Zi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+
PGF1dGhvcnM+PGF1dGhvcj5aaW1tZXIsIFJvbjwvYXV0aG9yPjxhdXRob3I+SGVucnksIEdhcnkg
VC48L2F1dGhvcj48YXV0aG9yPktobywgQWRhbTwvYXV0aG9yPjwvYXV0aG9ycz48L2NvbnRyaWJ1
dG9ycz48dGl0bGVzPjx0aXRsZT5UaGUgRWZmZWN0cyBvZiBTY2hvb2wgVHVybmFyb3VuZCBpbiBU
ZW5uZXNzZWUmYXBvcztzIEFjaGlldmVtZW50IFNjaG9vbCBEaXN0cmljdCBhbmQgSW5ub3ZhdGlv
biBab25lczwvdGl0bGU+PHNlY29uZGFyeS10aXRsZT5FZHVjYXRpb25hbCBFdmFsdWF0aW9uIGFu
ZCBQb2xpY3kgQW5hbHlzaXM8L3NlY29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48
ZnVsbC10aXRsZT5FZHVjYXRpb25hbCBFdmFsdWF0aW9uIGFuZCBQb2xpY3kgQW5hbHlzaXM8L2Z1
bGwtdGl0bGU+PC9wZXJpb2RpY2FsPjxwYWdlcz42NzAtNjk2PC9wYWdlcz48dm9sdW1lPjM5PC92
b2x1bWU+PG51bWJlcj40PC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTc8L3llYXI+PC9kYXRlcz48
dXJscz48L3VybHM+PC9yZWNvcmQ+PC9DaXRlPjxDaXRlPjxBdXRob3I+U3RydW5rPC9BdXRob3I+
PFllYXI+MjAxNjwvWWVhcj48UmVjTnVtPjE1MTwvUmVjTnVtPjxyZWNvcmQ+PHJlYy1udW1iZXI+
MTUxPC9yZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iZWQ1YXZy
eHBsc3R0ZmllcnhmMXB4cDB2Mjk5cDl2dHZkZWV0IiB0aW1lc3RhbXA9IjE1NTUwNzgyMTEiPjE1
MTwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3
PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+U3RydW5rLCBLYXRoZXJp
bmUgTy48L2F1dGhvcj48YXV0aG9yPk1hcnNoLCBKdWxpZSBBLjwvYXV0aG9yPjxhdXRob3I+SGFz
aGltLCBBeWVzaGEgSy48L2F1dGhvcj48YXV0aG9yPkJ1c2gtTWVjZW5hcywgU3VzYW48L2F1dGhv
cj48YXV0aG9yPldlaW5zdGVpbiwgVHJhY2V5PC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0
b3JzPjx0aXRsZXM+PHRpdGxlPlRoZSBJbXBhY3Qgb2YgVHVybmFyb3VuZCBSZWZvcm0gb24gU3R1
ZGVudCBPdXRjb21lczogRXZpZGVuY2UgYW5kIEluc2lnaHRzIGZyb20gdGhlIExvcyBBbmdlbGVz
IFVuaWZpZWQgU2Nob29sIERpc3RyaWN0PC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkVkdWNhdGlv
biBGaW5hbmNlIGFuZCBQb2xpY3k8L3NlY29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNh
bD48ZnVsbC10aXRsZT5FZHVjYXRpb24gRmluYW5jZSBhbmQgUG9saWN5PC9mdWxsLXRpdGxlPjwv
cGVyaW9kaWNhbD48cGFnZXM+MjUxLTI4MjwvcGFnZXM+PHZvbHVtZT4xMTwvdm9sdW1lPjxudW1i
ZXI+MzwvbnVtYmVyPjxkYXRlcz48eWVhcj4yMDE2PC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91cmxz
PjwvcmVjb3JkPjwvQ2l0ZT48L0VuZE5vdGU+AG==
ADDIN EN.CITE.DATA (Heissel and Ladd 2018, Schueler, Goodman, and Deming 2017, Zimmer, Henry, and Kho 2017, Strunk et al. 2016). This study evaluates a turnaround program in Denver Public Schools (DPS) which provided schools with managerial autonomy to lead innovative reforms. The aim of this study is to shed light on the impact of managerial autonomy on public school performance and the possibility of leading successful turnaround reforms. The possibility of turnarounds to improve performance in struggling urban schools would represent progress after decades of unsuccessful efforts.Many interventions have attempted to improve the performance of struggling urban schools with few encouraging results. States have increased educational aid to high-need schools by thousands in per-pupil dollars with little resulting improvement in national performance trends ADDIN EN.CITE <EndNote><Cite><Author>Hanushek</Author><Year>1997</Year><RecNum>26</RecNum><DisplayText>(Hanushek 1997, Yinger 2004)</DisplayText><record><rec-number>26</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534532974">26</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Hanushek, Eric</author></authors></contributors><titles><title>Assessing the Effects of School Resources on Student Performance: An Update</title><secondary-title>Educational Evaluation and Policy Analysis</secondary-title></titles><periodical><full-title>Educational Evaluation and Policy Analysis</full-title></periodical><pages>141-164</pages><volume>19</volume><number>2</number><dates><year>1997</year></dates><urls></urls></record></Cite><Cite><Author>Yinger</Author><Year>2004</Year><RecNum>16</RecNum><record><rec-number>16</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534444662">16</key></foreign-keys><ref-type name="Book">6</ref-type><contributors><authors><author>Yinger, John</author></authors></contributors><titles><title>Helping Children Left Behind: State Aid and the Pursuit of Educational Equity</title></titles><dates><year>2004</year></dates><pub-location>Cambridge, MA</pub-location><publisher>MIT University Press</publisher><urls></urls></record></Cite></EndNote>(Hanushek 1997, Yinger 2004). While ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>LaFortune</Author><Year>2018</Year><RecNum>3</RecNum><DisplayText>LaFortune, Rothstein, and Schanzenbach (2018)</DisplayText><record><rec-number>3</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534182127">3</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>LaFortune, Julien</author><author>Rothstein, Jesse</author><author>Schanzenbach, Diane Whitmore</author></authors></contributors><titles><title>School Finance Reform and the Distribution of Student Achievement</title><secondary-title>American Economic Journal: Applied Economics</secondary-title></titles><periodical><full-title>American Economic Journal: Applied Economics</full-title></periodical><pages>1-26</pages><volume>10</volume><number>2</number><dates><year>2018</year></dates><urls></urls></record></Cite></EndNote>LaFortune, Rothstein, and Schanzenbach (2018) now demonstrate that these investments produced achievement gains, their magnitude does not suggest that finance reforms could close achievement gaps without unprecedented investments. Other reforms focus on holding schools accountable for their performance on standardized examinations. While these accountability reforms have been shown to increase performance in certain contexts ADDIN EN.CITE <EndNote><Cite><Author>Chiang</Author><Year>2009</Year><RecNum>18</RecNum><DisplayText>(Chiang 2009, Rockoff and Turner 2010)</DisplayText><record><rec-number>18</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534445318">18</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Chiang, Hanley</author></authors></contributors><titles><title>How Accountability Pressure on Failing Schools Affects Student Achievement</title><secondary-title>Journal of Public Economics</secondary-title></titles><periodical><full-title>Journal of Public Economics</full-title></periodical><pages>1045-1057</pages><volume>93</volume><number>9-10</number><dates><year>2009</year></dates><urls></urls></record></Cite><Cite><Author>Rockoff</Author><Year>2010</Year><RecNum>19</RecNum><record><rec-number>19</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534445485">19</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Rockoff, Jonah</author><author>Turner, Lesley J.</author></authors></contributors><titles><title>Short-Run Impacts of Accountability on School Quality</title><secondary-title>American Economic Journal: Economics Policy</secondary-title></titles><periodical><full-title>American Economic Journal: Economics Policy</full-title></periodical><pages>119-47</pages><volume>2</volume><number>4</number><dates><year>2010</year></dates><urls></urls></record></Cite></EndNote>(Chiang 2009, Rockoff and Turner 2010), evaluations of two national accountability efforts exhibited failure to close achievement gaps or ambiguous impact on student achievement. ADDIN EN.CITE <EndNote><Cite><Author>Dee</Author><Year>2011</Year><RecNum>4</RecNum><DisplayText>(Dee and Jacob 2011, Hanushek and Raymond 2005)</DisplayText><record><rec-number>4</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534182515">4</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Dee, Thomas S.</author><author>Jacob, Brian A.</author></authors></contributors><titles><title>The Impact of No Child Left Behind on Student Achievement</title><secondary-title>Journal of Policy Analysis and Management</secondary-title></titles><periodical><full-title>Journal of Policy Analysis and Management</full-title></periodical><pages>418-446</pages><volume>30</volume><number>3</number><dates><year>2011</year></dates><urls></urls></record></Cite><Cite><Author>Hanushek</Author><Year>2005</Year><RecNum>17</RecNum><record><rec-number>17</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534445203">17</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Hanushek, Eric</author><author>Raymond, Margaret E.</author></authors></contributors><titles><title>Does School Accountability Lead to Improved Student Performance</title><secondary-title>Journal of Policy Analysis and Management</secondary-title></titles><periodical><full-title>Journal of Policy Analysis and Management</full-title></periodical><pages>297-327</pages><volume>24</volume><number>2</number><dates><year>2005</year></dates><urls></urls></record></Cite></EndNote>(Dee and Jacob 2011, Hanushek and Raymond 2005). Furthermore, even after achieving universal implementation of accountability systems in American public schools following the 2001 No Child Left Behind Act (NCLB), achievement gaps still persist.Some reformers advocate for charter schools, public schools which make decisions independently of school districts, allowing them to implement programs which are responsive to students’ needs. While they are subject to criticism, and often demonstrate negative or null impact on student achievement, PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CaWZ1bGNvPC9BdXRob3I+PFllYXI+MjAwNjwvWWVhcj48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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5CaWZ1bGNvPC9BdXRob3I+PFllYXI+MjAwNjwvWWVhcj48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ADDIN EN.CITE.DATA (Bifulco and Ladd 2006, Imberman 2011, Booker, Gilpatric, and Gronberg 2007) many studies now demonstrate that charter schools can improve student outcomes in a variety of contexts PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5Eb2JiaWU8L0F1dGhvcj48WWVhcj4yMDExPC9ZZWFyPjxS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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5Eb2JiaWU8L0F1dGhvcj48WWVhcj4yMDExPC9ZZWFyPjxS
ZWNOdW0+NjwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oRG9iYmllIGFuZCBGcnllciAyMDExLCBBbmdy
aXN0IGV0IGFsLiAyMDE2LCBCb29rZXIgZXQgYWwuIDIwMTEpPC9EaXNwbGF5VGV4dD48cmVjb3Jk
PjxyZWMtbnVtYmVyPjY8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRi
LWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYyOTlwOXZ0dmRlZXQiIHRpbWVzdGFtcD0iMTUz
NDE4MzE4NSI+Njwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFy
dGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+RG9iYmll
LCBXaWxsPC9hdXRob3I+PGF1dGhvcj5GcnllciwgIFJvbGFuZCBHLjwvYXV0aG9yPjwvYXV0aG9y
cz48L2NvbnRyaWJ1dG9ycz48dGl0bGVzPjx0aXRsZT5BcmUgSGlnaCBRdWFsaXR5IFNjaG9vbHMg
RW5vdWdoIHRvIEluY3JlYXNlIEFjaGlldmVtZW50IEFtb25nIHRoZSBQb29yPyBFdmlkZW5jZSBm
cm9tIHRoZSBIYXJsZW0gQ2hpbGRyZW4mYXBvcztzIFpvbmU8L3RpdGxlPjxzZWNvbmRhcnktdGl0
bGU+QW1lcmljYW4gRWNvbm9taWMgSm91cm5hbDogQXBwbGllZCBFY29ub21pY3M8L3NlY29uZGFy
eS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5BbWVyaWNhbiBFY29ub21p
YyBKb3VybmFsOiBBcHBsaWVkIEVjb25vbWljczwvZnVsbC10aXRsZT48L3BlcmlvZGljYWw+PHBh
Z2VzPjE1OC0xODc8L3BhZ2VzPjx2b2x1bWU+Mzwvdm9sdW1lPjxudW1iZXI+MzwvbnVtYmVyPjxk
YXRlcz48eWVhcj4yMDExPC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91cmxzPjwvcmVjb3JkPjwvQ2l0
ZT48Q2l0ZT48QXV0aG9yPkFuZ3Jpc3Q8L0F1dGhvcj48WWVhcj4yMDE2PC9ZZWFyPjxSZWNOdW0+
MjM8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVyPjIzPC9yZWMtbnVtYmVyPjxmb3JlaWduLWtl
eXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iZWQ1YXZyeHBsc3R0ZmllcnhmMXB4cDB2Mjk5cDl2dHZk
ZWV0IiB0aW1lc3RhbXA9IjE1MzQ0NDY0MjMiPjIzPC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10
eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1
dGhvcnM+PGF1dGhvcj5BbmdyaXN0LCBKb3NodWEgRC48L2F1dGhvcj48YXV0aG9yPkNvaG9kZXMs
IFNhcmFoIFIuPC9hdXRob3I+PGF1dGhvcj5EeW5hcnNraSwgU3VzYW4gTS48L2F1dGhvcj48YXV0
aG9yPlBhdGhhaywgUGFyYWcgQS48L2F1dGhvcj48YXV0aG9yPldhbHRlcnMsIENocmlzdG9waGVy
IFIuPC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjx0aXRsZXM+PHRpdGxlPlN0YW5k
IGFuZCBEZWxpdmVyOiBFZmZlY3RzIG9mIEJvc3RvbiZhcG9zO3MgQ2hhcnRlciBTY2hvb2xzIG9u
IENvbGxlZ2UgUHJlcGFyYXRpb24sIEVudHJ5IGFuZCBDaG9pY2U8L3RpdGxlPjxzZWNvbmRhcnkt
dGl0bGU+Sm91cm5hbCBvZiBMYWJvciBFY29ub21pY3M8L3NlY29uZGFyeS10aXRsZT48L3RpdGxl
cz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5Kb3VybmFsIG9mIExhYm9yIEVjb25vbWljczwvZnVs
bC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjI3NS0zMTg8L3BhZ2VzPjx2b2x1bWU+MzQ8L3Zv
bHVtZT48bnVtYmVyPjI8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAxNjwveWVhcj48L2RhdGVzPjx1
cmxzPjwvdXJscz48L3JlY29yZD48L0NpdGU+PENpdGU+PEF1dGhvcj5Cb29rZXI8L0F1dGhvcj48
WWVhcj4yMDExPC9ZZWFyPjxSZWNOdW0+MjI8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVyPjIy
PC9yZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iZWQ1YXZyeHBs
c3R0ZmllcnhmMXB4cDB2Mjk5cDl2dHZkZWV0IiB0aW1lc3RhbXA9IjE1MzQ0NDYxNDciPjIyPC9r
ZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3Jl
Zi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5Cb29rZXIsIEtldmluPC9hdXRo
b3I+PGF1dGhvcj5TYXNzLCBUaW0gUi48L2F1dGhvcj48YXV0aG9yPkdpbGwsIEJyaWFuPC9hdXRo
b3I+PGF1dGhvcj5aaW1tZXIsIFJvbjwvYXV0aG9yPjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48
dGl0bGVzPjx0aXRsZT5UaGUgRWZmZWN0cyBvZiBDaGFydGVyIEhpZ2ggU2Nob29scyBvbiBFZHVj
YXRpb25hbCBBdHRhaW5tZW50PC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkpvdXJuYWwgb2YgTGFi
b3IgRWNvbm9taWNzPC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlvZGljYWw+PGZ1bGwt
dGl0bGU+Sm91cm5hbCBvZiBMYWJvciBFY29ub21pY3M8L2Z1bGwtdGl0bGU+PC9wZXJpb2RpY2Fs
PjxwYWdlcz4zNzctNDE1PC9wYWdlcz48dm9sdW1lPjI5PC92b2x1bWU+PG51bWJlcj4yPC9udW1i
ZXI+PGRhdGVzPjx5ZWFyPjIwMTE8L3llYXI+PC9kYXRlcz48dXJscz48L3VybHM+PC9yZWNvcmQ+
PC9DaXRlPjwvRW5kTm90ZT5=
ADDIN EN.CITE.DATA (Dobbie and Fryer 2011, Angrist et al. 2016, Booker et al. 2011).The potential for charter schools to improve student outcomes may work through two mechanisms. The first is school choice: by allowing students to enroll in good schools and leave bad schools, competition for students may drive schools to improve performance. The second is independence from centralized district management, which allows charters to be responsive to student needs, making them more effectively and innovatively managed ADDIN EN.CITE <EndNote><Cite><Author>Chubb</Author><Year>1990</Year><RecNum>27</RecNum><DisplayText>(Chubb and Moe 1990)</DisplayText><record><rec-number>27</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534533242">27</key></foreign-keys><ref-type name="Book">6</ref-type><contributors><authors><author>Chubb, John E.</author><author>Moe, Terry M.</author></authors></contributors><titles><title>Politics, Markets and America's Schools</title></titles><dates><year>1990</year></dates><pub-location>Washington, DC.</pub-location><publisher>Brookings Institution Press</publisher><urls></urls></record></Cite></EndNote>(Chubb and Moe 1990). Empirical evidence suggests that innovative management practices can sometimes explain the positive impacts of effective charter schools. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Dobbie</Author><Year>2011</Year><RecNum>6</RecNum><DisplayText>Dobbie and Fryer (2011)</DisplayText><record><rec-number>6</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534183185">6</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Dobbie, Will</author><author>Fryer, Roland G.</author></authors></contributors><titles><title>Are High Quality Schools Enough to Increase Achievement Among the Poor? Evidence from the Harlem Children's Zone</title><secondary-title>American Economic Journal: Applied Economics</secondary-title></titles><periodical><full-title>American Economic Journal: Applied Economics</full-title></periodical><pages>158-187</pages><volume>3</volume><number>3</number><dates><year>2011</year></dates><urls></urls></record></Cite></EndNote>Dobbie and Fryer (2011) have linked the positive effects of high-quality charter schools to a set of innovative managerial strategies. In a project called Apollo 20, involving 9 district high schools and 11 district elementary schools in Houston, Fryer (2014) demonstrated that these strategies could be adopted by traditional public schools and produce similar positive effects: students posted significant annual gains in math but little progress in reading. Without the buy-in of parents and students opting into a charter school, Apollo 20 necessarily exhibited limitations (Abrams, 2016). The project nevertheless suggests that providing traditional public schools with the autonomy to implement managerial reforms could play an important role in turnaround efforts and lead to performance improvementsIn 2009, DPS implemented a program called Innovation Schools that allowed traditional public schools to waive district policies in order to pursue innovative managerial reforms, such as comprehensive changes to curriculum, scheduling and personnel management. This reform contributed to district-wide efforts to improve performance in struggling schools through enhanced accountability and innovative school governance, including expanded school choice. Innovation schools were accountable to DPS for faithful implementation of comprehensive reform plans and improved performance according to a district-wide accountability system which gave each school an annual performance rating and set targets for growth. The conceptual model of change underlying the Innovation Schools program was that providing traditional public schools with the autonomy to pursue independent managerial reforms would lead to improved school performance in the context of a struggling urban school district. Evaluation of the Innovation Schools program provides multiple contributions to the educational literature. First, it contributes to the literature on school turnarounds. School turnarounds are interventions that attempt to rapidly improve the performance of struggling public schools using comprehensive managerial reforms. The Innovation Schools program can be considered a turnaround-style reform, as it leverages managerial reform to rapidly improve school performance in the context of a very low-performing district. Estimating the impact of this program, including how program impacts develop over time, can expand the turnaround literature and suggest new strategies for turnaround reforms.School turnarounds have been evaluated in a number of contexts with mixed results. The most notable implementation of school turnaround occurred under the Title 1 School Improvement Grants (SIG) program, which provided grants to persistently low-performing schools which were conditional on managerial reform, often including replacement of staff and leadership. SIGs produced positive academic impacts in Ohio ADDIN EN.CITE <EndNote><Cite><Author>Carlson</Author><Year>2018</Year><RecNum>148</RecNum><DisplayText>(Carlson and Lavertu 2018)</DisplayText><record><rec-number>148</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1554330740">148</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Carlson, Deven E.