Archmage

Advanced Statistics (Psyc 305) Name Spring, 2016 -- Project #2

Nonparametric Statistics

Whenever an answer has many decimal places round it to two or three decimal places. P-values should be to three decimal places.

1. Using the salary.sav data file and the “select cases” function of SPSS restrict the file to only clerical workers (variable jobcat = 1). Then assess the parametric normality and homogeneity of variance assumptions for a potential analysis comparing the beginning salaries (variable salbeg) between males and females (variable sex). Report the findings of you assessment below:

2. Building on item 1 above, use a nonparametric statistic to test the null hypothesis that beginning salaries (variable salbeg) for men and women (variable sex) are equal for clerical workers (variable jobcat = 1). If the test is significant, then also report the effect size. Report the results of your analysis in APA style:

2 3. Once again using the salary.sav data file and the “select cases” function of SPSS restrict the

file to only clerical workers (variable jobcat = 1). Then assess the parametric normality and homogeneity of variance assumptions for a potential analysis comparing beginning salaries (variable salbeg) of the four gender/race groups (variable sexrace). Report the findings of you assessment below:

4. Building on item 3 above, use a nonparametric statistic to test the null hypothesis that beginning salaries (variable salbeg) for the four sex/race categories (variable sexrace) are equal for clerical workers (variable jobcat = 1). Report the results of your analysis in APA style. If the overall effect is significant, then do and report all pairwise significance tests along with the effect size for any comparison that is significant.

3 5. Using the gss.sav data file test the null hypothesis that men and woman (variable sex) are

equally likely to believe in life after death (variable postlife). Complete the following contingency table by inserting (in vertical order) observed frequencies, expected frequencies, and column percentages in each cell

Men Women

believe in life after death

O = E = % =

O = E = % =

do not believe in life after death

O = E = % =

O = E = % =

If the null hypothesis were true, what would be the best guess for the percentage of males who believe in life after death?

% of males =

÷ = df = Cramer’s V = 2 Write about significance/nonsignificance and effect size with statements suitable for a journal.

4 6. Using the gss.sav data file test the null hypothesis that people in various education

categories(variable degree) are equally likely to believe in life after death (variable postlife). Complete the following contingency table by inserting (in vertical order) observed frequencies, expected frequencies, and column percentages in each cell:

less than HS High School Jr College Bachelor Grad School

believe in life after death

O = E = % =

O = E = % =

O = E = % =

O = E = % =

O = E = % =

do not believe in life after death

O = E = % =

O = E = % =

O = E = % =

O = E = % =

O = E = % =

÷ = df = Cramer’s V = 2 Write about significance/nonsignificance and effect size with statements suitable for a journal.

5 7. Is the relationship between belief life after death (variable postlife) and highest degree

earned (variable degree) the same or different for males and females (variable sex)? In two sentences explain below how it is ithe same or different.