Exam #2 Design Variables and their Attributes



Quiz #4 Hw#1

1. The researcher ran an initial study which examined the effectiveness of a new program for teaching remedial mathematics to junior high school students. That study showed a linear trend across different amounts of training with the program (2, 4, 6, & 8 weeks). Excited by this finding, the researcher decided to explore whether the program would show similar results with high school and post-high school students. Volunteers from the three AGE GROUPs (jr. high, high, post-high) who has scored between 20% and 30% on a "remedial math assessment test" were randomly assigned to complete either 2, 4, 6 or 8 weeks of training (AMOUNT OR TRAINING) before re-taking the assessment test.

prior math training (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math training (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math training (with respect to the interaction )

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

2. The purpose of the study was to examine the possible joint effects of type of practice and the amount of practice upon putting performance. Each of the volunteer participants (who had no previous experience putting, not even carpet golf -- e.g., putt-putt golf) was randomly assigned to which type of practice they would receive, either "physical" or "imaginal" practice. All participants watched a 2-minute video of a professional golfer showing how to line-up and make a putt. Those in the "physical practice" condition were given an opportunity to practice putting and told to follow the procedure shown on the video. Those in the "imaginal practice" condition were told to sit quietly and imagine going through the procedure shown in the video, focusing upon both "seeing" and "feeling" each step in the procedure. Participants were permitted to complete as many practices as they wanted (and were told to keep count). When each participant felt they were well-prepared, they took 20 12-foot putts (real ball, real putter, real hole), and the DV was the number of putts that went into the hole. When preparing for data analysis, the participants were divided into those who made 0-10 practices, those who made 11 or more practices.

Number of practices taken by each participant (with respect to the main effect of amount of practice)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Number of practices taken by each participant (with respect to the main effect of type of practice)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Number of practices taken by each participant (with respect to the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Previous experience putting (with respect to main effect of practice type)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Previous experience putting (with respect to main effect of practice amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Previous experience putting (with respect to the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Putting instruction (with respect to the main effect of practice type)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Putting instruction (with respect to the main effect of practice amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Putting instruction (with respect to the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Motivation to perform well (with respect to main effect of practice type)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Motivation to perform well (with respect to main effect of practice amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Motivation to perform well (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

3. The researcher ran an initial study that showed a linear trend across different amounts of training with the program (2, 4, 6, & 8 weeks). Excited by this finding, the researcher decided to explore whether the program would show similar results with high school and post-high school students. Volunteers from the three age groups (jr. high, high, post-high) who has scored between 20% and 30% on a "remedial math assessment test" were randomly assigned to complete either 2, 4, 6 or 8 weeks of training before re-taking the assessment test. Because volunteers were used, the researcher was forced to make the homework assigned at each class to be “optional”. He did ask students to record how many problems from each homework they completed, and used this information to form groups of three students, one from each age group, who had completed the same average number of assignments.

prior math training (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math training (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math training (with respect to interaction )

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of training amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

amount of homework completed (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

amount of homework completed (with respect to main effect of type of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

amount of homework completed (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

4. Here’s how the study was actually run… The researcher ran an initial study that showed a linear trend across different amounts of training with the program (2, 4, 6, & 8 weeks). Excited by this finding, the researcher decided to explore whether the program would show similar results with high school and post-high school students. Volunteers from the three age groups (jr. high, high, post-high) who has scored between 20% and 30% on a "remedial math assessment test" were asked to complete a questionnaire that measured their motivation to succeed and got a vocabulary-based assessment of their IQ (this instrument is the “industry standard” for this type of research). Because volunteers were used, the researcher was forced to allow each participant to choose whether they would be in the 2, 4, 6 or 8-week treatment and to make the homework assigned at each class to be “optional”. Participants were asked to record how many problems from each homework they completed. Sets of 12 participants (one each who were each combination of age group and amounts of training) were identified who had equivalent IQ scores. Attempts to further equate these sets of 12 individuals on motivation was impossible, but we were able to identify groups of 3 (one from each age group) who were equivalent on motivation and who reported doing the same number of homework assignments.

IQ (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

IQ (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

IQ (with respect to interaction )

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

motivation (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

amount of homework completed (with respect to main effect of type of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

amount of homework completed (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of training amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

prior math skill (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Answers

Story #1.

prior math training (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

participants were randomly assigned to which amount of training condition they did, so prior math training (and all other subject variables) were ( randomized ( balanced ( controlled

prior math training (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

participants were not randomly assigned to which age group they were in; prior math training was not a matching variable, and it is not reasonably a constant so ( confound ( initial equivalence problem

prior math training (with respect to the interaction )

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any variable that is a confound for either of the main effects is a confound of the interaction effect

prior math skill & motivation ( will have exactly the same set of descriptors for each term, because they too are subject variables

Story #2

Number of practices taken by each participant (with respect to the main effect of amount of practice)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this is a manipulated/procedural variable because it could/should have been under the control of the researcher it is one of the IVs

Number of practices taken by each participant (with respect to the main effect of type of practice)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this procedural variable is one of the IVs. each effect is controlled for the other effects, because of balancing (1/2 those in each condition are split between the conditions of the other variable) ( control ( balancing