</author><author>Lavertu, Stephane</author></authors></contributors><titles><title>School Improvement Grants in Ohio: Effects on Student Achievement and School Administration</title><secondary-title>Educational Evaluation and Policy Analysis</secondary-title></titles><periodical><full-title>Educational Evaluation and Policy Analysis</full-title></periodical><pages>287-315</pages><volume>40</volume><number>3</number><dates><year>2018</year></dates><urls></urls></record></Cite></EndNote>(Carlson and Lavertu 2018) and California ADDIN EN.CITE <EndNote><Cite><Author>Dee</Author><Year>2012</Year><RecNum>34</RecNum><DisplayText>(Dee 2012)</DisplayText><record><rec-number>34</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1535037143">34</key></foreign-keys><ref-type name="Unpublished Work">34</ref-type><contributors><authors><author>Dee, Thomas S.</author></authors></contributors><titles><title>School Turnarounds: Evidence from the 2009 Stimulus</title><secondary-title>NBER Working Papers</secondary-title></titles><number>17990</number><dates><year>2012</year></dates><urls></urls></record></Cite></EndNote>(Dee 2012), but evaluation of a national sample of SIG schools showed null academic impacts and indications of failed implementation ADDIN EN.CITE <EndNote><Cite><Author>Dragoset</Author><Year>2017</Year><RecNum>35</RecNum><DisplayText>(Dragoset et al. 2017)</DisplayText><record><rec-number>35</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1535037306">35</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>Dragoset, Lisa</author><author>Thomas, Jaime</author><author>Herrmann, Mariesa</author><author>Deke, John</author><author>James-Burdumy, Susanne</author><author>Graczewski, Cheryl</author><author>Boyle, Andrea</author><author>Upton, Rachel</author><author>Tanenbaum, Courtney</author><author>Giffin, Jessica</author></authors><tertiary-authors><author>Mathematica Policy Research</author></tertiary-authors></contributors><titles><title>School Improvement Grants: Implementation and Effectiveness</title></titles><dates><year>2017</year></dates><publisher>U.S. Department of Education</publisher><urls><related-urls><url>;(Dragoset et al. 2017). Turnarounds have been evaluated in other contexts with some showing large impacts on academic performance ADDIN EN.CITE <EndNote><Cite><Author>Schueler</Author><Year>2017</Year><RecNum>152</RecNum><DisplayText>(Schueler, Goodman, and Deming 2017)</DisplayText><record><rec-number>152</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1555078541">152</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Schueler, Beth E.</author><author>Goodman, Joshua S.</author><author>Deming, David J.</author></authors></contributors><titles><title>Can States Take Over and Turn Around School Districts? Evidence from Lawrence, Massachusetts</title><secondary-title>Educational Evaluation and Policy Analysis</secondary-title></titles><periodical><full-title>Educational Evaluation and Policy Analysis</full-title></periodical><pages>311-332</pages><volume>39</volume><number>2</number><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>(Schueler, Goodman, and Deming 2017) but others showing mixed academic impacts and negative impacts on school climate PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5aaW1tZXI8L0F1dGhvcj48WWVhcj4yMDE3PC9ZZWFyPjxS
ZWNOdW0+MTU4PC9SZWNOdW0+PERpc3BsYXlUZXh0PihaaW1tZXIsIEhlbnJ5LCBhbmQgS2hvIDIw
MTcsIFN0cnVuayBldCBhbC4gMjAxNiwgSGVpc3NlbCBhbmQgTGFkZCAyMDE4KTwvRGlzcGxheVRl
eHQ+PHJlY29yZD48cmVjLW51bWJlcj4xNTg8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5
IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYyOTlwOXZ0dmRlZXQiIHRp
bWVzdGFtcD0iMTU1NjExNDE1NyI+MTU4PC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5h
bWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+
PGF1dGhvcj5aaW1tZXIsIFJvbjwvYXV0aG9yPjxhdXRob3I+SGVucnksIEdhcnkgVC48L2F1dGhv
cj48YXV0aG9yPktobywgQWRhbTwvYXV0aG9yPjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0
bGVzPjx0aXRsZT5UaGUgRWZmZWN0cyBvZiBTY2hvb2wgVHVybmFyb3VuZCBpbiBUZW5uZXNzZWUm
YXBvcztzIEFjaGlldmVtZW50IFNjaG9vbCBEaXN0cmljdCBhbmQgSW5ub3ZhdGlvbiBab25lczwv
dGl0bGU+PHNlY29uZGFyeS10aXRsZT5FZHVjYXRpb25hbCBFdmFsdWF0aW9uIGFuZCBQb2xpY3kg
QW5hbHlzaXM8L3NlY29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRs
ZT5FZHVjYXRpb25hbCBFdmFsdWF0aW9uIGFuZCBQb2xpY3kgQW5hbHlzaXM8L2Z1bGwtdGl0bGU+
PC9wZXJpb2RpY2FsPjxwYWdlcz42NzAtNjk2PC9wYWdlcz48dm9sdW1lPjM5PC92b2x1bWU+PG51
bWJlcj40PC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTc8L3llYXI+PC9kYXRlcz48dXJscz48L3Vy
bHM+PC9yZWNvcmQ+PC9DaXRlPjxDaXRlPjxBdXRob3I+U3RydW5rPC9BdXRob3I+PFllYXI+MjAx
NjwvWWVhcj48UmVjTnVtPjE1MTwvUmVjTnVtPjxyZWNvcmQ+PHJlYy1udW1iZXI+MTUxPC9yZWMt
bnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iZWQ1YXZyeHBsc3R0Zmll
cnhmMXB4cDB2Mjk5cDl2dHZkZWV0IiB0aW1lc3RhbXA9IjE1NTUwNzgyMTEiPjE1MTwva2V5Pjwv
Zm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlw
ZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+U3RydW5rLCBLYXRoZXJpbmUgTy48L2F1
dGhvcj48YXV0aG9yPk1hcnNoLCBKdWxpZSBBLjwvYXV0aG9yPjxhdXRob3I+SGFzaGltLCBBeWVz
aGEgSy48L2F1dGhvcj48YXV0aG9yPkJ1c2gtTWVjZW5hcywgU3VzYW48L2F1dGhvcj48YXV0aG9y
PldlaW5zdGVpbiwgVHJhY2V5PC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjx0aXRs
ZXM+PHRpdGxlPlRoZSBJbXBhY3Qgb2YgVHVybmFyb3VuZCBSZWZvcm0gb24gU3R1ZGVudCBPdXRj
b21lczogRXZpZGVuY2UgYW5kIEluc2lnaHRzIGZyb20gdGhlIExvcyBBbmdlbGVzIFVuaWZpZWQg
U2Nob29sIERpc3RyaWN0PC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkVkdWNhdGlvbiBGaW5hbmNl
IGFuZCBQb2xpY3k8L3NlY29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10
aXRsZT5FZHVjYXRpb24gRmluYW5jZSBhbmQgUG9saWN5PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNh
bD48cGFnZXM+MjUxLTI4MjwvcGFnZXM+PHZvbHVtZT4xMTwvdm9sdW1lPjxudW1iZXI+MzwvbnVt
YmVyPjxkYXRlcz48eWVhcj4yMDE2PC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91cmxzPjwvcmVjb3Jk
PjwvQ2l0ZT48Q2l0ZT48QXV0aG9yPkhlaXNzZWw8L0F1dGhvcj48WWVhcj4yMDE4PC9ZZWFyPjxS
ZWNOdW0+MTUzPC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj4xNTM8L3JlYy1udW1iZXI+PGZv
cmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYy
OTlwOXZ0dmRlZXQiIHRpbWVzdGFtcD0iMTU1NTA3ODgwMyI+MTUzPC9rZXk+PC9mb3JlaWduLWtl
eXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmli
dXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5IZWlzc2VsLCBKZW5uaWZlciBBLjwvYXV0aG9yPjxhdXRo
b3I+TGFkZCwgSGVsZW4gRi48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxl
cz48dGl0bGU+U2Nob29sIFR1cm5hcm91bmQgaW4gTm9ydGggQ2Fyb2xpbmE6IEEgUmVncmVzc2lv
biBEaXNjb250aW51aXR5IEFuYWx5c2lzPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkVjb25vbWlj
cyBvZiBFZHVjYXRpb24gUmV2aWV3PC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlvZGlj
YWw+PGZ1bGwtdGl0bGU+RWNvbm9taWNzIG9mIEVkdWNhdGlvbiBSZXZpZXc8L2Z1bGwtdGl0bGU+
PC9wZXJpb2RpY2FsPjx2b2x1bWU+NjI8L3ZvbHVtZT48bnVtYmVyPjMwMi0zMjA8L251bWJlcj48
ZGF0ZXM+PHllYXI+MjAxODwveWVhcj48L2RhdGVzPjx1cmxzPjwvdXJscz48L3JlY29yZD48L0Np
dGU+PC9FbmROb3RlPn==
ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5aaW1tZXI8L0F1dGhvcj48WWVhcj4yMDE3PC9ZZWFyPjxS
ZWNOdW0+MTU4PC9SZWNOdW0+PERpc3BsYXlUZXh0PihaaW1tZXIsIEhlbnJ5LCBhbmQgS2hvIDIw
MTcsIFN0cnVuayBldCBhbC4gMjAxNiwgSGVpc3NlbCBhbmQgTGFkZCAyMDE4KTwvRGlzcGxheVRl
eHQ+PHJlY29yZD48cmVjLW51bWJlcj4xNTg8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5
IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYyOTlwOXZ0dmRlZXQiIHRp
bWVzdGFtcD0iMTU1NjExNDE1NyI+MTU4PC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5h
bWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+
PGF1dGhvcj5aaW1tZXIsIFJvbjwvYXV0aG9yPjxhdXRob3I+SGVucnksIEdhcnkgVC48L2F1dGhv
cj48YXV0aG9yPktobywgQWRhbTwvYXV0aG9yPjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0
bGVzPjx0aXRsZT5UaGUgRWZmZWN0cyBvZiBTY2hvb2wgVHVybmFyb3VuZCBpbiBUZW5uZXNzZWUm
YXBvcztzIEFjaGlldmVtZW50IFNjaG9vbCBEaXN0cmljdCBhbmQgSW5ub3ZhdGlvbiBab25lczwv
dGl0bGU+PHNlY29uZGFyeS10aXRsZT5FZHVjYXRpb25hbCBFdmFsdWF0aW9uIGFuZCBQb2xpY3kg
QW5hbHlzaXM8L3NlY29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRs
ZT5FZHVjYXRpb25hbCBFdmFsdWF0aW9uIGFuZCBQb2xpY3kgQW5hbHlzaXM8L2Z1bGwtdGl0bGU+
PC9wZXJpb2RpY2FsPjxwYWdlcz42NzAtNjk2PC9wYWdlcz48dm9sdW1lPjM5PC92b2x1bWU+PG51
bWJlcj40PC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTc8L3llYXI+PC9kYXRlcz48dXJscz48L3Vy
bHM+PC9yZWNvcmQ+PC9DaXRlPjxDaXRlPjxBdXRob3I+U3RydW5rPC9BdXRob3I+PFllYXI+MjAx
NjwvWWVhcj48UmVjTnVtPjE1MTwvUmVjTnVtPjxyZWNvcmQ+PHJlYy1udW1iZXI+MTUxPC9yZWMt
bnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iZWQ1YXZyeHBsc3R0Zmll
cnhmMXB4cDB2Mjk5cDl2dHZkZWV0IiB0aW1lc3RhbXA9IjE1NTUwNzgyMTEiPjE1MTwva2V5Pjwv
Zm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlw
ZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+U3RydW5rLCBLYXRoZXJpbmUgTy48L2F1
dGhvcj48YXV0aG9yPk1hcnNoLCBKdWxpZSBBLjwvYXV0aG9yPjxhdXRob3I+SGFzaGltLCBBeWVz
aGEgSy48L2F1dGhvcj48YXV0aG9yPkJ1c2gtTWVjZW5hcywgU3VzYW48L2F1dGhvcj48YXV0aG9y
PldlaW5zdGVpbiwgVHJhY2V5PC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3JzPjx0aXRs
ZXM+PHRpdGxlPlRoZSBJbXBhY3Qgb2YgVHVybmFyb3VuZCBSZWZvcm0gb24gU3R1ZGVudCBPdXRj
b21lczogRXZpZGVuY2UgYW5kIEluc2lnaHRzIGZyb20gdGhlIExvcyBBbmdlbGVzIFVuaWZpZWQg
U2Nob29sIERpc3RyaWN0PC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkVkdWNhdGlvbiBGaW5hbmNl
IGFuZCBQb2xpY3k8L3NlY29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10
aXRsZT5FZHVjYXRpb24gRmluYW5jZSBhbmQgUG9saWN5PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNh
bD48cGFnZXM+MjUxLTI4MjwvcGFnZXM+PHZvbHVtZT4xMTwvdm9sdW1lPjxudW1iZXI+MzwvbnVt
YmVyPjxkYXRlcz48eWVhcj4yMDE2PC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91cmxzPjwvcmVjb3Jk
PjwvQ2l0ZT48Q2l0ZT48QXV0aG9yPkhlaXNzZWw8L0F1dGhvcj48WWVhcj4yMDE4PC9ZZWFyPjxS
ZWNOdW0+MTUzPC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj4xNTM8L3JlYy1udW1iZXI+PGZv
cmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYy
OTlwOXZ0dmRlZXQiIHRpbWVzdGFtcD0iMTU1NTA3ODgwMyI+MTUzPC9rZXk+PC9mb3JlaWduLWtl
eXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmli
dXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5IZWlzc2VsLCBKZW5uaWZlciBBLjwvYXV0aG9yPjxhdXRo
b3I+TGFkZCwgSGVsZW4gRi48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxl
cz48dGl0bGU+U2Nob29sIFR1cm5hcm91bmQgaW4gTm9ydGggQ2Fyb2xpbmE6IEEgUmVncmVzc2lv
biBEaXNjb250aW51aXR5IEFuYWx5c2lzPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkVjb25vbWlj
cyBvZiBFZHVjYXRpb24gUmV2aWV3PC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlvZGlj
YWw+PGZ1bGwtdGl0bGU+RWNvbm9taWNzIG9mIEVkdWNhdGlvbiBSZXZpZXc8L2Z1bGwtdGl0bGU+
PC9wZXJpb2RpY2FsPjx2b2x1bWU+NjI8L3ZvbHVtZT48bnVtYmVyPjMwMi0zMjA8L251bWJlcj48
ZGF0ZXM+PHllYXI+MjAxODwveWVhcj48L2RhdGVzPjx1cmxzPjwvdXJscz48L3JlY29yZD48L0Np
dGU+PC9FbmROb3RlPn==
ADDIN EN.CITE.DATA (Zimmer, Henry, and Kho 2017, Strunk et al. 2016, Heissel and Ladd 2018). These mixed results suggest expanding the turnaround literature could refine understandings of the factors leading to successful or unsuccessful reforms. Evaluation of the Innovation Schools program provides an opportunity to explore the understudied relationship between managerial autonomy and school performance in public education, and how this autonomy interacts with turnaround strategies. Autonomy can be understood as decentralization of control from higher levels of governance, in this case school districts, to lower levels, in this case schools. The operational principle underlying provision of autonomy is that lower levels of governance can use discretion to implement policies that are more responsive to conditions at the point of service delivery. Since the key policy lever in this program is a waiver from district policies, the program allows schools to employ discretion to implement independent managerial reforms that are more responsive to student needs than standard district practices. Managerial autonomy was linked to performance improvements in a study of British public schools ADDIN EN.CITE <EndNote><Cite><Author>Clark</Author><Year>2009</Year><RecNum>90</RecNum><DisplayText>(Clark 2009)</DisplayText><record><rec-number>90</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1540311431">90</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Clark, Damon</author></authors></contributors><titles><title>The Performance and Competitive Effects of School Autonomy</title><secondary-title>Journal of Political Economy</secondary-title></titles><periodical><full-title>Journal of Political Economy</full-title></periodical><pages>745-783</pages><volume>117</volume><number>4</number><dates><year>2009</year></dates><urls></urls></record></Cite></EndNote>(Clark 2009), however studies of public schools in Chicago and Boston showed null effects on academic performance ADDIN EN.CITE <EndNote><Cite><Author>Abdulkadiroglu</Author><Year>2011</Year><RecNum>24</RecNum><DisplayText>(Abdulkadiroglu et al. 2011, Steinberg 2014)</DisplayText><record><rec-number>24</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534446649">24</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Abdulkadiroglu, Atila</author><author>Angrist, Joshua D.</author><author>Dynarski, Susan M.</author><author>Kane, Thomas J.</author></authors></contributors><titles><title>Accountability and Flexibility in Public Schools: Evidence from Boston's Charters and Pilots</title><secondary-title>Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>Quarterly Journal of Economics</full-title></periodical><pages>699-748</pages><volume>126</volume><number>2</number><dates><year>2011</year></dates><urls></urls></record></Cite><Cite><Author>Steinberg</Author><Year>2014</Year><RecNum>115</RecNum><record><rec-number>115</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1541978356">115</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Steinberg, Matthew P.</author></authors></contributors><titles><title>Does Greater Autonomy Improve School Performance? Evidence from a Regression Discontinuity Analysis in Chicago</title><secondary-title>Education Finance and Policy</secondary-title></titles><periodical><full-title>Education Finance and Policy</full-title></periodical><pages>1-35</pages><volume>9</volume><number>1</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>(Abdulkadiroglu et al. 2011, Steinberg 2014). Evaluation of the Innovation Schools reform can assess the efficacy of autonomy-based interventions as a school turnaround strategy. I evaluate the DPS Innovation Schools intervention, providing estimates of treatment effects on academic outcomes and how these effects develop over time. Using a two-way fixed-effects difference-in-differences design, this study assesses how school performance changes in schools transitioning to Innovation status, a result which assesses the efficacy of the program as a managerial reform to improve struggling schools. Results suggest that standardized test scores in math, reading and writing improved by 0.1 to 0.3 standard deviations (sd) in public schools that transitioned to Innovation status, indicating the program rapidly improved performance. Effects greater than 0.2 sd are considered large relative to other studies in the educational literature, according to the framework of ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Kraft</Author><Year>2020</Year><RecNum>193</RecNum><DisplayText>Kraft (2020)</DisplayText><record><rec-number>193</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1591362279">193</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kraft, Matthew A.