Question 1
STUDY QUESTIONS GROUP DIFFERENCES Factual Questions 1. By visual inspection alone, at - which site did the average stu- 5. For Site B, should the null dent have a higher score on hypothesis be rejected for the the pretest? main effect of time? If yes, at 2. For which site are the two what probability level? lines more nearly parallel? 6. For Site B, should the null (Note: Parallel lines indicate no - hypothesis be rejected for interaction.) the main effect of condition 3. For Site A, should the null (treatment vs. control)? If yes, `hypothesis be rejected for the at what probability level? main effect of time? If yes, at 7. For Site B, should the null what probability level? hypothesis be rejected for the 4. For Site A, should the null interaction? If yes, at what - hypothesis be rejected for the probability level? interaction? If yes, at what probability level? Questions for Discussion 8. For Site B, why do you think 10. Why might Site A find a the researcher did not give the interaction effect, i.e., that the p value for the interaction? treatment program improve 9. Consider Site A. Would the scores, while Site B found n - results have been as informative interaction? if the researcher had not tested for an interaction? Explain.EXCERPT FROM THE RESEARCH ARTICLE Appli Within each site, I performed repeated-measures analyses on oral and sight vocabulary, with condition (treatment and control) as the between-subjects factor and time (pre- and posttest) as the within- subjects factor. Figure 1 shows graphic changes in oral vocabulary, with Site A pairs of pre- and posttest mean scores plotted in the lower half and Site B pairs of pre- and posttest mean scores plotted in the upper half. As suggested graphically in Figure 1, oral vocab- ulary gains appeared to be affected differentially by condition in Site A but not in Site B. In Site A, results of the repeated-measures ANOVA performed on oral vocabulary revealed a significant effect on time, F(1, 114)=12.95, p<.001 a significant timex condition interaction effect f p and nonsignificant effect. in site b results of the repeated-measures anova performed on oral vocabulary revealed time but x to summarize effects as plotted figure indicate that teachers use elements reading: had positive not b. treatment control mean raw score pretest posttest reading diagnostic assessment full composite by highest possible is ii improvement program statistical guide eview an experiment which there more than one independent variable nova said employ factorial design. two-way common statisti- new xercise cal procedure for designs. discussed exercise main are impact itself inter- action two variables combined. designs can have number different patterns results. indeed at least eight possibilities possibility only was no between independ- ent variables. another neither their own led consult statistics textbook if you want see all laid out. background note study below states within e eated-m analyses used. when researchers term repeated measures they usually anova. phrase used because repeats same measure with participants times. precisely this uses both measures. it two- variable. abulary measured>
