Oneway ANOVA Homework with Trend Analysis
A psychologist studying social behavior among children examined the number of times they freely shared their toys. The psychologist videotaped the children in numerous different settings and then computed their proportions of sharing behaviors. Four age levels of children were videotaped, and the results reported below. Consider the observations to be independent both between and within groups.
3 yrs 5yrs 7yrs 9yrs
32 15 54 80
10 20 65 78
25 10 70 84
5 14 58 65
35 16 10 92
8 16 68 64
34 20 75
30
40
4
8
1. Enter the data into SPSS and conduct a between-subjects oneway ANOVA. Test the homogeneity of population variances assumption and also examine the boxplots or histograms for any anomalies.
2. You should notice at least one outlier and also a clear violation of the homogeneity assumption. Review Howell’s discussions of these problems (sections 11.9 and 11.10). Discuss amongst yourselves how to correct for these problems, and at least try the following:
1) remove the outlier(s) and re-run the analysis. Summarize the results.
2) try a data transformation and re-run the analysis. Summarize the results.
Would Winsorizing remedy the problems in this data set? Is there anything about the data set that would cause concern for Winsorizing?
3. Complete problem 12.25 (a through e) from Howell. You may use SPSS to conduct all of the analyses. Note that Howell is asking you to compare results when the IV is coded in SPSS in two different ways; first, as it is scaled (0, 1, 10, 100, 250, and 500); and second, generically as 1, 2, 3, 4, 5, and 6.