</author></authors></contributors><titles><title>Interpreting Effect Sizes of Education Interventions</title><secondary-title>Educational Researcher</secondary-title></titles><periodical><full-title>Educational Researcher</full-title></periodical><pages>241-253</pages><volume>49</volume><number>4</number><dates><year>2020</year></dates><urls></urls></record></Cite></EndNote>Kraft (2020). However, exploration of how these effects develop over time suggests program effects peaked during year 2 of implementation and declined steeply in following years. The finding that turnaround effects are susceptible to fade-out is a new insight that can inform future turnaround evaluations, and is a major contribution of this study. This study provides evidence on the effects of managerial autonomy on public school performance and the potential for turnaround reforms to drive improvements in struggling urban schools. Denver Innovation SchoolsIn 2008, the State of Colorado passed the Innovation Schools Act, which allowed public schools to submit formal plans to improve student achievement through innovative managerial reforms. Upon approval, Innovation Schools were provided waivers exempting them from district policies, allowing greater autonomy over staffing and other operational practices. While the Innovation Schools Act allowed all Colorado school districts to create programs, Denver Public Schools (DPS) was the first to implement it and authorized more Innovation Schools than any other district. While 3 other districts implemented programs during the period of this study, only 1 had a significant number of Innovation schools, models were inconsistent between districts, and the majority of schools opened late in the period of this study (during the 2012-13 academic year). For these reasons, this study focuses only on the Denver reform. The Innovation Schools program was part of a broader DPS strategy using school choice and accountability to drive performance improvements. These reforms were implemented to address a performance crisis in the district, which drove enrollment declines that threatened the sustainability of district programming. The Denver reforms reversed enrollment loss, with enrollment growing from about 80,000 to 90,000 during the period of this study spanning the 2006-07 to 2013-14 academic years. Evaluation of Denver’s reform model suggests that school choice initiatives associated with the reform improved achievement and shows that the DPS context and portfolio system offers enough variation in governance regimes and a large enough sample size for analysis as a self-contained ecosystem ADDIN EN.CITE <EndNote><Cite><Author>Abdulkadiroglu</Author><Year>2017</Year><RecNum>28</RecNum><DisplayText>(Abdulkadiroglu et al. 2017)</DisplayText><record><rec-number>28</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534618148">28</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Abdulkadiroglu, Atila</author><author>Angrist, Joshua D.</author><author>Narita, Yusuke</author><author>Pathak, Parag A.</author></authors></contributors><titles><title>Research Design Meets Market Design: Using Centralized Assignment for Impact Evaluation</title><secondary-title>Econometrica</secondary-title></titles><periodical><full-title>Econometrica</full-title></periodical><pages>1373-1432</pages><volume>85</volume><number>5</number><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>(Abdulkadiroglu et al. 2017).Denver Innovation Schools are public schools that are provided exemptions from district policies in order to pursue innovative managerial reforms. While Innovation Schools are one of many school models available to students in the DPS, they are not charter schools. Some Innovation Schools were traditional public schools that petitioned a change to Innovation status, while others were new schools that opened as Innovation Schools. By 2014, approximately 30 schools had attained Innovation status out of approximately 200 in DPS. Figure 1 charts the openings and transitions of these schools. Innovation Schools are distinct from charter schools in their level of managerial autonomy. While Innovation Schools attain some autonomies granted charter schools under Colorado law, districts maintain oversight over the schools and their leadership. Schools receive this status by submitting a comprehensive reform plan subject to district approval. All traditional public schools in DPS are eligible to submit Innovation Plans, and new schools may submit plans to open with the enhanced autonomy offered under the program. While denial of a proposal is possible, there is no documentation of denials during the period of the study. Once the district approves an Innovation Plan, schools are subject to performance review and renewal every 3 years. During the period of this study, no Innovation Schools dropped out of the program. ADDIN EN.CITE <EndNote><Cite><Author>Colorado Department of Education</Author><Year>2015</Year><RecNum>11</RecNum><DisplayText>(Colorado Department of Education 2015)</DisplayText><record><rec-number>11</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534191502">11</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>Colorado Department of Education,</author></authors></contributors><titles><title>2015 Innovation Report</title></titles><dates><year>2015</year></dates><publisher>Colorado Department of Education</publisher><urls><related-urls><url>;(Colorado Department of Education 2015)Under Innovation Plans, schools request waivers from district policies. The most common waivers are for policies related to human resources practices, school day and year length, and curriculum requirements. Schools may not waive policies related to accountability requirements, student safety, or teacher retirement. Table 1 lists the types of waivers received and percentage of DPS Innovation Schools receiving them ADDIN EN.CITE <EndNote><Cite><Author>Colorado Department of Education</Author><Year>2013</Year><RecNum>168</RecNum><DisplayText>(Colorado Department of Education 2013)</DisplayText><record><rec-number>168</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1559848290">168</key></foreign-keys><ref-type name="Web Page">12</ref-type><contributors><authors><author>Colorado Department of Education,</author></authors></contributors><titles><title>Colorado Innovation School Act</title></titles><dates><year>2013</year></dates><urls><related-urls><url>;(Colorado Department of Education 2013). Innovation Schools receive waivers from policies relating to budgeting, including teacher collective bargaining, which they use to adjust teacher compensation, allocate resources for critical hires, compensate teachers for added instructional time, and contract out for services like food, maintenance and security. The powers to waive teacher collective bargaining were granted under the Innovation Schools Act. Schools receive instructional waivers which they use to implement innovative curricula including project-based and student-centered learning, augment instruction with educational technologies, and offer electives and enrichment such as concurrent enrollment programs. Innovation schools use human resource management waivers to create their own teacher evaluation systems, pay for performance schemes, and to protect their teachers from being reassigned to other schools. All Innovation Schools sought scheduling waivers to allow them to extend the school day and school year, most commonly to provide an extra 30 minutes per day of instruction in math and ELA. Finally, some schools received waivers from restrictions on receiving non-tax revenue, which allowed them to create foundations and engage in external fundraising ADDIN EN.CITE <EndNote><Cite><Author>Colorado Department of Education</Author><Year>2010</Year><RecNum>127</RecNum><DisplayText>(Colorado Department of Education 2010)</DisplayText><record><rec-number>127</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1543418721">127</key></foreign-keys><ref-type name="Web Page">12</ref-type><contributors><authors><author>Colorado Department of Education,</author></authors></contributors><titles><title>Annual Report Concerning Districts of Innovation</title></titles><dates><year>2010</year></dates><urls><related-urls><url>;(Colorado Department of Education 2010). Some used these extra funds to engage outside consultants to help with strategy and program development. One example was the Denver Summit Schools Network (DSSN), which contracted with the Blueprint Schools Network (BPSN) (), a private nonprofit educational management organization. BPSN worked in partnership with Harvard University’s EdLabs to lead turnaround reforms in struggling public schools according management principles informed by the ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014) research agenda. The DSSN was initiated in the 2011-12 academic year with 11 schools (2 of which transitioned from traditional public schools to Innovation status and are included in the treatment group of this study) and operated for the duration of the study period. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014) evaluated the impacts of Denver DSSN schools, some of which were Innovation Schools, and found positive impacts. DSSN schools paid $800,000 per year to the BPSN, funded in part by $6.7 million in School Improvement Grant (SIG) funding and $4.2 million in private fundraising (none of the SIG funded DSSN schools are included in the treatment group of this study, and only 1 treatment school was SIG-funded (Trevista at Horace Mann)). DSSN schools benefitted from 75 full time math tutors, students began the school year 6 days early and spent an extra hour in class, and most of the staff and teachers in the schools were replaced. ADDIN EN.CITE <EndNote><Cite><Author>Tomassini</Author><Year>2012</Year><RecNum>124</RecNum><DisplayText>(Tomassini 2012, Robles 2011)</DisplayText><record><rec-number>124</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1543092453">124</key></foreign-keys><ref-type name="Newspaper Article">23</ref-type><contributors><authors><author>Tomassini, Jason</author></authors></contributors><titles><title>Denver Turnaround Initiative Showing Achievement Gains</title><secondary-title>Education Week</secondary-title></titles><dates><year>2012</year><pub-dates><date>June 5, 2012</date></pub-dates></dates><urls><related-urls><url> app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1543277002">126</key></foreign-keys><ref-type name="Newspaper Article">23</ref-type><contributors><authors><author>Robles, Yesenia</author></authors></contributors><titles><title>Anxious Eyes on Denver's Far-Northeast Region as School Starts</title><secondary-title>The Denver Post</secondary-title></titles><dates><year>2011</year><pub-dates><date>August 10, 2011</date></pub-dates></dates><urls><related-urls><url>;(Tomassini 2012, Robles 2011) To understand the impacts of different types of Innovation Schools and make a contribution beyond ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014), I estimate separate models that assess whether there were heterogenous effects for DSSN- and SIG-funded schools. In addition to district oversight of Innovation Plans, Innovation Schools are subject to the Denver-specific accountability system called the School Performance Framework (SPF). The SPF rates schools annually based on standardized testing performance. In 2014, 10 Innovation schools were rated at the lowest performance rating and only 1 graduated from a turnaround rating in the prior period. However, most of the Innovation Schools in Turnaround Status were schools that had opened as Innovation Schools. Of the 13 schools that transitioned from traditional public school to Innovation status, only 2 were rated turnaround, and some showed evidence of improvement. The experience of schools transitioning to Innovation Status is most relevant to understanding the success of the program as a managerial intervention to turn around struggling schools. Improving performance under SPF factors heavily into evaluation and renewal of Innovation Plans every 3 years ADDIN EN.CITE <EndNote><Cite><Author>Colorado Department of Education</Author><Year>2015</Year><RecNum>11</RecNum><DisplayText>(Colorado Department of Education 2015)</DisplayText><record><rec-number>11</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534191502">11</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>Colorado Department of Education,</author></authors></contributors><titles><title>2015 Innovation Report</title></titles><dates><year>2015</year></dates><publisher>Colorado Department of Education</publisher><urls><related-urls><url>;(Colorado Department of Education 2015).The reforms in DPS mirror a national trend towards portfolio management (PM) reforms, which have been implemented in cities including Los Angeles, Chicago, Indianapolis, and New Orleans. In these reforms, school districts allow low performing schools varying levels of autonomy over staffing, curriculum, budgeting and operational decision-making. This autonomy is generally coupled with accountability requirements: if schools fail to improve performance, their autonomous status may be revoked ADDIN EN.CITE <EndNote><Cite><Author>Marsh</Author><Year>2013</Year><RecNum>150</RecNum><DisplayText>(Marsh, Strunk, and Bush 2013)</DisplayText><record><rec-number>150</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1555078053">150</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Marsh, Julie A.</author><author>Strunk, Katherine o.</author><author>Bush, Susan</author></authors></contributors><titles><title>Portfolio District Reform Meets School Turnaround: Early Implementation Findings from the Los Angeles Public School Choice Initiative</title><secondary-title>Journal of Educational Administration</secondary-title></titles><periodical><full-title>Journal of Educational Administration</full-title></periodical><pages>498-527</pages><volume>51</volume><number>4</number><dates><year>2013</year></dates><urls></urls></record></Cite></EndNote>(Marsh, Strunk, and Bush 2013). PM reforms were evaluated in Los Angeles with disappointing achievement results ADDIN EN.CITE <EndNote><Cite><Author>Strunk</Author><Year>2016</Year><RecNum>151</RecNum><DisplayText>(Strunk et al. 2016)</DisplayText><record><rec-number>151</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1555078211">151</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Strunk, Katherine O.</author><author>Marsh, Julie A.</author><author>Hashim, Ayesha K.</author><author>Bush-Mecenas, Susan</author><author>Weinstein, Tracey</author></authors></contributors><titles><title>The Impact of Turnaround Reform on Student Outcomes: Evidence and Insights from the Los Angeles Unified School District</title><secondary-title>Education Finance and Policy</secondary-title></titles><periodical><full-title>Education Finance and Policy</full-title></periodical><pages>251-282</pages><volume>11</volume><number>3</number><dates><year>2016</year></dates><urls></urls></record></Cite></EndNote>(Strunk et al. 2016). However, PM reforms in New Orleans, where the district was almost entirely converted to charter schools with significant managerial autonomy, demonstrated performance improvements ADDIN EN.CITE <EndNote><Cite><Author>Harris</Author><Year>2016</Year><RecNum>169</RecNum><DisplayText>(Harris and Larsen 2016)</DisplayText><record><rec-number>169</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1561143345">169</key></foreign-keys><ref-type name="Report">27</ref-type><contributors><authors><author>Harris, Douglas N.</author><author>Larsen, Matthew</author></authors></contributors><titles><title>The Effects of New Orleans Post-Katrina School Reforms on Student Academic Outcomes</title></titles><dates><year>2016</year></dates><publisher>Education Research Alliance for New Orleans</publisher><urls><related-urls><url>;(Harris and Larsen 2016). PM reforms are closely related to school turnarounds, which leverage managerial reforms to drive rapid improvement in struggling schools. Turnarounds have shown evidence of success in a number of instances PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5DYXJsc29uPC9BdXRob3I+PFllYXI+MjAxODwvWWVhcj48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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5DYXJsc29uPC9BdXRob3I+PFllYXI+MjAxODwvWWVhcj48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ADDIN EN.CITE.DATA (Carlson and Lavertu 2018, Schueler, Goodman, and Deming 2017, Dee 2012), though the results in sum have been mixed PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IZWlzc2VsPC9BdXRob3I+PFllYXI+MjAxODwvWWVhcj48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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5IZWlzc2VsPC9BdXRob3I+PFllYXI+MjAxODwvWWVhcj48
UmVjTnVtPjE1MzwvUmVjTnVtPjxEaXNwbGF5VGV4dD4oSGVpc3NlbCBhbmQgTGFkZCAyMDE4LCBE
cmFnb3NldCBldCBhbC4gMjAxNywgWmltbWVyLCBIZW5yeSwgYW5kIEtobyAyMDE3KTwvRGlzcGxh
eVRleHQ+PHJlY29yZD48cmVjLW51bWJlcj4xNTM8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48
a2V5IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYyOTlwOXZ0dmRlZXQi
IHRpbWVzdGFtcD0iMTU1NTA3ODgwMyI+MTUzPC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBl
IG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhv
