2020-intro to statistics in sociology analysis of variance-assignment docx PDF

Preview

Summary

2020-intro to statistics in sociology analysis of variance-assignment docx...

Description

Sociology 312 Introduction to Statistics in Sociology Professor Quan Mai Assignment 4: Analysis of Variance This assignment is due November 15 at 11:55PM. Turn in your computer printout and your typed discussion of the three problems. Be sure that you include pages that show all of your work; you can scan or take a photo of hand-written work to upload it. Please paste images of tables and your hand-written work into ONE .doc file. All materials must be uploaded to Sakai. 1. Use SPSS and the General Social Survey (GSS) data set provided to the class to perform a hypothesis test about differences in means among three or more groups (or categories) of individuals living in the United States. To perform this analysis of variance, you need to choose variables you are interested in that are appropriate for conducting this type of analysis. Look carefully through the class GSS codebook (available on sakai), decide on two appropriate variables for your analysis, and develop a hypothesis concerning the relationship between these two variables. Use SPSS to run the analysis needed to test your null hypothesis and save the results in your output. (You cannot use the same two variables provided in the example in recitation.) Write up all of the steps to perform your hypothesis test (with an alpha of .01). Be sure to include equations and written statements for the null and alternative hypothesis. Describe your summary conclusion in complete sentences. Lastly, discuss the strength of the relationship based upon the value and meaning of the appropriate measure.

Step 1: H0 – There is no difference between how often someone attends a religious service and if they decide to get an abortion. Ha – There is a significant difference between how often someone attends a religious service and if they decide to get an abortion.

Step 2: A= .01 Df between = 1 Df within = 1225 Critical F = 131.077 Step 4: Calculate sample F = 914.213/6.975= 131.069 Critical F > sample F so accept H0 A woman’s decision to go to church is not affected by the amount of times she attends religious services. Eta sq= BSS/TSS = 914.213/9458.148= .097 9.7% of the total variation in whether a woman gets an abortion is explained by how often she attends religious services. This is a weak relationship.

2. Explain in words what the data for your dependent variable would look like (i.e., how would the values of this variable be distributed) if there was a perfect relationship between your two variables. If there was a perfect relationship, the graph would look like a bell curve. The relationship would be very strong and those women who attended religious services most often would be least likely to get an abortion.

3. You are currently conducting research on the effects of school and work on stress. For this research, you collected data on stress levels, measured on a scale from 1 (least stressed) to 25 most stressed. You have data from 15 young adults divided into three groups: those who both work and attend college; those who only attend college; and those who only work. Data for the three groups are shown in the table below. Work and School

School Only

Work Only

18.0

13.8

11.6

22.2

16.6

13.4

12.4

9.2

9.4

15.8

10.4

8.6

13.2

12.8

9.6

You predict that these three groups of young adults experience different levels of stress. Write up the appropriate step-by-step hypothesis test to determine whether there is a relationship between school/work and stress (alpha of .01). Be sure to include equations and written statements for the null and alternative hypothesis. Describe your summary conclusion in complete sentences. If it is appropriate, discuss the strength of the relationship based upon the value and meaning of the relevant measure or relationship strength. Step 1: H0 – People who attend work and school do not differ in stress levels from people who only attend work or only attend school. Ha – People who attend work and school face a significant difference in stress levels from people who only attend work or only attend school. Step 2: A=.01 Df= (r-1)(c-1) = 2*2=4 Critical x sq = 13.277 Step 4: X squared sample = 21.09 Step 5: sample > critical, reject H0 People who attend work and school face significant differences in stress levels from people who only attend work or only attend school. (teepee symbol) = (81.6+79.6) – ((4+2) + (6+5) + (4+3))) / (81.6-79.6) 161.2 – 24 / 2 = 68.6 One would make 68.6% fewer errors in determining how much stress people faced with work and school than just work or school. This is a strong relationship....