Home > Teaching > SYG 3300 > Assignments > Lab 5

Lab 5: Comparing groups; graphical displays; significance tests

Note:  This assignment is due by 9:00 a.m. Friday, June 16.  You may submit it to me at my office Friday morning, or you may fax it to me at (603) 963-0087.  If you would like me to return your work, please include a self-addressed envelope with sufficient postage.

1. Does religious behavior have an effect on life expectancy? Do people who attend church often live longer than those who attend church infrequently? We can use the GSS datasets to find a preliminary answer to these questions. Use the variables ATTEND and AGE in the dataset CH10END.

1. Produce boxplots to examine the distribution of AGE for the two groups of respondents who vary in church attendance. Are there differences in the shapes of these distributions? Judging from the boxplots, does it appear that the two groups differ in terms of age?
2. Take a closer look at the difference in age between those who attend church at least once a month and those who attend church less than once a month. What is the mean age for each group in our sample?
3. Is there evidence to conclude that, in the population, frequency of church attendance is associated with differences in life expectancy? Report and interpret the t-statistic and its associated significance value.

2. It's possible that any observed relationship between church attendance and longevity could be spurious. One possibility is that both increased church attendance and increased life expectancy are dependent on level of education, an indicator of social status. For a preliminary look at this possibility, compare mean years of education (EDUC) of those who attend church more than once a month and those who attend less than once a month.

1. a. What is the sample mean years of education for each group? Is the mean years of education for those who attend church regularly greater than, less than, or the same as for those who attend church less regularly?
2. b. What is the 95% confidence interval for the true difference in mean years of education? Interpret this confidence interval.

3. Choose a comparison that interests you and conduct the appropriate analyses. Make sure that your grouping variable contains just two groups. Examples include RACED, SEX, AGER2, and RINC91D. Then, choose a comparison variable at the scale level of measurement. Produce boxplots, report sample means, and indicate whether the observed difference in sample means is likely to reflect a true difference in population means. Include the appropriate statistics to support your inference to the population.