cnM+PGF1dGhvcj5IZWlzc2VsLCBKZW5uaWZlciBBLjwvYXV0aG9yPjxhdXRob3I+TGFkZCwgSGVs
ZW4gRi48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxlcz48dGl0bGU+U2No
b29sIFR1cm5hcm91bmQgaW4gTm9ydGggQ2Fyb2xpbmE6IEEgUmVncmVzc2lvbiBEaXNjb250aW51
aXR5IEFuYWx5c2lzPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkVjb25vbWljcyBvZiBFZHVjYXRp
b24gUmV2aWV3PC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlvZGljYWw+PGZ1bGwtdGl0
bGU+RWNvbm9taWNzIG9mIEVkdWNhdGlvbiBSZXZpZXc8L2Z1bGwtdGl0bGU+PC9wZXJpb2RpY2Fs
Pjx2b2x1bWU+NjI8L3ZvbHVtZT48bnVtYmVyPjMwMi0zMjA8L251bWJlcj48ZGF0ZXM+PHllYXI+
MjAxODwveWVhcj48L2RhdGVzPjx1cmxzPjwvdXJscz48L3JlY29yZD48L0NpdGU+PENpdGU+PEF1
dGhvcj5EcmFnb3NldDwvQXV0aG9yPjxZZWFyPjIwMTc8L1llYXI+PFJlY051bT4zNTwvUmVjTnVt
PjxyZWNvcmQ+PHJlYy1udW1iZXI+MzU8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFw
cD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxzdHRmaWVyeGYxcHhwMHYyOTlwOXZ0dmRlZXQiIHRpbWVz
dGFtcD0iMTUzNTAzNzMwNiI+MzU8L2tleT48L2ZvcmVpZ24ta2V5cz48cmVmLXR5cGUgbmFtZT0i
UmVwb3J0Ij4yNzwvcmVmLXR5cGU+PGNvbnRyaWJ1dG9ycz48YXV0aG9ycz48YXV0aG9yPkRyYWdv
c2V0LCBMaXNhPC9hdXRob3I+PGF1dGhvcj5UaG9tYXMsIEphaW1lPC9hdXRob3I+PGF1dGhvcj5I
ZXJybWFubiwgTWFyaWVzYTwvYXV0aG9yPjxhdXRob3I+RGVrZSwgSm9objwvYXV0aG9yPjxhdXRo
b3I+SmFtZXMtQnVyZHVteSwgU3VzYW5uZTwvYXV0aG9yPjxhdXRob3I+R3JhY3pld3NraSwgQ2hl
cnlsPC9hdXRob3I+PGF1dGhvcj5Cb3lsZSwgQW5kcmVhPC9hdXRob3I+PGF1dGhvcj5VcHRvbiwg
IFJhY2hlbDwvYXV0aG9yPjxhdXRob3I+VGFuZW5iYXVtLCBDb3VydG5leTwvYXV0aG9yPjxhdXRo
b3I+R2lmZmluLCBKZXNzaWNhPC9hdXRob3I+PC9hdXRob3JzPjx0ZXJ0aWFyeS1hdXRob3JzPjxh
dXRob3I+TWF0aGVtYXRpY2EgUG9saWN5IFJlc2VhcmNoPC9hdXRob3I+PC90ZXJ0aWFyeS1hdXRo
b3JzPjwvY29udHJpYnV0b3JzPjx0aXRsZXM+PHRpdGxlPlNjaG9vbCBJbXByb3ZlbWVudCBHcmFu
dHM6IEltcGxlbWVudGF0aW9uIGFuZCBFZmZlY3RpdmVuZXNzPC90aXRsZT48L3RpdGxlcz48ZGF0
ZXM+PHllYXI+MjAxNzwveWVhcj48L2RhdGVzPjxwdWJsaXNoZXI+VS5TLiBEZXBhcnRtZW50IG9m
IEVkdWNhdGlvbjwvcHVibGlzaGVyPjx1cmxzPjxyZWxhdGVkLXVybHM+PHVybD5odHRwczovL2Zp
bGVzLmVyaWMuZWQuZ292L2Z1bGx0ZXh0L0VENTcyMjE1LnBkZjwvdXJsPjwvcmVsYXRlZC11cmxz
PjwvdXJscz48L3JlY29yZD48L0NpdGU+PENpdGU+PEF1dGhvcj5aaW1tZXI8L0F1dGhvcj48WWVh
cj4yMDE3PC9ZZWFyPjxSZWNOdW0+MTU4PC9SZWNOdW0+PHJlY29yZD48cmVjLW51bWJlcj4xNTg8
L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRiLWlkPSJlZDVhdnJ4cGxz
dHRmaWVyeGYxcHhwMHYyOTlwOXZ0dmRlZXQiIHRpbWVzdGFtcD0iMTU1NjExNDE1NyI+MTU4PC9r
ZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0aWNsZSI+MTc8L3Jl
Zi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5aaW1tZXIsIFJvbjwvYXV0aG9y
PjxhdXRob3I+SGVucnksIEdhcnkgVC48L2F1dGhvcj48YXV0aG9yPktobywgQWRhbTwvYXV0aG9y
PjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0bGVzPjx0aXRsZT5UaGUgRWZmZWN0cyBvZiBT
Y2hvb2wgVHVybmFyb3VuZCBpbiBUZW5uZXNzZWUmYXBvcztzIEFjaGlldmVtZW50IFNjaG9vbCBE
aXN0cmljdCBhbmQgSW5ub3ZhdGlvbiBab25lczwvdGl0bGU+PHNlY29uZGFyeS10aXRsZT5FZHVj
YXRpb25hbCBFdmFsdWF0aW9uIGFuZCBQb2xpY3kgQW5hbHlzaXM8L3NlY29uZGFyeS10aXRsZT48
L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5FZHVjYXRpb25hbCBFdmFsdWF0aW9uIGFu
ZCBQb2xpY3kgQW5hbHlzaXM8L2Z1bGwtdGl0bGU+PC9wZXJpb2RpY2FsPjxwYWdlcz42NzAtNjk2
PC9wYWdlcz48dm9sdW1lPjM5PC92b2x1bWU+PG51bWJlcj40PC9udW1iZXI+PGRhdGVzPjx5ZWFy
PjIwMTc8L3llYXI+PC9kYXRlcz48dXJscz48L3VybHM+PC9yZWNvcmQ+PC9DaXRlPjwvRW5kTm90
ZT4A
ADDIN EN.CITE.DATA (Heissel and Ladd 2018, Dragoset et al. 2017, Zimmer, Henry, and Kho 2017). ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014)’s evaluation of DSSN schools, which included Innovation Schools, found significant effects in math and null effects in reading. ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Abdulkadiroglu</Author><Year>2017</Year><RecNum>28</RecNum><DisplayText>Abdulkadiroglu et al. (2017)</DisplayText><record><rec-number>28</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534618148">28</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Abdulkadiroglu, Atila</author><author>Angrist, Joshua D.</author><author>Narita, Yusuke</author><author>Pathak, Parag A.</author></authors></contributors><titles><title>Research Design Meets Market Design: Using Centralized Assignment for Impact Evaluation</title><secondary-title>Econometrica</secondary-title></titles><periodical><full-title>Econometrica</full-title></periodical><pages>1373-1432</pages><volume>85</volume><number>5</number><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>Abdulkadiroglu et al. (2017) evaluated the DPS context in a study of charter school effects, which included a supplementary model isolating effects of Innovation Schools that found null negative and null positive results depending on specification. This study will evaluate 13 schools that transitioned from traditional public schools to Innovation Schools between the 2006-07 and 2013-14 academic years. This variation in Innovation School status within schools allows estimation of a causal effect using a two-way fixed-effects difference-in-differences design. These estimates differ from the estimates of ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014) and ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Abdulkadiroglu</Author><Year>2017</Year><RecNum>28</RecNum><DisplayText>Abdulkadiroglu et al. (2017)</DisplayText><record><rec-number>28</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534618148">28</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Abdulkadiroglu, Atila</author><author>Angrist, Joshua D.</author><author>Narita, Yusuke</author><author>Pathak, Parag A.</author></authors></contributors><titles><title>Research Design Meets Market Design: Using Centralized Assignment for Impact Evaluation</title><secondary-title>Econometrica</secondary-title></titles><periodical><full-title>Econometrica</full-title></periodical><pages>1373-1432</pages><volume>85</volume><number>5</number><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>Abdulkadiroglu et al. (2017). This study includes a longer panel than either study, with both a longer pre-treatment and post-treatment period. The analytic sample only includes 2 of 7 Innovation Schools evaluated by Fryer, whose study leveraged student-level rather than school-level variation in Innovation Schools status and could therefore analyze schools that opened as Innovation Schools and experienced no variation in treatment. It provides new estimates for 11 schools not evaluated by Fryer, all of which experienced transition to Innovation status, and whose experience can suggest the ability of the program improve performance in struggling schools, consistent with a turnaround model. The study uses a different identification strategy than Abdulkadiroglu and colleagues, who used lottery randomization to estimate student-level achievement effects. It also uses a different sample; lottery randomization allowed them to include Innovation Schools without variation in treatment status in their estimates. Their counterfactual indicates how a student’s performance would differ if he or she attended another district school rather than an Innovation School, while my counterfactual indicates whether Innovation Schools improved from their pre-treatment performance level. It is simultaneously possible for Innovation Schools to improve following treatment, but fail to reach parity with other district schools, which could lead to a negative estimate in the Abdulkadiroglu model. Thus, my identification strategy is most relevant to understanding how the program changes performance within schools over time, rather than how attending an Innovation School impacts student achievement compared to other schools.Since my analysis examines schools that transitioned from a traditional public school to Innovation status, my results reveal whether the Innovation Schools program can turn around struggling public schools, rather than how the performance of students entering an Innovation School changes. My analysis probes the sensitivity of Innovation School effects to a greater extent than prior studies, which addressed the program briefly in supplementary analyses. I pay careful attention to parallel trends assumptions, non-random selection into treatment, treatment heterogeneity and inferential difficulties arising from the small number of schools experiencing change in treatment status. I demonstrate that managerial autonomy leading to innovative reforms may be an effective strategy to turn around low-performing urban schools, but that performance improvements may deteriorate without continuous effort to sustain results. DataThis study uses data collected alongside annual accountability requirements by the Colorado Department of Education (CDE), which are publicly available on their data website (cde.state.co.us/cdereval). I extract data based on Enrollment, Demographics, Resources, Staffing and Achievement from the 2006-07 to 2013-14 Academic Years, resulting in a final sample of 148 schools in DPS, including all schools which transitioned to Innovation status.The study estimates the effect of Innovation Schools treatment on academic achievement, using test scores on the CSAP (Colorado Student Assessment) and TCAP (Transitional Colorado Assessment Program) exams, Colorado’s year end standardized tests in writing, reading and math, taken by students in grades 3-10. The CSAP changed to TCAP in the 2011-12 academic year, but the scale and distribution of test scores changed very little and there is little evidence of differential changes pertaining to treated and untreated schools. These measures were available as school level means through the CDE’s data portal ADDIN EN.CITE <EndNote><Cite><Author>Colorado Department of Education</Author><Year>n.d.</Year><RecNum>192</RecNum><DisplayText>(Colorado Department of Education n.d.)</DisplayText><record><rec-number>192</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1591303823">192</key></foreign-keys><ref-type name="Web Page">12</ref-type><contributors><authors><author>Colorado Department of Education,</author></authors></contributors><titles><title>SchoolView</title></titles><number>June 22, 2019</number><dates><year>n.d.</year></dates><urls><related-urls><url>;(Colorado Department of Education n.d.). As school-level averages of an exam with uniform scale between grades, these measures are comparable across all schools including elementary, middle, and high schools. To remove temporal and distributional trends, I standardize these measures with respect to full sample (all DPS schools) by year consistent with prevailing norms for standardization of test scores in the educational policy literature. I also test for effects on non-academic and organizational outcomes using the following dependent variables: enrollment, student-teacher ratio, teacher salary (adjusted to 2016 dollars) and discipline (a measure of combined suspensions and expulsions during the academic year). I am missing data from the 2006-07 academic year for some of these variables. The organizational models, accordingly have one less pre-treatment year. The study uses a binary indicator of treatment status coded 1 following approval of an Innovation Plan and 0 otherwise. This is equivalent to an interaction between a binary indicator of membership in the Innovation Schools treatment group, and binary indicator for each school’s post-treatment period. While there were 30 Innovation Schools in DPS during the period of the study, only 13 demonstrate variation in treatment status indicating transition from a traditional public school to an Innovation school (4 elementary Schools, 5 schools with elementary and middle grades, 1 middle school, 1 school with middle and high school grades, and 2 high schools). The other 17 schools were new schools that opened as Innovation Schools, and thus do not have a pre-treatment period. Since fixed-effects models use only within-panel variation, Innovation Schools without treatment variation will not contribute to identification and will enter into the estimate as equivalent to untreated schools. To preserve homogeneity of treated and comparison groups, I exclude all Innovation Schools with no variation in treatment status from the analysis. The results with this specification are fundamentally equivalent to models that include Innovation Schools with time-invariant treatment status. The final treatment group sample includes all 13 DPS schools that transitioned from a traditional public school to an Innovation School during the period of the study, and therefore provide within-school variation in treatment status to leverage in a two-way fixed-effects design.Due to limitations of publicly available data, I have few control variables to include in my models. However, I can include enrollment, student teacher ratio, % free lunch eligibility, % black students and % Hispanic students. Given model assumptions hold, lack of rich controls does not prevent identification, though I will be unable to tell how the program impacted the proportion of students with disabilities or English language learners and how this contributed to treatment effects. This is a limitation of my analysis. However, I can test for changes in composition based on % free lunch, % black and % Hispanic students (Appendix Table 1) and find no evidence of changes in student composition.Methods Ordinary least squares estimates of Innovation School impacts on school performance may be biased since participation in the program was not randomly assigned and may therefore be correlated with unobserved characteristics of participating schools. To address the endogenous nature of the reform, I specify a two-way fixed-effects difference-in-differences model according to the following specification:yst=γ0Innovation School*Postst+ γ1Xst+θs+τt+ εst(Equation 1)In this equation, yst is an outcome of interest for school s in year t, Xst is a vector of school level demographic and organizational characteristics, and εst is a stochastic error term for school s in year t. The identification strategy leverages within-school transition from a traditional public school to an Innovation School. The measure Innovation School*Postst equals one following transition to Innovation School status and zero prior for each treated school, and zero for all untreated schools in all periods. Since transition occurs at different times for different Innovation Schools, this indicator is equivalent to an interaction between a time invariant indicator of treatment group membership and each school’s unique post-treatment period. θs is a vector of school fixed-effects which absorb all time invariant characteristics of each school, including the binary indicator of treatment group membership included in traditional difference-in-difference models. τt is a vector of year fixed-effects, which absorb temporal factors shared across schools, including the binary indicators for all post-treatment periods for all treatment cohorts included in traditional difference-in-differences models. Since the school fixed-effects absorb time invariant treatment status and the year fixed-effects absorb all possible post treatment periods, γ0 can be interpreted as a difference-in-differences estimator. This functional form attempts to mitigate selection into Innovation Schools treatment to capture treatment effects with reduced bias, and allows for estimation in the presence of staggered treatment initiation. The identification assumptions of this model are that treatment is assigned exogenously conditional on school and year fixed-effects and that treated and untreated schools would have similar outcome trends in the absence of treatment. I assess the plausibility of these assumptions by showing robustness of effects to different comparison groups, by testing for parallel trends violations in the pre-treatment period with placebo tests, event-study regressions and graphical analyses, and by probing robustness to school-specific linear time trends and lagged dependent variables. For the placebo tests I estimate treatment effect models with “placebo” treatment dummies assigned to the treatment group in each pre-treatment year when no actual intervention existed. These dummies are equivalent to an interaction between a time-invariant treatment group indicator and a full vector of year fixed-effects, consistent with the following specification (All terms equivalent to Equation 1, γ0 is a vector of placebo coefficients):yst=γ0Innovation School*Year=2007-08Year=2012-13Yearst+ γ1Xst+θs+τt+ εst(Equation 2)Treated schools are removed from the sample as they enter a genuine treatment period; the panel terminates in the 2012-13 academic year. Estimating a significant effect in a placebo period, where no treatment existed, could indicate a parallel trend violation ADDIN EN.CITE <EndNote><Cite><Author>Mora</Author><Year>2017</Year><RecNum>15</RecNum><DisplayText>(Mora and Reggio 2017)</DisplayText><record><rec-number>15</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534274027">15</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Mora, Ricardo</author><author>Reggio, Iliana</author></authors></contributors><titles><title>alternative Diff-in-Diffs estimators with Several Pre-Treatment Periods</title><secondary-title>Econometrics Reviews</secondary-title></titles><periodical><full-title>Econometrics