homework lab

Lab 7 Problem: A researcher uses a repeated-measures study to examine the effectiveness of a therapy procedure for treating depression. A sample of n = 8 depressed patients is obtained and each individual is given a depression test immediately before therapy, immediately after therapy, and once again 3 months after therapy. The data are to be examined with a repeated-measures ANOVA. Questions:

1) What does significantly different mean for this scenario?

2) What is the null hypothesis?

3) What is the alternative hypothesis?

4) What is our test statistic?

a. Which variance is affected when the sample means get farther apart? How?

b. Which variance is affected when scores within a group get farther apart from one another? How?

c. What is our expected test-value under the assumed null?

d. Is it a 1-tailed or 2-tailed test?

Page 1 of 4

Lab 7 5) Data:

Before Therapy After Therapy 3 Months Later Total n1=8 n2=8 n3=8 N = 24 T1=136.3 T2 = 134 T3 = 85.1 G = 355.4 SS1 = 25.54 SS2 = 30.14 SS3 = 17.90 k = 3 ΣX2 =2347.75 ΣX2 = 2274.64 ΣX2 = 923.15 ΣX2 =5545.54 M1 = 17.04 M2 = 16.75 M3 =10.64

6) Calculate SS’s:

a. SSTOTAL =

b. SSWITHIN =

c. SSBETWEEN =

7) Compute dfs: a. dfTOTAL =

b. dfWITHIN =

c. dfBETWEEN =

8) Calculate MS and df necessary for repeated-measures steps:

a. SSBETWEEN SUBJECTS (OR SSINDIVIDUAL DIFFERENCES) =

b. SSERROR =

c. dfBETWEEN SUBJECTS (OR dfINDIVIDUAL DIFFERENCES) =

d. dfERROR =

9) Calculate MS’s: a. MSBETWEEN =

b. MSERROR =

10) Calculate F-ratio: a. F =

Page 2 of 4

Lab 7 11) What is our critical value?

12) What do we conclude?

13) What is the calculated effect size?

14) Compute post-hoc tests: a. Tukey HSD

b. Scheffe

i. Immediately Before vs. Immediately After

Source SS df MS F Between Within B/S (ID) Error Total

ii. Do we reject the null hypothesis?

iii. Immediately Before vs. 3 months later

Source SS df MS F Between Within B/S (ID) Error Total

Page 3 of 4

Lab 7 iv. Do we reject the null hypothesis?

v. Immediately After vs. 3 months later

Source SS df MS F Between Within B/S Error Total

vi. Do we reject the null hypothesis?

Page 4 of 4