Number of practices taken by each participant (with respect to the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

again, each effect is controlled for the other effects ( control ( balancing

Previous experience putting (with respect to main effect of practice type)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

since no one had any practice, this variable is a constant with a value of “0”, and so, practice was ( eliminated -> control

Previous experience putting (with respect to main effect of practice amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

again, since no one had any practice, this variable is a constant with a value of “0”, and so, practice was ( eliminated ( control

Previous experience putting (with respect to the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any variable that is a control for both main effects will be a control for the interaction effect ( eliminated ( control

Putting instruction (with respect to the main effect of practice type)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

Everybody in every condition got the same 2-minute instructional video ( held constant ( control

A constant for everybody is a control for all effects – both main effects and the interaction

Motivation to perform well (with respect to main effect of practice type)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

participants were randomly assigned to which type of training they received, so motivation (and all other subject variables) were ( randomized ( balanced ( controlled

Motivation to perform well (with respect to main effect of practice amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

participants were not randomly assigned to which amount of practice group they were in; motivation was not a matching variable, and it is not reasonably a constant so ( confound ( initial equivalence problem

Motivation to perform well (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any variable that is a confound for either of the main effects is a confound of the interaction effect

Story #3

prior math training (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

participants were randomly assigned to which amount of training condition they did, so prior math training (and all other subject variables) were ( randomized ( balanced ( controlled

prior math training (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

participants were not randomly assigned to which age group they were in; prior math training was not a matching variable (that age groups were matched on #homeworks doesn’t help with the control of this subject variable), and it is not reasonably a constant so ( confound ( initial equivalence problem

prior math training (with respect to the interaction )

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any variable that is a confound for either of the main effects is a confound of the interaction effect

prior math skill & motivation ( will have exactly the same set of descriptors for each term, because they too are subject variables ( that there was post hoc matching on amount of homework doesn’t provide control for these variables!

amount of homework completed (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this is a procedural variable – while the amount of homework each participant completed was determined by that participant (rather than held constant, etc.), this was a variable that “occurred” during the procedure and should have been controlled/manipulated by the researcher (but wasn’t). We can not reasonably assume that people in the different amount of training conditions did the same amount of homework so, manipulated/procedural variable ( confound ( ongoing equivalence problem

amount of homework completed (with respect to main effect of type of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this procedural variable was the basis for forming post hoc equivalent groups (since this is a procedural variable it may be called “yoking” instead of matching, but …), so ( manipulated/procedural variable ( matched/yoked ( balanced ( control

amount of homework completed (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any procedural variable that is a confound for either of the main effects is a confound of the interaction effect

Story #4

IQ (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

IQ was a post-hoc matching variable – since groups of 12 (1 from each design condition) were matched, we get, subject variable ( matched ( balanced ( controlled for all effects in this design

IQ (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

IQ was a post-hoc matching variable – since groups of 12 (1 from each design condition) were matched, we get, subject variable ( matched ( balanced ( controlled for all effects in this design

IQ (with respect to interaction )

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

IQ was a post-hoc matching variable – since groups of 12 (1 from each design condition) were matched, we get, subject variable ( matched ( balanced ( controlled for all effects in this design

motivation (with respect to main effect of amount of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this subject variable was used as a post hoc matching variable, but the matches were not made across the conditions of this IV, so it is subject variable ( confound ( initial equivalence problem

motivation (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this subject variable was used as a post hoc matching variable across the conditions of this IV, so subject variable ( matched ( balanced ( control variable

motivation (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any variable that is a confound for either of the main effects is a confound of the interaction effect

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this is a procedural variable – while the amount of homework each participant completed was determined by that participant (rather than held constant, etc.), this was a variable that “occurred” during the procedure and should have been controlled/manipulated by the researcher (but wasn’t). We can not reasonably assume that people in the different amount of training conditions did the same amount of homework so, manipulated/procedural variable ( confound ( ongoing equivalence problem

amount of homework completed (with respect to main effect of type of training)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this procedural variable was the basis for forming post hoc equivalent groups (since this is a procedural variable it may be called “yoking” instead of matching, but …), so ( manipulated/procedural variable ( matched/yoked ( balanced ( control

amount of homework completed (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any procedural variable that is a confound for either of the main effects is a confound of the interaction effect

prior math skill (with respect to main effect of age group)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this subject variable isn’t involved in any of the post hoc matching, there wasn’t random assignment and it is not reasonably a constant, so subject variable ( confound ( initial equivalence confound

prior math skill (with respect to main effect of training amount)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

this subject variable isn’t involved in any of the post hoc matching, there wasn’t random assignment and it is not reasonably a constant, so subject variable ( confound ( initial equivalence confound

prior math skill (with respect to main effect of the interaction)

subject manipulated IV DV confound control ongoing eq problem

balanced eliminated held constant randomized matched initial eq problem

any subject variable that is a confound for either of the main effects is a confound of the interaction effect

So, causal interpretability isn’t possible with these data, however…

Because of matching on IQ, Motivation and Amount of homework, we are relatively sure that the main effect of Age was not produced by any of these variables (but confounding by a gazillion other variables is possible).

Because of matching on IQ , we are relatively sure that the main effect of Amount of Training was not produced by any of these variables (but confounding by a gazillion other variables is possible with Motivation and Amount of homework as known problems – remember that we couldn’t match on these because of the group differences).

The interaction is “safe” from IQ confounding, only – which is still better than nothing!

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download