INSTRUCTIONS FOR HOMEWORK
YOU SHOULD INDICATE AT THE BEGINNING OF ANY EXPLANATION FOR EACH EXERCISE WHAT THE STATISTICAL TEST IS ACTUALLY MEASURING. IT WILL HELP YOU BETTER UNDERSTAND THE UTILITY OF EACH TEST
1) Test whether there is an association between a person’s gender and the prestige of their occupation. Use the GSS2018 data set to perform an independent samples t-test on SEX and PRESTIG10 depending on the dataset. Report the following:
Mean prestige score for men ___________________
Mean prestige score for women ___________________
t-test equality of means significance level ___________________
Is the relationship statistically significant Yes No?
Based on these data, would you say that gender is associated with occupational prestige? What could explain this relationship? Remember to explain this relationship based on what the test measures.
2) You will report the information listed below for the GSS 2018 data set: Perform an independent samples t-test to compare the mean socioeconomic index (SEI10) of those who have had a born-again experience with those who have not (REBORN). NOTE: SEI: Socioeconomic index scores reflect the education, income, and prestige associated with different occupations.
Mean SEI of those non-born again ___________________
Mean SEI of those born again ___________________
t-test equality of means significance level ___________________
Is the relationship statistically significant Yes No?
Based on these data, would you say that the religious experience of being born again is associated with socioeconomic status? Remember to explain this relationship based on what the test measures.
What were you expecting and what did you find?
3) Perform a paired t-test to compare the respondent’s mother’s occupational prestige score (MAPRES10) to the respondent’s father’s occupational prestige score (PAPRES10) using the GSS2018 data set.
Respondent’s mother’s occupational prestige score? ___________________
Respondent’s father’s occupational prestige score? ___________________
Significance for the Paired Samples Test? ___________________
Is the relationship statistically significant Yes No?
Is prestige related to generation? What were you expecting to find, and did you find it?
USE STATES10 DATA FOR THE NEXT SET OF QUESTIONS
4. Perform a paired t-test to compare the median earnings of male full-time workers (EMS168) to the median earnings of female full-time workers (EMS169) using the STATES10 data set.
Mean earnings of men? ___________________Mean earnings of women? ___________________Significance for the Paired Samples Test? ___________________Is the relationship statistically significant Yes No Are earnings related to gender? Remember to explain this relationship based on what the test measures. What were you expecting to find, and did you find it?
5. Using a variable called WAGEGAP that is the difference between median earnings of male full-time workers (EMS168) and the median earnings of female full-time workers (EMS169),
Create a histogram of WAGEGAP’s distribution (Remember Histogram is under either frequencies-click on chart from the dialog box OR under GRAPH). GRAPH IS PART OF THE RIBBON OF COMMANDS ALONG THE TOP
Describe the shape of the distribution.
In which state do women have earnings closest to men’s? ___________________
In which state do women’s women have earnings most disparate from men’s? ___________________
(If you are having trouble with this, you can sort the data (sort cases) in DATA VIEW from lowest to highest by the WAGEGAP variable and …)What might account for the variation in wage gaps observed across states?
6. From the States10 dataset preform a paired sample t-test and report the results using Overdose Deaths 1999 and 2005 (Pair 1) and Overdose Deaths 2005 and 2017 (Pair 2).
Pair 1
Mean ________? Mean ________?
Significance for the Paired Samples Test? ___________________Is the relationship statistically significant Yes No
Pair 2
Mean ________?
Mean ________?
Significance for the Paired Samples Test? ___________________Is the relationship statistically significant Yes No
What do the t-test results suggest about deaths by overdose over the two-decade period?