UNIT 13: Earnings: White Anglos, Blacks, Latinos, and Asians, and men and women

In the last unit we looked at the men-women gaps, and the white-black gaps in earnings, among year-round full time workers. This time we are going to add two more minorities, Latinos and Asians, to the earnings comparisons. And we will look at changes in earnings depending on the age of workers.

In the AI1 folder (DOS: subdirectory) open datafile EARN9.DAT, from the 1990 Census, and start your log. As usual, do Info to put the file name at the top of your output, and then All Marginals.

The Earning marginals show the usual thinning out of earners as earnings go up. But to work with this variable, we need to simply it, as we did before. Open the Modify menu, pick Combine, and select the variable Earning.

First, pick <15K, then 15-25K, then No more. Label this Lt25K. Second, (and be careful not to include your Lt25K in your picks) pick 25-35K, then 35-50K, then 50K+, then No more, and label this new combined category 25Kup. If you make a mistake during the Modifying, you will need to go to Open and open the datafile again and start over.

Crosstab Racelat / Earning, pct across, and make a slope graph, "Fig. 13.1: Percent earning $25,000 and up, full-time year round workers, by Race-ethnicity. Source: 1990 Census. <EARN9.DAT." You may do the graph by hand, use a computer program, or have CHIP make the graph, with Plot, Slope, and pick 25Kup as the category to graph. Put it on the screen with the full 0-100 vertical scale first, then do it again, and change the 100 to a percentage just above the highest point you saw the first time. Remember that this graph does not go into your output file. You have to print it at once on an attached printer. See Manual on this printing.

Q1. A fairly simple pattern emerges. Describe the pattern of differences in earnings among the four race-ethnic groups.

Q2. What difference do you expect age, thinking of the categories in the Marginals, to make to people's earnings?

Let's go right to a comparison of our four groups, over the 6 age categories. There are two ways to do this, depending on whether you make your own graph, or let CHIP make it for you. If you make your owngraph, Crosstab Age / Earning, Control RaceLat, and percent across. You will get four subtables of Age / Earning, one for each population group. Make Fig. 13.2, "Percent earning $25,000 or more, by race-ethnicity and age, full-time year round workers. Source: 1990 Census, <EARN9.DAT." This is a Fishnet graph. Horizontal scale, the 6 age categories, and one line of the 25Kup percents for each of the four race-ethnicities.

When CHIP makes the Fishnet graph, you have to set up the Crosstab differently. The rule is, there will be a line for each category of the independent variable. And the control variable is the one whose categories will form the horizontal scale across the bottom. A little tricky. So Crosstab Racelat / Earning, control Age. Go to slope, and pick Fishnet. As before, run it once to the screen so you can see it, then again, this time probably picking a new upper-limit percentage. And print it straight to the printer. You will need to write a title over these CHIPmade graphs. Not really publishable, but it does save work. What not to do: After you have set up the crosstab, go to Plot and pick Slope. Slope is only for two-variable graphs, without control. If you use it on a three-variable graph, it doesn't make a fishnet, it makes four different slope graphs. and we don't want that, because we want to have the lines all in one frame, so we can compare them.

Q3. What happens to the pay pattern in Fig. 1, when we look at separate age groups? What does the picture suggest? That minorities take longer to learn their jobs, but in time they catch up and make as much as whites? What explanations might there be for the pattern you see?

Now let's look at men-women differences in earnings. Follow the same directions for graphing. Crosstab, and graph Gender / Earning, for Fig. 13.3: "Percentage earning $25,000 and up, by gender, full-time year round workers. Source: Census 1990. <EARN9.DAT"

Q4. Compare the gender gap in pay to the various race-ethnicity gaps (that is, difference from white Anglos).

For Figure 13.4, "Percentage earning $25,000 or more, by gender and age, full-time year round workers, etc. (full title please), Crosstab Age / Earning, control Gender if you are making your own graph. If you are graphing with CHIP, then it's Gender / Earning, Control Age.

Q4. As we move up the age levels, what happens to the gender gap in pay?

Finally, let's compare the genders within the race-ethnic groups. Crosstab Racelat / Earning, Control Gender if you make your own Fishnet, and title it "Figure 13.5, Percent earning $25,000 or more, by Gender and Race-ethnicity, full-time year-round workers. Source: Census 1990. <EARN9.DAT.

Q5. What is the pattern of earnings, comparing the genders and the racenicicties together?