Although the MIDUS2011 data set is quite large, we can filter and sample the data set to simulate a “small-N” situation. For these exercises, be sure to have the MIDUS2011 data set open.
- To simulate a “small-N” situation, let’s look at the relationship between income and life satisfaction only among participants who are 40 years old.
- DATA > SELECT CASES > Click the “If” button > Select the RA1PRAGE variable in the left box, and then click the transfer arrow button to move it to the right box. Click the "=" button and then type “40” and click Continue.
- Now let’s run our analysis.
- ANALYZE > CORRELATE > BIVARIATE
- Drag the variables RA1PB16 (income) and RA1SR1 (life satisfaction) over to the “variable” field
- Under the Correlation Coefficients heading, check the Pearson box
- Click OK.
- Describe what you have found. How do you think these results would differ if we had a greater N? You can find out by using the DATA selection procedure in 1a to select “all cases” and then repeating the correlation analysis.
- At the top banner, select Analysis > Correlation Matrix
- Type the variables RA1PB16 (income) and RA1SR1 (life satisfaction) into the boxes under “Variables to Correlate”
- Type RA1PRAGE(40) into the “selection filters” box
- Check to be sure that the box beside “Main statistic to display” says “Pearson correlation”
- Choose Run Correlation.
- Describe what you have found. How do you think these results would differ if we had a greater N? You can find out by removing the selection filter in 1b and then repeating the correlation analysis.