Ww2.justanswer.com



I. Paired Sample Design: 30 pts

Download the data set bph-samp.sav and use SPSS to complete the following calculations:

1) Run a paired t test to compare if there is a mean change in QoL at the baseline and at 3 months.

Ho: µD = 0

Ha: µD ≠ 0

[pic]

[pic]

[pic]

Since the p-value (.006) is smaller than 0.05, we reject the null hypothesis. We have sufficient evidence to prove that there is significant difference between the means at the baseline and at 3 months.

2) Run a one-sample t test to test the same hypothesis as in (1) but on the variable DELTA.

Ho: µDelta = 0

Ha: µDelta ≠ 0

[pic]

[pic]

Since the p-value (.006) is smaller than 0.05, we reject the null hypothesis. We have sufficient evidence to prove that the mean delta is different than zero.

The two methods gave the same result!

3) Check if the QoL at the baseline and at 3 months follow Normal distributions.

Ho: QoL at the baseline = normal

Ha: QoL at the baseline normal

[pic]

Since the p-value is greater than 0.05 we fail to reject the null hypothesis. So QoL at the baseline can be assumed normal.

Ho: QoL at 3 months = normal

Ha: QoL at 3 months normal

[pic]

Since the p-value is greater than 0.05 we fail to reject the null hypothesis. So QoL at the baseline can be assumed normal according to Kolmogorov-Simirnov test.

Note that the data doesn't pass the normality test in Shapiro-Wilk. The sample size is small to decide the normality at this point. We may want to do a nonparametric test.

4) If you have a concern about the small sample size and perhaps non-normal data, choose an appropriate nonparametric test to compare the median QoL score at the baseline and at 3 months. (Hint: You need to research nonparametric tests.)

[pic]

[pic]

Both Wilcoxon Signed Ranks Test and Sign Test prove that the median scores at the baseline and at 3 months are different.

Sign test may be more suitable for this case because the histograms don't seem to be symmetric.

II. Independent Sample Design: 30pts

Download the data set lactation.sav and use SPSS to complete the following calculations.

1) Produce a side-by-side boxplot for the percentage of bone loss in the breast feeding group vs. the control group.

[pic]

2) Produce a side-by-side histogram for the percentage of bone loss in the breast feeding group vs. the control group. (Hint: You need to search for SPSS functions not covered in the lecture to produce the histograms side-by-side.)

[pic]

3) Test if the mean percentage of bone loss in the two groups are the same using the right version of the t test based on the SPSS output.

H0: µ1 = µ2

Ha: µ1 ≠ µ2

[pic]

[pic]

Since the p-value is very small we reject the null hypothesis of equal means. We have sufficient evidence to prove that the control and the breast-feeding groups have different means.

4) Calculate a 95% confidence interval for the mean difference in the percentage of bone loss between the two groups.

95% confidence interval for the mean difference is calculated in the previous table as (2.7621, 5.0305)

III. Cross-Tabulation: 30pts

Download the data set bd1.sav and use SPSS to complete the following calculations.

1) Test the association between esophageal cancer and alcohol consumption (using the original alcohol consumption variable). Write down the hypotheses, the test used, the p-value and the interpretation.

H0: There is no association between esophageal cancer and alcohol consumption

Ha: There is association

[pic]

Since the p-value (.000) is very small ( ................
................

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

Google Online Preview   Download