Reviews</full-title></periodical><volume>(forthcoming)</volume><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>(Mora and Reggio 2017). I also include event-study regressions by estimating two-way fixed-effects models with treatment dummies indicating each treatment school’s year relative to treatment, ranging from 7 years pre-treatment to 5 years post treatment. The treatment group for these models is equivalent to the main models, consisting of schools who transitioned to Innovation School status. These indicators are coded as one for each Innovation School in its nth year relative to treatment, and zero for Innovation schools not in their nth year relative to treatment or untreated schools. This is equivalent to an interaction between the time invariant treatment group indicator and a binary indicator of each school’s year relative to treatment, consistent with the following specification (All terms equivalent to Equation 1, γ0 is a vector of event-study coefficients): yst=γ0Innovation School*i= -7, i ≠-15Treatment Year ist+ γ1Xst+θs+τt+ εst(Equation 3)Innovation Schools in the first treatment cohort (treated in the 2009-2010 academic year) would be coded as one for 3 pre-treatment indicators and all 5 post-treatment indicators, given an 8-year panel spanning the 2006-07 to 2013-14 academic years. Innovation schools in the final treatment cohort (treated in the 2013-14 academic year) would be coded as one for 7 pre-treatment indicators and 1 post-treatment indicator. Because this requires excluding a year as the reference category, I estimate two different specifications. The primary model excludes the last pre-treatment year (year negative 1) excluded. For robustness, I estimate a second model with the third pre-treatment year excluded (year negative 3), since this is the earliest pre-treatment year shared by every treated school. In the absence of pre-treatment trend violations, we should expect to see null coefficients in the pre-treatment years, especially close to the treatment window, and treatment effects in the post treatment year similar to those in the main models.I also estimate models with school-specific linear time trends. Introduction of a panel-specific trend in difference-in-difference models allows for linear violations of the parallel trends assumption; schools can be moving in opposing paths and identification assumptions are maintained, provided there are no quadratic deviations (a more plausible assumption than no linear deviation) ADDIN EN.CITE <EndNote><Cite><Author>Mora</Author><Year>2017</Year><RecNum>15</RecNum><DisplayText>(Mora and Reggio 2017)</DisplayText><record><rec-number>15</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534274027">15</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Mora, Ricardo</author><author>Reggio, Iliana</author></authors></contributors><titles><title>alternative Diff-in-Diffs estimators with Several Pre-Treatment Periods</title><secondary-title>Econometrics Reviews</secondary-title></titles><periodical><full-title>Econometrics Reviews</full-title></periodical><volume>(forthcoming)</volume><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>(Mora and Reggio 2017). Finally, I estimate models with a lagged dependent variable, which accounts for prior year achievement in the estimates and may provide more robust treatment effect estimates in the presence of parallel trends violations. With a relatively long panel, and an intention to probe robustness to prior achievement rather than estimate causal effects, concerns about Nickell bias likely do not warrant a dynamic panel approach to these supplementary analyses. These models are provided in a robustness checks section.Since schools are not treated randomly, but rather adopt Innovation Schools treatment by choice, and because there are significant differences between treated and untreated schools on observable characteristics, I estimate my models in 2 different subsamples. The first subsample compares Innovation Schools to all other schools. The second subsample is chosen using propensity score matching. I estimate propensity scores using a probit model with all covariates and dependent variables employed in the main models, estimating the propensity scores separately for each treatment cohort starting with the first, restricting the sample to the last pre-treatment year for each treatment cohort, and excluding matched schools from prior cohorts from the potential matches of future cohorts to avoid excessive overlap. I select the 5 nearest neighbor matches with replacement. I then absorb the entire panel of any school that is a nearest neighbor match for any treated school according to this strategy. This yields a matched sample of 62 clusters with the comparison group including the five nearest propensity score matches to each treated school, with some control schools serving as matches to multiple treated schools. I also checked robustness to 2 other comparison groups selected on the basis of free lunch eligibility and test scores, and the results were fundamentally equivalent (results available on request).The literature on school turnarounds shows that effects often grow in the years following implementation ADDIN EN.CITE <EndNote><Cite><Author>Carlson</Author><Year>2018</Year><RecNum>148</RecNum><DisplayText>(Carlson and Lavertu 2018, Sun, Penner, and Loeb 2017)</DisplayText><record><rec-number>148</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1554330740">148</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Carlson, Deven E.</author><author>Lavertu, Stephane</author></authors></contributors><titles><title>School Improvement Grants in Ohio: Effects on Student Achievement and School Administration</title><secondary-title>Educational Evaluation and Policy Analysis</secondary-title></titles><periodical><full-title>Educational Evaluation and Policy Analysis</full-title></periodical><pages>287-315</pages><volume>40</volume><number>3</number><dates><year>2018</year></dates><urls></urls></record></Cite><Cite><Author>Sun</Author><Year>2017</Year><RecNum>149</RecNum><record><rec-number>149</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1555077644">149</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Sun, Min</author><author>Penner, Emily K.</author><author>Loeb, Susanna</author></authors></contributors><titles><title>Resource- and Approach-Driven Multidimensional Change: Three-Year Effects of School Improvement Grants</title><secondary-title>American Educational Research Journal</secondary-title></titles><periodical><full-title>American Educational Research Journal</full-title></periodical><pages>607-643</pages><volume>54</volume><number>4</number><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>(Carlson and Lavertu 2018, Sun, Penner, and Loeb 2017). Alternatively, if turnaround schools decrease effort following an initial implementation period, treatment effects may decrease as time progresses. The event-study models described above allow heterogeneity of treatment effects by number of years since treatment initiation. The estimates on the indicators for the first through fifth years of treatment allow interpretation of the cumulative effects of the program in each year since treatment initiation. I also explore whether the effects of the program were larger or smaller in certain schools. To explore this heterogeneity, I engaged in list-wise deletion of schools from the treatment group. If removing the school decreased either the effect size or significance level of the main math effect, indicating that the school’s performance was contributing to the positive average treatment effect, I added the school to my “High Performer” group. If removing the school increased both the effect size and significance level, I added it to my “Low Performer” group. This allows for interpretation of whether effects of the program were evenly distributed, or whether they masked both successes and failures. Since many of the Innovation Schools were externally managed by the Blueprint Schools Network, a CMO-like private nonprofit management group, it is important to understand how effects of the program are driven by these externally managed schools. To assess this dynamic, I estimate separate treatment effects for Blue Print and non-Blue Print schools. I include these three results in a section on treatment heterogeneity following the main results.All models are estimated with Huber-White robust standard errors clustered by school, to address heteroscedasticity and autocorrelation within schools. However, this has the potential to be inaccurate given the small number of treated clusters in my sample. Significant attention has focused on inference in difference-in-differences models ADDIN EN.CITE <EndNote><Cite><Author>Bertrand</Author><Year>2004</Year><RecNum>12</RecNum><DisplayText>(Bertrand, Duflo, and Mullainathan 2004)</DisplayText><record><rec-number>12</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534271450">12</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Bertrand, Marianne</author><author>Duflo, Esther</author><author>Mullainathan, Sendhil</author></authors></contributors><titles><title>How Much Should We Trust Differences-In-Differences Estimates?</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>249-275</pages><volume>119</volume><number>1</number><dates><year>2004</year></dates><urls></urls></record></Cite></EndNote>(Bertrand, Duflo, and Mullainathan 2004), particularly when the number of clusters or treated clusters are small PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5NYWNraW5ub248L0F1dGhvcj48WWVhcj4yMDE4PC9ZZWFy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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5NYWNraW5ub248L0F1dGhvcj48WWVhcj4yMDE4PC9ZZWFy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ADDIN EN.CITE.DATA (Mackinnon and Webb 2018, Mackinnon 2016, Cameron, Gelbach, and Miller 2008). With a small proportion of treated clusters, cluster robust standard errors can lead to over-rejection in hypothesis testing. Wild bootstrap clustered standard errors can provide more accurate estimates of the standard error with a small number of treated clusters, but can also lead to under-rejection. For this reason, I conduct hypothesis testing on the coefficients of interest using wild bootstrap clustered standard errors and provide the associated p-values corresponding to each estimate. Providing both cluster robust and wild bootstrap hypothesis testing should provide an upper and lower bound for the level of significance of my estimates.ResultsDescriptive StatisticsSummary statistics are provided in Table 2. DPS is a low-performing urban school district, which does not exceed 50% proficiency in writing, reading or mathematics. Students in the district are overwhelmingly poor and minority, with average free or reduced-price lunch eligibility of 71%, average Hispanic composition of 58% and average Black composition of 14%. Schools have 533 students on average; the average student teacher ratio is 17.45, which was above the national average of 16.1 in 2014. DPS has approximately 90,000 students, making them the thirty-fourth largest public-school district in the nation in 2013. Teachers make approximately $53,000 per year, on par with the national average. During the period of the study, DPS spent approximately $6,000 per pupil and Colorado was in the bottom 25% of states by per pupil expenditures. Table 3 presents summary statistics comparing Innovation Schools in the treatment group (those who transitioned to Innovation Status) and non-Innovation schools in the 3-year pretreatment period from the 2006-07 to 2008-09 academic years. Innovation Schools are poorer, more diverse and lower-achieving that non-treated schools. Table 3 also presents comparisons of Innovation Schools to the propensity matched sample used in the analysis. Innovation Schools are almost perfectly matched to comparison schools in this sample on organizational and demographic characteristics. They do have more black and fewer Hispanic students, though only the difference in % black students is statistically significant at the .10 level.Main ResultsIn the following section I estimate difference-in-differences models with school and year fixed-effects, and controls for enrollment, student-teacher ratio, % free lunch eligibility, % black students and % Hispanic students with both cluster robust standard errors and Wild bootstrap clustered standard errors (results are robust to estimation without controls, and these estimates are available upon request). Table 4 presents the main results, showing treatment coefficients for writing, reading and math from left to right, with effects shown for each dependent variable in the full sample and propensity score matched sample. All subsequent treatment effect tables will follow this general format. Effects in writing are statistically insignificant but meaningful at about .1 standard deviations (sd) in test score gains. Effects in reading are significant at the 0.05 level in the full sample and at the 0.10 level in the PSM sample (and at the 0.10 level with wild bootstrap hypothesis testing), and about 0.15 sd in magnitude. Effects in math are significant and measure about 0.25 standard deviations. All are significant at the 0.05 level with cluster robust standard errors, and at the 0.05 level with wild bootstrap clustered errors (p= 0.043, 0.041). According to guidelines established by ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Kraft</Author><Year>2020</Year><RecNum>193</RecNum><DisplayText>Kraft (2020)</DisplayText><record><rec-number>193</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1591362279">193</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kraft, Matthew A.</author></authors></contributors><titles><title>Interpreting Effect Sizes of Education Interventions</title><secondary-title>Educational Researcher</secondary-title></titles><periodical><full-title>Educational Researcher</full-title></periodical><pages>241-253</pages><volume>49</volume><number>4</number><dates><year>2020</year></dates><urls></urls></record></Cite></EndNote>Kraft (2020), effect sizes of educational interventions greater than 0.2 standard deviations are considered large, and fall above the 80th percentile of empirical effect sizes in a review of over 200 studies. As can be seen in systematic reviews of the evaluation literature of educational accountability studies, it is common for educational interventions to show larger results in math than in reading and writing, possibly because math skills are developed in the classroom to a greater extent ADDIN EN.CITE <EndNote><Cite><Author>Figlio</Author><Year>2011</Year><RecNum>79</RecNum><DisplayText>(Figlio and Loeb 2011)</DisplayText><record><rec-number>79</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1539890072">79</key></foreign-keys><ref-type name="Book Section">5</ref-type><contributors><authors><author>Figlio, David</author><author>Loeb, Susanna</author></authors><secondary-authors><author>Hanushek, Eric A.