Overview
Dataset: GSS 1972-2024, use YEAR >= 2000. Key variables: HEALTH, MNTLHLTH, PHYSHLTH, REGION, EDUC, RINCOME, AGE, SEX.
Instructions: Use SPSS with syntax. Clean data (e.g., handle missing values).
Question 1: Aggregate Health (3 points) Aggregate HEALTH by REGION to add health_mean and health_sd. Compute z_health = (HEALTH - health_mean) / health_sd.
Provide syntax. Report z_health mean and SD. Interpret: How does this highlight health disparities?
Question 2: Deviation Score (3 points) Aggregate total_unhealthy = MNTLHLTH + PHYSHLTH by REGION to add mean_unhealthy. Compute unhealthy_dev = total_unhealthy - mean_unhealthy.
Provide syntax. Summarize unhealthy_dev by REGION. Interpret: How can this aid epidemiology studies?
Question 3: Multi-Level Aggregation (3 points) Aggregate by REGION for mean_health, mean_educ, n_cases. Then aggregate by SEX within REGION. Match back to flag HEALTH > mean_health + health_sd.
Provide syntax. Report number of flagged cases. Interpret: How can this guide interventions?
Question 4: Restructure Data (3 points) Restructure GSS panel (or simulate with YEAR: 2000, 2010, 2020) to wide format for HEALTH. Compute health_change = health_2020 - health_2000.
Provide syntax. Report health_change descriptives. Interpret: What health trends does this show?
Question 5: Advanced Stats (3 points) Run ANOVA on mean_health by REGION or logistic regression on poor_health (HEALTH >=3=1) with SEX, REGION, EDUC.
Provide syntax and key output (e.g., p-value). Interpret: How do findings relate to health disparities?
Submit your SPSS output, interpretations, and any calculations using pdf/word doc. Use the GSS dataset for all questions.
Question 1 (5 points):
Using the GSS data, perform an independent samples t-test to compare the mean self-rated health (HEALTH) between males and females (SEX). Include the homogeneity of variance test (Levene's). Report the t-value, degrees of freedom, p-value, and confidence interval. Interpret the results in the context of public health disparities. If variances are unequal, use the appropriate adjustment.
Question 2 (5 points):
Conduct a one-way ANOVA to examine differences in general happiness (HAPPY) across racial groups (RACE). Check assumptions including homogeneity of variance. If the ANOVA is significant, perform a Tukey post hoc test to identify which groups differ. Report the F-statistic, p-value, and key post hoc findings. Discuss implications for public health policy.
Question 3 (5 points):
Test for linearity in the relationship between age (AGE) and self-rated health (HEALTH) using the Means procedure or Curve Estimation. If nonlinearity is present, calculate and interpret eta squared to assess its strength. Create a scatterplot with a fitted line. Explain how this might influence public health interventions for aging populations.