</author><author>Machin, Stephen</author><author>Woessman, Ludger</author></secondary-authors></contributors><titles><title>School Accountability</title><secondary-title>Handbook of the Economics of Education</secondary-title></titles><pages>383-421</pages><volume>3</volume><dates><year>2011</year></dates><pub-location>The Netherlands: North-Holland</pub-location><publisher>Elsevier</publisher><urls></urls></record></Cite></EndNote>(Figlio and Loeb 2011). Treatment HeterogeneityTable 5 estimates event-study models which reveal treatment effect heterogeneity by years since treatment initiation. (Model coefficients are plotted in figures 5-7. Since coefficients are roughly equivalent in both estimation samples, I only plot the full sample models. Appendix Table 3 and figures 8-10 show the same results with an alternate reference category (earliest shared pre-treatment year excluded) with equivalent results.) These models show large effects developing into year two of program implementation, but indicate trouble sustaining results. The math results show positive effects of approximately 0.15 sd in the first year of treatment, which grow to approximately 0.3 sd in the second and third years. Effects in the fourth year remain large, though marginally significant, but effects regress to approximately zero in the 5th year. In writing, effects increase to statistical significance in the second year of treatment, measuring approximately 0.2 sd, but then recede dramatically in the following years, fully regressing to zero. Effects in reading grow to significance in the second year as well, measuring approximately 0.15 sd, and then follow a similar receding pattern. These results suggest that the effects of Innovation Schools treatment grow over time, especially leading into the second and third years of implementation, but raise concerns about fade-out suggesting difficulty sustaining the success of the program. The treatment group in these models is not stable across post treatment years; for example, while all treatment schools have a first post-treatment year, only the first treated cohort has a fifth post-treatment year due to the staggered timing of treatment initiation. Results by treatment cohort (available on request) show a similar fade-out pattern across treatment cohorts, though the second treatment cohort showed more success at sustaining results. I also perform exploratory analysis of heterogeneity in the size of treatment effects. The listwise deletion procedure outlined in the methods section left me with a treatment group of 8 “High Performer” schools that showed dramatic improvements under the Innovation Schools program (see Table 6). These schools were Cole Arts and Sciences Academy, Godsman Elementary School, Green Valley Elementary School, McGlone Elementary School, Montclair School of Academics and Enrichment, Trevista at Horace Mann and Valdez Elementary School and Whitter K-8 school. The effects on these schools were large and highly significant ranging from 0.25 sd in writing to 0.4 sd in math, all significant at the 0.01 level. Most of these schools are elementary schools and middle schools, with younger students who may be more receptive to interventions. I estimate treatment effects for the remaining group, labeled “Low Performers.” This group of 5 schools showed negative effects in all subjects, some of which are statistically significant at the 0.10 level. This group includes Centennial (A School for Expeditionary Learning), Grant Beacon Middle School, Manual High School, Summit Academy, and Martin Luther King Jr. Early College. Centennial only had 1 year of treatment, so their poor performance could have been a result of transitional disruptions. Most of these schools include high school grades, where students might be more resistant to interventions or more likely to be impacted by disruptions. These models do not exclude performance declines driven by other factors, and these schools may have had similar declines in the absence of Innovation Schools treatment. These exploratory analyses should be interpreted as descriptive, as its uncertain whether parallel trends assumptions hold for these restricted performance groups. Three of 8 high performers belonged to the DSSN (Green Valley Elementary School and McGlone Elementary School) or were SIG-funded (Trevista at Horace Mann). This group was selected because every school contributed to the positive effects of the program, but it is important to assess the dependency of the results on these externally supported schools. Main results are not robust to exclusion of DSSN and SIG schools, but are also not robust to excluding the 5 high-performing district schools and including only externally supported schools (results available upon request). Table 7 modifies the “high-performer” group to exclude DSSN and SIG schools to assess changes in effect sizes and significance. Exclusion of DSSN schools reduces effect size by 0.05 to 0.1 sd in all subjects, but results remain highly significant and 0.2 sd or larger. Further exclusion of the SIG funded school leads to similar results, though effects in Reading are larger than in the full sample. These results suggest that DSSN schools may have done slightly better than other high-performing Innovation Schools, and the SIG funded school did about the same or slightly worse. This suggests that external funding or management does not explain the effects of the program in high-performing Innovation Schools.Robustness ChecksTo probe the robustness of these findings, I first interrogate the parallel trends assumption. First, I investigate this condition graphically. In figures 2-4, I present graphs of writing, reading and math scores over time in both comparison and treatment groups and in the multiple samples. I include a vertical line in 2010 which is the first year of Innovation Schools treatment. The graphs show some common tendencies across all variable. The Innovation Schools cohort seems to have experienced a large increase in performance in the first year of the study, 2006-07, which is 3 years pre-treatment. After that, the cohort levels out and achieves a generally common trend with the comparison group, until 2009-10 when a sizeable treatment effect appears to develop. Then, in the later years of the panel, the visible treatment effect appears to recede, consistent with the findings of the event-study models. The finding that the results of the treatment effect models can be clearly visualized in the graphical plots lends confidence in the main findings. Some concerns remain about pre-treatment spikes and irregularities in the test score plots, though it should be emphasized that these are unadjusted trend lines, and not causal models. Upon closer investigation, the “performance spike” results in part because one of the treatment schools, Manual High School was closed in the 2006-07 year, and therefore did not have data. Since Manual High School has higher mean test scores than the other Innovation Schools, its addition in 2007-08 increased mean performance, which accounts for much of the spike. Trend plots that exclude Manual High School (available on request) show roughly parallel trends in the pre-treatment period for the writing and math variables, though a less pronounced spike persists in reading. The trend plots show suggestive evidence to support identification assumptions, with a stable or downward trajectory immediately pre-treatment followed by a rapid performance increase post-treatment, but to bolster this analysis, I test statistically for parallel pre-trends in the following section using the placebo testing process outlined in the methods sectionThe placebo testing procedure used in this paper hinges on the premise that, since placebo treatments are assigned in the pre-treatment period when no actual treatment occurred, their inclusion in the model should result in a null coefficient. Null results on all placebo coefficients suggests support for parallel trends assumptions. The placebo test results are included in Appendix Table 2. These tests perform fairly well, with no placebo coefficient rejected at the 0.05 level in any sample for any variable in the first five years of placebo treatment. In the sixth year we see rejection in all models, but the treated group for this coefficient only includes 1 school and one school-year observation. The success of these tests provides statistical evidence in support of parallel trends assumptions. Event-study regressions in Table 5 & Appendix Table 3 show a similar trend. The models with year negative one excluded perform well, with only 3 out of 42 coefficients rejected at the 0.05 level and 1 rejected at the 0.10 level, results that could be expected due to random chance. All rejected coefficients are in year negative 7 and negative 6, which only include 1 and 2 treated schools which are more than 5 years from treatment initiation, when pre-trend violations are unlikely to influence results so far in the future. The models with year negative 3 omitted perform somewhat worse, though the rejections are concentrated in years negative 5-7, when the treatment groups only include, 1, 2 and 5 treated schools, and where the schools are far from the treatment window. In the 4 pre-treatment years immediately preceding treatment there are only 2 rejections at the 0.10 level out of 18 coefficients, a result that would be expected due to random chance. Event-study plots in Figures 5-10 show very similar trends, with pre-treatment plots stable in the first 5 pre-treatment year and overlapping with zero virtually everywhere, with treatment effects developing into the second post treatment year and declining thereafter. These tests support the validity of parallel trends assumptions and suggest treatment effects aren’t being driven by selection of pre-treatment period in the main model or event-study regressions. Next, I estimate models with school-specific linear time trends (Appendix Table 4). The results in math yield slightly to this test. However, both models are significant at the 0.10 level, as are estimates in writing. Effect sizes are roughly equivalent to the main models, which suggests that trend violations are not meaningfully biasing the coefficient estimates in the main models. We may interpret the reduced significance as resulting in part from larger standard errors, which are produced by the extra demands that a full vector of school-specific trends (1 per panel) places on the data. Effects may also be less precisely estimated as de-trending the data removes useful variation along with confounding trends. The robustness of the effect sizes suggests evidence of a true performance gain in Innovation Schools, though increased demands on the data cause these estimates to be less precise. Finally, I estimate models with lagged dependent variables in Appendix Table 5 (since these require lagged data, they have one fewer pre-treatment year). In these models, inclusion of the lagged test scores decreases the writing and reading slightly, though they remain about 0.10 sd. Math results remain significant at the 0.05 level, though they are about 0.05 sd smaller. It is unclear whether any diminishing of effect size or significance is due to inclusion of the lagged dependent variable or shortening of the pre-treatment period, though results are largely consistent with the main models. These findings suggest that the effects in the main models, especially in math, are not being driven by prior year achievement trends. Non-Academic and Organizational OutcomesTo suggest generalizable mechanisms by which the Innovation Schools reform led to performance improvements, I test the effect of Innovation Schools treatment on a number of organizational outcomes (Table 8). While many of the mechanisms reported under the program, such as expanded instructional time, curriculum reform, teacher evaluation, or managerial autonomy in general are unlikely to be observable, some of the available metrics could yield insights if Innovation Schools grew, decreased class sizes, increased salaries through pay for performance, or increased discipline of students. These results should be interpreted as purely descriptive, and I do not explore the plausibility of parallel trends assumptions for these analyses. Innovation Schools may have grown following treatment, with an imprecise effect of 17.58 students in the full sample. However, analysis of changes in student composition, measured by % free lunch eligibility, % black students and % Hispanic students (Appendix Table 1) do not show evidence of significant changes, which suggests enrollment growth did not impact treatment effect estimates through changes in student demographics. Student Teacher Ratios show null positive results, suggesting class size reductions were not a mechanism by which Innovation Schools improved performance. Teacher Salary shows a significant negative relationship, which is the best evidence of organizational mechanisms produced in this analysis. While teacher turnover rates or teacher experience were not available, if lower salaries resulted from turnover leading to replacement with less experienced teachers, these teacher salary effects could indicate large changes in the composition of the teaching staff, consistent with other turnover reforms. Event-study analysis by cohort suggests these salary declines may have been larger in early years of the reform (results available upon request), suggesting performance improvements were associated with early reforms which also faded out over time. Finally, discipline shows null negative results, indicating that at the very least, Innovation Schools did not improve performance through harsh discipline. These organizational analyses are not highly informative in understanding the success of the Innovation Schools. It appears that many causal mechanisms of the program, including the ones listed above, or mechanisms fitting broadly under the umbrella of managerial autonomy, are likely to be unobservable.ConclusionFor decades, efforts to improve performance in struggling urban schools have fallen short of hopes. Studies of high-quality charters have linked positive achievement effects to innovative managerial practices ADDIN EN.CITE <EndNote><Cite><Author>Dobbie</Author><Year>2011</Year><RecNum>6</RecNum><DisplayText>(Dobbie and Fryer 2011)</DisplayText><record><rec-number>6</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534183185">6</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Dobbie, Will</author><author>Fryer, Roland G.</author></authors></contributors><titles><title>Are High Quality Schools Enough to Increase Achievement Among the Poor? Evidence from the Harlem Children's Zone</title><secondary-title>American Economic Journal: Applied Economics</secondary-title></titles><periodical><full-title>American Economic Journal: Applied Economics</full-title></periodical><pages>158-187</pages><volume>3</volume><number>3</number><dates><year>2011</year></dates><urls></urls></record></Cite></EndNote>(Dobbie and Fryer 2011) and experimental analysis by ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014) find that transplanting managerial innovations from charter schools to traditional public schools can drive performance improvements. In this paper, I investigate a similar managerial reform in DPS. Using a difference-in-differences design, I identify large positive effects of the Innovation Schools reform on math, reading, and writing test scores. However, I also find that program impacts peaked in year two of implementation and then faded out dramatically afterward. This finding suggests turnaround reforms are capable of producing large performance improvements, but that results may be difficult to sustain without continued focus in later years of implementation.This study contributes a methodologically robust evaluation of the effects of the Innovation Schools program on traditional public schools that transition to Innovation status. I address methodological issues related to parallel trends assumptions, non-random selection into treatment, and inferential difficulties resulting from a small number of treated clusters. Through graphical analyses and placebo testing, I find that the intervention was mostly free of pre-treatment trend violations. Results are generally robust to multiple specification, including with propensity score matched comparison groups, event-study regressions, school-specific linear time trends and lagged dependent variable. By using variation from schools that transition to Innovation Status, I contribute evidence that this reform has the potential to turn around struggling urban schools, and that coupling managerial autonomy and innovation has potential as an effective turnaround strategy. However, this study also provides cautionary new evidence that turnaround impacts may be difficult to sustain, and can be lost in later years of implementation. This finding presents an important consideration for policymakers and school leaders implementing turnaround reforms and should be explored in other contexts.I find that the program resulted in improvements of 0.1 to 0.3 standard deviations on end of year standardized test scores, effects which are often above the 80th percentile (0.2 sd) of over 200 findings reviewed by ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Kraft</Author><Year>2020</Year><RecNum>193</RecNum><DisplayText>Kraft (2020)</DisplayText><record><rec-number>193</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1591362279">193</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kraft, Matthew A.</author></authors></contributors><titles><title>Interpreting Effect Sizes of Education Interventions</title><secondary-title>Educational Researcher</secondary-title></titles><periodical><full-title>Educational Researcher</full-title></periodical><pages>241-253</pages><volume>49</volume><number>4</number><dates><year>2020</year></dates><urls></urls></record></Cite></EndNote>Kraft (2020). Where significant, results are generally robust to rigorous hypothesis testing with wild bootstrap clustered standard errors. The calculation of treatment effects in standard deviations (sd) allows comparison to the results of ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014). While Fryer estimates effects of 0.172 (0.065) sd in math and 0.076 (0.052) sd in reading, I find significant effects of approximately 0.25 sd in math and 0.15 sd in reading, with marginally significant effects of approximately 0.1 sd in writing. These effects are larger than those of ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Fryer</Author><Year>2014</Year><RecNum>8</RecNum><DisplayText>Fryer (2014)</DisplayText><record><rec-number>8</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534186062">8</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Fryer, Roland G.</author></authors></contributors><titles><title>Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments</title><secondary-title>The Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>The Quarterly Journal of Economics</full-title></periodical><pages>1355-1407</pages><volume>129</volume><number>3</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>Fryer (2014), and achieve significance in reading, where he found null effects. However, evidence from both graphical plots and heterogeneity analyses suggests that the effects of the reform may fade out over time. Changes in teacher salary suggest large shifts in workforce composition in earlier years of the reform which may have also faded over time, suggesting an association between performance increases and organizational changes that were not sustained over time. While turnarounds may be able to rapidly increase performance in struggling schools, sustaining those improvements may require different approaches. Future research on school turnarounds should investigate whether this fade-out is a regular phenomenon in other contexts.The results contrast with those of ADDIN EN.CITE <EndNote><Cite AuthorYear="1"><Author>Abdulkadiroglu</Author><Year>2017</Year><RecNum>28</RecNum><DisplayText>Abdulkadiroglu et al. (2017)</DisplayText><record><rec-number>28</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534618148">28</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Abdulkadiroglu, Atila</author><author>Angrist, Joshua D.</author><author>Narita, Yusuke</author><author>Pathak, Parag A.</author></authors></contributors><titles><title>Research Design Meets Market Design: Using Centralized Assignment for Impact Evaluation</title><secondary-title>Econometrica</secondary-title></titles><periodical><full-title>Econometrica</full-title></periodical><pages>1373-1432</pages><volume>85</volume><number>5</number><dates><year>2017</year></dates><urls></urls></record></Cite></EndNote>Abdulkadiroglu et al. (2017) who found null impacts of Innovation Schools on student achievement. To reconcile these differences, it is important to remember the differences between the two results in terms of sample and identification strategy. Abdulkadiroglu and colleagues employ a sample including both Innovation Schools that transitioned from traditional public schools and a larger group that opened as Innovation Schools. The sample in this study includes only those Innovation Schools that transitioned to Innovation status, which allows estimation of a parameter indicating the efficacy of the program as a managerial intervention to turn around struggling public schools. Established public schools that initiate reforms under the program may use autonomy in different ways or benefit from prior experience leading to better results; brand new schools are likely to suffer from growing pains and experience higher risk of failure than those with an established track record. Furthermore, the identification strategy in this study leverages transition over time to Innovation status, so results can be interpreted as within-schools improvements from baseline performance. While some researchers may find the lottery randomization used by Abdulkadirolgu and colleagues more compelling, it estimates a different parameter, which should be interpreted as a comparison between the performance of two identical students during the same time period, one of whom attends an Innovation School and one who does not. If Innovation Schools have lower average performance than comparison schools holding all confounding factors constant, the Abdulkadiroglu method will find a negative impact on student achievement. However, my within-schools difference-in-differences model holds differences in mean performance constant, instead comparing schools on their rate of change from the pre-treatment period. It is possible that Innovation Schools could demonstrate considerable growth from their pre-treatment level over time, even if their ultimate performance mean remains lower than other district schools. Since turn around reforms intend to initiate performance growth in very low-performing schools, it is important to know that this growth is possible, even if it will not lead to parity with other district schools. These findings suggest the Innovation Schools reform can produce growth in school performance. However, policy makers should be aware that autonomy and innovation involves risk, and the effects of the intervention were heterogenous, driven by a group of successful schools whose progress masked a smaller group with negative impacts from possibly failed reforms. The potential of the program to lead to both successes and failures was born out in anecdotal experiences of Innovation Schools. For instance, Manual High School experienced notable problems when splitting the school into 3 smaller schools hosted in the same building, and its performance declined after transition to Innovation status. ADDIN EN.CITE <EndNote><Cite><Author>Robles</Author><Year>2011</Year><RecNum>126</RecNum><DisplayText>(Robles 2011)</DisplayText><record><rec-number>126</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1543277002">126</key></foreign-keys><ref-type name="Newspaper Article">23</ref-type><contributors><authors><author>Robles, Yesenia</author></authors></contributors><titles><title>Anxious Eyes on Denver's Far-Northeast Region as School Starts</title><secondary-title>The Denver Post</secondary-title></titles><dates><year>2011</year><pub-dates><date>August 10, 2011</date></pub-dates></dates><urls><related-urls><url>;(Robles 2011) The proposition of managerial autonomy-based interventions as a high-risk high-reward venture should be understood by policy-makers considering similar programs. The analysis also sheds light on the importance of external support, such as the private Blueprint Schools Network and the School Improvement Grant program. While these supports possibly contributed to the effects of the program, they do not explain them. Externally supported schools did about the same as other high-performing Innovation Schools. This finding makes a contribution to understandings of privatization in public education and lends new insight into the role of Charter Management Organization-like private actors in turnaround efforts in traditional public schools. The analysis in this paper is subject to limitations. Most notable is the small sample of treated schools. While the estimated effects are compelling and robust, they are only observed in a small sample of 13 treated schools. The extent to which the experience of 13 schools can be generalized nationwide should be considered with caution. Since similar models have been implemented in other cities, scholars should compare results from those programs to the results seen Denver. Further limitations stem from the limits of publicly available data. I was not able to control for demographic percentages of students with disabilities or limited English proficiency, and I was not able to assess how their populations changed in Innovation Schools. Since schools subject to performance pressures have been documented to manipulate enrollments of these populations, it would have been better to have these variables available ADDIN EN.CITE <EndNote><Cite><Author>Figlio</Author><Year>2011</Year><RecNum>79</RecNum><DisplayText>(Figlio and Loeb 2011)</DisplayText><record><rec-number>79</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1539890072">79</key></foreign-keys><ref-type name="Book Section">5</ref-type><contributors><authors><author>Figlio, David</author><author>Loeb, Susanna</author></authors><secondary-authors><author>Hanushek, Eric A.</author><author>Machin, Stephen</author><author>Woessman, Ludger</author></secondary-authors></contributors><titles><title>School Accountability</title><secondary-title>Handbook of the Economics of Education</secondary-title></titles><pages>383-421</pages><volume>3</volume><dates><year>2011</year></dates><pub-location>The Netherlands: North-Holland</pub-location><publisher>Elsevier</publisher><urls></urls></record></Cite></EndNote>(Figlio and Loeb 2011). However, I was able to test for changes in student composition based on poverty and black and Hispanic composition and found no evidence of sorting. Similarly, I am limited in only being able to use high-stakes tests as a measure of performance. Its well-documented that schools can sometimes respond to interventions by increasing performance on high-stakes tests, without demonstrating gains on more general measures of performance ADDIN EN.CITE <EndNote><Cite><Author>Figlio</Author><Year>2011</Year><RecNum>79</RecNum><DisplayText>(Figlio and Loeb 2011)</DisplayText><record><rec-number>79</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1539890072">79</key></foreign-keys><ref-type name="Book Section">5</ref-type><contributors><authors><author>Figlio, David</author><author>Loeb, Susanna</author></authors><secondary-authors><author>Hanushek, Eric A.</author><author>Machin, Stephen</author><author>Woessman, Ludger</author></secondary-authors></contributors><titles><title>School Accountability</title><secondary-title>Handbook of the Economics of Education</secondary-title></titles><pages>383-421</pages><volume>3</volume><dates><year>2011</year></dates><pub-location>The Netherlands: North-Holland</pub-location><publisher>Elsevier</publisher><urls></urls></record></Cite></EndNote>(Figlio and Loeb 2011). Future analysis should investigate whether turnaround reforms lead to general performance improvements, or only improvement on high-stakes assessments.This study contributes to the literature on the impacts of managerial autonomy in public schools. Results suggest a reform involving decentralization of school management from districts to schools produced large performance gains as a turnaround strategy. The relationship between managerial autonomy and public school performance is understudied, with the finding of positive effects in British public schools ADDIN EN.CITE <EndNote><Cite><Author>Clark</Author><Year>2009</Year><RecNum>90</RecNum><DisplayText>(Clark 2009)</DisplayText><record><rec-number>90</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1540311431">90</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Clark, Damon</author></authors></contributors><titles><title>The Performance and Competitive Effects of School Autonomy</title><secondary-title>Journal of Political Economy</secondary-title></titles><periodical><full-title>Journal of Political Economy</full-title></periodical><pages>745-783</pages><volume>117</volume><number>4</number><dates><year>2009</year></dates><urls></urls></record></Cite></EndNote>(Clark 2009) and two findings of null impacts in the American context ADDIN EN.CITE <EndNote><Cite><Author>Steinberg</Author><Year>2014</Year><RecNum>115</RecNum><DisplayText>(Steinberg 2014, Abdulkadiroglu et al. 2011)</DisplayText><record><rec-number>115</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1541978356">115</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Steinberg, Matthew P.</author></authors></contributors><titles><title>Does Greater Autonomy Improve School Performance? Evidence from a Regression Discontinuity Analysis in Chicago</title><secondary-title>Education Finance and Policy</secondary-title></titles><periodical><full-title>Education Finance and Policy</full-title></periodical><pages>1-35</pages><volume>9</volume><number>1</number><dates><year>2014</year></dates><urls></urls></record></Cite><Cite><Author>Abdulkadiroglu</Author><Year>2011</Year><RecNum>24</RecNum><record><rec-number>24</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1534446649">24</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Abdulkadiroglu, Atila</author><author>Angrist, Joshua D.</author><author>Dynarski, Susan M.</author><author>Kane, Thomas J.</author></authors></contributors><titles><title>Accountability and Flexibility in Public Schools: Evidence from Boston's Charters and Pilots</title><secondary-title>Quarterly Journal of Economics</secondary-title></titles><periodical><full-title>Quarterly Journal of Economics</full-title></periodical><pages>699-748</pages><volume>126</volume><number>2</number><dates><year>2011</year></dates><urls></urls></record></Cite></EndNote>(Steinberg 2014, Abdulkadiroglu et al. 2011) being the most notable results. This study contributes to what will hopefully be a growing body of literature on this question.This paper also contributes to literatures on school choice, demonstrating that charter-like managerial interventions can improve student outcomes in struggling urban schools. As studies now consistently find that charter schools can have positive effects on student achievement in multiple contexts PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BYmR1bGthZGlyb2dsdTwvQXV0aG9yPjxZZWFyPjIwMTc8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==
ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BYmR1bGthZGlyb2dsdTwvQXV0aG9yPjxZZWFyPjIwMTc8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==
ADDIN EN.CITE.DATA (Abdulkadiroglu et al. 2017, Angrist et al. 2016, Dobbie and Fryer 2011, Booker et al. 2011), successful programs should be analyzed for lessons that can improve traditional public schools. Lessons on curriculum, staffing and organizational management can inform efforts to improve performance in struggling public schools. In the case of Innovation Schools in DPS, implementing charter-like management practices such as increased instructional time, student-based learning and customized evaluation schemes can lead to performance improvements. These are generalizable lessons that can be exported to other contexts and brought to scale. While Innovation Schools in the DSSN raised $4.2 million in external fundraising, fundraising may be difficult to sustain over time or may be more difficult in communities with limited resources, as has been experienced by resource-intensive charters such as KIPP as they have expanded operations ADDIN EN.CITE <EndNote><Cite><Author>Abrams</Author><Year>2016</Year><RecNum>191</RecNum><DisplayText>(Abrams 2016)</DisplayText><record><rec-number>191</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1591300967">191</key></foreign-keys><ref-type name="Book">6</ref-type><contributors><authors><author>Abrams, Samuel E.</author></authors></contributors><titles><title>Education and the Commercial Mindset</title></titles><dates><year>2016</year></dates><publisher>Harvard University Press</publisher><urls></urls></record></Cite></EndNote>(Abrams 2016). Furthermore, many of the reforms in Denver would have been impossible without the ability to waive teacher collective bargaining rights, which would be potentially prohibitive in many institutional contexts. Finally, the study contributes to a growing literature on turnaround reforms PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5TY2h1ZWxlcjwvQXV0aG9yPjxZZWFyPjIwMTc8L1llYXI+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=
ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5TY2h1ZWxlcjwvQXV0aG9yPjxZZWFyPjIwMTc8L1llYXI+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=
ADDIN EN.CITE.DATA (Schueler, Goodman, and Deming 2017, Carlson and Lavertu 2018, Heissel and Ladd 2018). Identifying reforms that can improve performance in struggling urban schools has been an elusive goal in education policy. While there are mixed results, a number of studies now demonstrate that it is possible to rapidly improve public school performance by leveraging comprehensive managerial interventions. This study is limited in its ability to explore mechanisms, and thus cannot provide statistical evidence to distinguish results from other contexts where reforms did not lead to performance improvements. However, the study shows that turnaround is possible in a new context, providing detailed exposition of an autonomy-based model that has demonstrated some efficacy. While effects are heterogenous and fade out over time, and questions remain about the possibility of scaling up from a limited reform, turnarounds may be possible in other settings, especially if further studies can identify determinants of successful reforms. References ADDIN EN.REFLIST Abdulkadiroglu, Atila, Joshua D. Angrist, Susan M. Dynarski, and Thomas J. Kane. 2011. "Accountability and Flexibility in Public Schools: Evidence from Boston's Charters and Pilots." Quarterly Journal of Economics 126 (2):699-748.Abdulkadiroglu, Atila, Joshua D. Angrist, Yusuke Narita, and Parag A. Pathak. 2017. "Research Design Meets Market Design: Using Centralized Assignment for Impact Evaluation." Econometrica 85 (5):1373-1432.Abrams, Samuel E. 2016. Education and the Commercial Mindset: Harvard University Press.Angrist, Joshua D., Sarah R. Cohodes, Susan M. Dynarski, Parag A. Pathak, and Christopher R. Walters. 2016. "Stand and Deliver: Effects of Boston's Charter Schools on College Preparation, Entry and Choice." Journal of Labor Economics 34 (2):275-318.Bertrand, Marianne, Esther Duflo, and Sendhil Mullainathan. 2004. "How Much Should We Trust Differences-In-Differences Estimates?" The Quarterly Journal of Economics 119 (1):249-275.Bifulco, Robert, and Helen F. Ladd. 2006. "The Impacts of Charter Schools on Student Achievement: Evidence from North Carolina." Education Finance and Policy 1 (1):50-90.Booker, Kevin, Scott M. Gilpatric, and Timothy Gronberg. 2007. "The Impact of Charter School attendance on Student Performance." Journal of Public Economics 91 (5-6):849-876.Booker, Kevin, Tim R. Sass, Brian Gill, and Ron Zimmer. 2011. "The Effects of Charter High Schools on Educational Attainment." Journal of Labor Economics 29 (2):377-415.Cameron, A. Colin, Jonah B. Gelbach, and Douglas L. Miller. 2008. "Bootstrap-Based Improvements for Inference with Clustered Errors." Review of Economics and Statistics 90 (3):414-427.Carlson, Deven E., and Stephane Lavertu. 2018. "School Improvement Grants in Ohio: Effects on Student Achievement and School Administration." Educational Evaluation and Policy Analysis 40 (3):287-315.Chiang, Hanley. 2009. "How Accountability Pressure on Failing Schools Affects Student Achievement." Journal of Public Economics 93 (9-10):1045-1057.Chubb, John E., and Terry M. Moe. 1990. Politics, Markets and America's Schools. Washington, DC.: Brookings Institution Press.Clark, Damon. 2009. "The Performance and Competitive Effects of School Autonomy." Journal of Political Economy 117 (4):745-783.Colorado Department of Education. 2010. "Annual Report Concerning Districts of Innovation." Department of Education. 2013. "Colorado Innovation School Act." Department of Education. 2015. 2015 Innovation Report. Colorado Department of Education.Colorado Department of Education. n.d. "SchoolView." accessed June 22, 2019. , Thomas S. 2012. "Title." NBER Working Papers.Dee, Thomas S., and Brian A. Jacob. 2011. "The Impact of No Child Left Behind on Student Achievement." Journal of Policy Analysis and Management 30 (3):418-446.Dobbie, Will, and Roland G. Fryer. 2011. "Are High Quality Schools Enough to Increase Achievement Among the Poor? Evidence from the Harlem Children's Zone." American Economic Journal: Applied Economics 3 (3):158-187.Dragoset, Lisa, Jaime Thomas, Mariesa Herrmann, John Deke, Susanne James-Burdumy, Cheryl Graczewski, Andrea Boyle, Rachel Upton, Courtney Tanenbaum, and Jessica Giffin. 2017. School Improvement Grants: Implementation and Effectiveness. U.S. Department of Education.Figlio, David, and Susanna Loeb. 2011. "School Accountability." In Handbook of the Economics of Education, edited by Eric A. Hanushek, Stephen Machin and Ludger Woessman, 383-421. The Netherlands: North-Holland: Elsevier.Fryer, Roland G. 2014. "Injecting Charter School Best Practices into Traditional Public Schools: Evidence from Field Experiments." The Quarterly Journal of Economics 129 (3):1355-1407.Hanushek, Eric. 1997. "Assessing the Effects of School Resources on Student Performance: An Update." Educational Evaluation and Policy Analysis 19 (2):141-164.Hanushek, Eric, and Margaret E. Raymond. 2005. "Does School Accountability Lead to Improved Student Performance." Journal of Policy Analysis and Management 24 (2):297-327.Harris, Douglas N., and Matthew Larsen. 2016. The Effects of New Orleans Post-Katrina School Reforms on Student Academic Outcomes. Education Research Alliance for New Orleans.Heissel, Jennifer A., and Helen F. Ladd. 2018. "School Turnaround in North Carolina: A Regression Discontinuity Analysis." Economics of Education Review 62 (302-320).Imberman, Scott A. 2011. "The Effect of Charter Schools on Achievement and Behavior of Public School Students." Journal of Public Economics 95 (7-8):850-863.Kraft, Matthew A. 2020. "Interpreting Effect Sizes of Education Interventions." Educational Researcher 49 (4):241-253.LaFortune, Julien, Jesse Rothstein, and Diane Whitmore Schanzenbach. 2018. "School Finance Reform and the Distribution of Student Achievement." American Economic Journal: Applied Economics 10 (2):1-26.Mackinnon, James G. 2016. "Wild Bootstrap Inference for Wildly Different Cluster Sizes." Journal of Applied Econometrics 32 (2).Mackinnon, James G., and Matthew D. Webb. 2018. "The Wild Bootstrap for Few (Treated) Clusters." Econometrics Journal 21:114-135.Marsh, Julie A., Katherine o. Strunk, and Susan Bush. 2013. "Portfolio District Reform Meets School Turnaround: Early Implementation Findings from the Los Angeles Public School Choice Initiative." Journal of Educational Administration 51 (4):498-527.Mora, Ricardo, and Iliana Reggio. 2017. "alternative Diff-in-Diffs estimators with Several Pre-Treatment Periods." Econometrics Reviews (forthcoming).Robles, Yesenia. 2011. "Anxious Eyes on Denver's Far-Northeast Region as School Starts." The Denver Post, August 10, 2011. , Jonah, and Lesley J. Turner. 2010. "Short-Run Impacts of Accountability on School Quality." American Economic Journal: Economics Policy 2 (4):119-47.Schueler, Beth E., Joshua S. Goodman, and David J. Deming. 2017. "Can States Take Over and Turn Around School Districts? Evidence from Lawrence, Massachusetts." Educational Evaluation and Policy Analysis 39 (2):311-332.Steinberg, Matthew P. 2014. "Does Greater Autonomy Improve School Performance? Evidence from a Regression Discontinuity Analysis in Chicago." Education Finance and Policy 9 (1):1-35.Strunk, Katherine O., Julie A. Marsh, Ayesha K. Hashim, Susan Bush-Mecenas, and Tracey Weinstein. 2016. "The Impact of Turnaround Reform on Student Outcomes: Evidence and Insights from the Los Angeles Unified School District." Education Finance and Policy 11 (3):251-282.Sun, Min, Emily K. Penner, and Susanna Loeb. 2017. "Resource- and Approach-Driven Multidimensional Change: Three-Year Effects of School Improvement Grants." American Educational Research Journal 54 (4):607-643.Tomassini, Jason. 2012. "Denver Turnaround Initiative Showing Achievement Gains." Education Week, June 5, 2012. , John. 2004. Helping Children Left Behind: State Aid and the Pursuit of Educational Equity. Cambridge, MA: MIT University Press.Zimmer, Ron, Gary T. Henry, and Adam Kho. 2017. "The Effects of School Turnaround in Tennessee's Achievement School District and Innovation Zones." Educational Evaluation and Policy Analysis 39 (4):670-696.Tables:Table 1: Waivers Requested by DPS Innovation Schools ADDIN EN.CITE <EndNote><Cite><Author>Colorado Department of Education</Author><Year>2013</Year><RecNum>168</RecNum><DisplayText>(Colorado Department of Education 2013)</DisplayText><record><rec-number>168</rec-number><foreign-keys><key app="EN" db-id="ed5avrxplsttfierxf1pxp0v299p9vtvdeet" timestamp="1559848290">168</key></foreign-keys><ref-type name="Web Page">12</ref-type><contributors><authors><author>Colorado Department of Education,</author></authors></contributors><titles><title>Colorado Innovation School Act</title></titles><dates><year>2013</year></dates><urls><related-urls><url>;(Colorado Department of Education 2013)Statutory Provision WaivedPercentage of SchoolsMoney?Budgetary Control100CurriculumEducational Program and Textbook76Content Standards80Personnel?Employee Performance Evaluations92Personnel Selection and Pay100Employee Dress Code80Principal Training92Termination of Personnel96Teacher Licensing88Teacher Contracts92Teacher Probation92Teacher Transfers 92Grounds for Dismissal96Procedure for Dismissal88Teacher Salaries92Time?Longer School Day and School Year100Table 2: Descriptives???VariableObsMeanStd. Dev.MinMax??Writing 1,015491.5740.33419.88624.56Writing Z1,0150.001.00-1.723.21Reading1,015583.8547.01493.58704.82Reading Z1,0150.001.00-2.082.57Math1,015490.1152.45390.50642.37Math Z1,0150.001.00-1.952.91??Innovation Schools1,0150.040.200.001.00??Enrollment1,015532.88314.1242.002435.00Student Teacher Ratio1,01517.458.573.37264.00Teacher Salary90152947.737471.8931093.5282848.02Discipline90447.9468.470.00494.00% Free Lunch1,0150.710.270.031.00% Black1,0150.140.140.000.86% Hisp1,0150.580.280.040.98Table 3: Comparison Between Innovation and Non-Innovation Schools in 3-year Pre-Treatment Period (2006-07 to 2008-09)???Non-Innovation Schools (NIS)(105 Schools)PSM Matches (PSM)(44 Schools)Innovation Schools (IS) (12 Schools)p-value (NIS vs. IS)p-value (PSM vs. IS)N30813031??Writing486.67 (40.53)462.28 (28.74)466.11 (30.65)0.0060.51Writing Z0.05 (1.01)-0.56 (0.71)-0.47 (0.76)0.0060.54Reading579.63 (49.03)549.89 (38.38)550.92 (37.30)0.0020.89Reading Z0.05 (1.00)-0.55 (0.78)-0.54 (0.76)0.0020.91Math483.80 (50.32)451.97 (37.73)453.20 (40.71)0.0010.87Math Z0.06 (1.00)-0.58 (0.75)-0.55 (0.80)0.0010.89??Enrollment534.10 (320.83)533.59 (329.19)514.00 (238.51)0.730.76Student Teacher Ratio18.46 (14.50)17.28 (3.99)16.80 (2.53)0.520.53Teacher Salary (Thousands)54.77 (69.39)54.72 (51.60)53.11 (46.88)0.280.19Discipline55.93 (84.35)59.17 (86.17)77.18 (100.84)0.270.40% Free Lunch0.65 (0.27)0.81 (0.16)0.78 (0.09)0.0060.44% Black0.15 (0.15)0.15 (0.18)0.21 (0.19)0.0300.082% Hispanic0.55 (0.30)0.69 (0.23)0.63 (0.21)0.140.21Table 4: Effects of Innovation Schools on Academic Outcomes??(1)(2)(3)(4)(5)(6)VARIABLESWritingWritingReadingReadingMathMath???????Innovation Schools0.111+0.1080.136*0.113+0.232*0.225*?(0.0669)(0.0660)(0.0649)(0.0598)(0.0969)(0.0970)?0.1460.01350.062+0.095+0.043*0.041*??Full Samplexxx?Matched SamplexxxSchool FExxxxxxYear FExxxxxxObservations1,0154641,0154641,015464R-squared0.0830.0900.0860.1340.0920.084Number of Panels148621486214862Cluster robust standard errors in parentheses?Wild bootstrap p-values in third row?** p<0.01, * p<0.05, + p<0.1????Table 5: Event-study Regressions (Year Negative One Omitted)?(1)(2)(3)(4)(5)(6)VARIABLESWritingWritingReadingReadingMathMath??Year Negative Seven0.03620.05310.09790.188*-0.0171-0.0250?(0.0583)(0.0724)(0.0758)(0.0842)(0.0842)(0.0996)Year Negative Six0.190*0.214*0.08170.152-0.0878-0.0760?(0.0908)(0.0989)(0.140)(0.132)(0.142)(0.143)Year Negative Five0.03180.0269-0.0329-0.0185-0.191-0.163?(0.0883)(0.0873)(0.136)(0.136)(0.118)(0.123)Year Negative Four0.02610.0254-0.01410.009300.02580.0370?(0.0673)(0.0715)(0.0857)(0.0815)(0.0862)(0.0899)Year Negative Three-0.0648-0.0697-0.01210.00958-0.0155-0.0127?(0.0794)(0.0836)(0.0953)(0.0980)(0.106)(0.113)Year Negative Two-0.00993-0.0169-0.128+-0.121-0.01170.000351?(0.0441)(0.0450)(0.0759)(0.0736)(0.0903)(0.0935)Year Negative One(Omitted)(Omitted)(Omitted)(Omitted)(Omitted)(Omitted)??Year One0.03110.02560.03330.02710.1290.135?(0.0501)(0.0528)(0.0551)(0.0563)(0.0961)(0.0971)Year Two0.218**0.215**0.172*0.164*0.310*0.304*?(0.0695)(0.0703)(0.0767)(0.0741)(0.120)(0.123)Year Three0.1280.1270.1110.09950.284+0.273+?(0.125)(0.126)(0.141)(0.138)(0.155)(0.159)Year Four0.05520.03420.1280.1070.1750.189?(0.138)(0.140)(0.155)(0.159)(0.187)(0.195)Year Five-0.0537-0.07180.04210.0122-0.0135-0.00280?(0.143)(0.147)(0.175)(0.195)(0.211)(0.221)??Full Samplexxx?Matched SamplexxxSchool FExxxxxxYear FExxxxxxObservations1,0154641,0154641,015464R-squared0.0940.1190.0940.1540.1030.106Number of Panels148621486214862Cluster robust standard errors in parentheses?** p<0.01, * p<0.05, + p<0.1????Table 6: Heterogenous Effects of Innovation Schools on Academic Outcomes (By Effect Size and Direction)?Full SampleMatched SampleFull Sample Matched Sample??Writing0.228**0.224**-0.136-0.152+?(0.0653)(0.0658)(0.0836)(0.0801)Reading0.254**0.222**-0.116-0.132?(0.0614)(0.0551)(0.0991)(0.101)Math0.419**0.408**-0.160+-0.168+?(0.0919)(0.0894)(0.0947)(0.0927)High Performers (8 Schools)xx?Low Performers (5 Schools)xxSchool FExxxxYear FExxxxObservations980429955404Number of Panels1435714054Cluster robust standard errors in parentheses** p<0.01, * p<0.05, + p<0.1Table 7: Heterogenous Effects of Innovation Schools on Academic Outcomes (By External Management and SIG Funding)?Full SampleMatched SampleFull Sample Matched Sample??Writing0.197*0.194*0.216*0.211*?(0.0827)(0.0816)(0.0939)(0.0940)Reading0.255**0.205**0.297**0.239**?(0.0818)(0.0712)(0.0845)(0.0754)Math0.302**0.297**0.322**0.321**?(0.0683)(0.0749)(0.0757)(0.0833)??Without DSSNxx?Without DSSN & SIGxxSchool FExxxxYear FExxxxObservations964413958407Number of Panels1415514054Cluster robust standard errors in parentheses** p<0.01, * p<0.05, + p<0.1Table 8: Effects of Innovation Schools on Organizational Outcomes?Full SampleMatched Sample??Enrollment17.589.894?(16.76)(18.06)Student-Teacher Ratio1.6320.371?(1.339)(0.515)Teacher Salary-2,475*-2,477*?(1,117)(1,057)Discipline-8.961-8.606?(10.94)(11.42)??All Innovation SchoolsxxSchool FExxYear FExxObservations904412Number of Panels14862Cluster robust standard errors in parentheses** p<0.01, * p<0.05, + p<0.1Figure 2: Trend Graphs (Writing Scores, Main Sample) Note. These figures are binned scatter plots. This graph plots average Writing Score for treatment and control schools from 2007 and 2014.Figure 3: Trend Graphs (Reading Scores, Main Sample) Note. These figures are binned scatter plots. This graph plots average Reading score for treatment and control schools from 2007 and 2014.Figure 4: Trend Graphs (Math Scores, Main Sample) Note. These figures are binned scatter plots. This graph plots average Math score for treatment and control schools from 2007 and 2014.Appendix Tables:Appendix Table 1: Student Sorting Models (Treatment Effects on Demographic Controls)?(1)(2)(3)(4)(5)(6)VARIABLES% Free Lunch% Free Lunch% Black% Black% Hisp% Hisp???????Innovation Schools0.00574-0.00575-0.0118-0.0154-0.0126-0.0180?(0.0202)(0.0217)(0.0105)(0.00980)(0.0103)(0.0132)??Full Samplexxx?Matched SamplexxxSchool FExxxxxxYear FExxxxxxObservations1,0154641,0154641,015464R-squared0.4030.4320.4090.5660.4200.534Number of Panels148621486214862Cluster robust standard errors in parentheses?** p<0.01, * p<0.05, + p<0.1????Appendix Table 2: Placebo Tests???(1)(2)(3)(4)(5)(6)VARIABLESWritingWritingReadingReadingMathMath???????Placebo 2008-0.00010.02480.1090.07220.04170.0504?(0.0502)(0.0553)(0.0682)(0.0856)(0.0617)(0.0686)Placebo 20090.03830.008400.0827+0.02120.09240.103?(0.0500)(0.0599)(0.0428)(0.0588)(0.0679)(0.0738)Placebo 2010-0.00397-0.02120.0540-0.01730.06780.0825?(0.0614)(0.0719)(0.0925)(0.0986)(0.0845)(0.0993)Placebo 20110.01170.02000.1200.07230.162+0.175+?(0.0807)(0.0863)(0.0848)(0.0944)(0.0869)(0.103)Placebo 2012-0.0906-0.08420.00131-0.04120.09870.118?(0.0590)(0.0679)(0.0659)(0.0752)(0.0740)(0.0814)Placebo 2013-0.285**-0.291**-0.361**-0.432**-0.240**-0.264**?(0.0490)(0.0661)(0.0525)(0.0758)(0.0608)(0.0882)Full Samplexxx?Matched SamplexxxSchool FExxxxxxYear FExxxxxxObservations837371837371837371R-squared0.0750.0870.0760.1160.0890.085Number of Panels143621436214362Cluster robust standard errors in parentheses???** p<0.01, * p<0.05, + p<0.1????Appendix Table 3: Event-study Regressions (Year Negative Three Omitted)?(1)(2)(3)(4)(5)(6)VARIABLESWritingWritingReadingReadingMathMath??Year Negative Seven0.101+0.123*0.110*0.179**-0.00159-0.0123?(0.0522)(0.0598)(0.0470)(0.0559)(0.0597)(0.0702)Year Negative Six0.255**0.283**0.09380.142-0.0723-0.0633?(0.0760)(0.0825)(0.106)(0.0980)(0.108)(0.110)Year Negative Five0.09660.0967+-0.0208-0.0280-0.175**-0.150**?(0.0589)(0.0565)(0.0782)(0.0816)(0.0489)(0.0529)Year Negative Four0.09100.0951-0.00196-0.0002730.04130.0496?(0.0602)(0.0636)(0.0550)(0.0522)(0.0590)(0.0588)Year Negative Three(Omitted)(Omitted)(Omitted)(Omitted)(Omitted)(Omitted)??Year Negative Two0.05490.0528-0.116+-0.131+0.003770.0130?(0.0571)(0.0583)(0.0688)(0.0769)(0.0773)(0.0778)Year Negative One0.06480.06970.0121-0.009580.01550.0127?(0.0794)(0.0836)(0.0953)(0.0980)(0.106)(0.113)Year One0.09590.09540.04540.01760.1440.148?(0.0888)(0.0965)(0.0906)(0.0945)(0.128)(0.133)Year Two0.283**0.285**0.184*0.154*0.326*0.316*?(0.0875)(0.0920)(0.0829)(0.0750)(0.130)(0.133)Year Three0.193+0.197+0.1230.08990.299*0.286*?(0.115)(0.117)(0.131)(0.127)(0.120)(0.128)Year Four0.1200.1040.1400.09700.1910.201?(0.124)(0.124)(0.142)(0.134)(0.167)(0.177)Year Five0.0112-0.002060.05420.002660.002010.00986?(0.137)(0.137)(0.170)(0.172)(0.191)(0.201)??Full Samplexxx?Matched SamplexxxSchool FExxxxxxYear FExxxxxxObservations1,0154641,0154641,015464R-squared0.0940.1190.0940.1540.1030.106Number of Panels148621486214862Cluster robust standard errors in parentheses?** p<0.01, * p<0.05, + p<0.1????Appendix Table 4: School-Specific Linear Time Trends (Main Models)?(1)(2)(3)(4)(5)(6)VARIABLESWritingWritingReadingReadingMathMath???????Innovation School0.135+0.122+0.1240.1190.206+0.194+?(0.0739)(0.0718)(0.0846)(0.0768)(0.117)(0.115)??Full Samplexxx?Matched SamplexxxSchool FExxxxxxYear FExxxxxxSchool-Specific TrendsxxxxxxObservations1,0154641,0154641,015464R-squared0.4690.4400.4820.4700.5310.495Number of Panels148621486214862Cluster robust standard errors in parentheses?** p<0.01, * p<0.05, + p<0.1????Appendix Table 5: Lagged Dependent Variable Models (Main Models)?(1)(2)(3)(4)(5)(6)VARIABLESWritingWritingReadingReadingMathMath???????Innovation School0.08840.08670.07110.07670.177*0.179*?(0.0538)(0.0550)(0.0498)(0.0504)(0.0694)(0.0728)??Full Samplexxx?Matched SamplexxxSchool FExxxxxxYear FExxxxxxSchool-Specific TrendsxxxxxxObservations865402865402865402R-squared0.2050.1860.2190.1930.2900.237Number of Panels142621426214262Cluster robust standard errors in parentheses?** p<0.01, * p<0.05, + p<0.1???? ................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- wordpress passing data between pages
- wordpress business templates
- wordpress rss feed not working
- wordpress jquery is not defined
- create wordpress blog
- wordpress roles editor
- wordpress full rss feed
- wordpress rss feed settings
- wordpress rss feed plugin
- wordpress display rss feed
- wordpress rss feed link
- wordpress rss feed to post