Data Mining Exercises

Marriage and Family Changes in the United States

Marriage in America and most other postindustrial societies is not what it used to be. We have all heard about rising divorce rates, and many other marriage customs—whether we get married at all, whether couples cohabitate, and whether we have smaller families or even remain childfree—are also transforming. Once again, the General Social Survey (GSS), with its yearly or biennial surveys of our adult population, can be very helpful in trying to explore these changes.

Part I 

First, let’s look at how the pattern of divorce has changed across the last several decades.   There are different ways that sociologists measure divorce and talk about changes in the pattern of divorce across time. You may have heard it said that one-half of all marriages today end in divorce. This statistic reflects the probability that a marriage formed today will eventually end in divorce, given certain assumptions about current rates. A more accurate measure of divorce is to calculate a standardized annual rate of divorce per 1,000 people. By measuring the incidence for a standard unit of population, we are able to control for changes in the overall size of the population and compare rates at different points in time. Table 1 provides information on both the number of divorces in the United States and the rates of divorce; the information is taken from the Statistical Abstract of the United States, an annual publication of the U.S. Census Bureau. 

Table 1: The Number and Rates of Divorce in the United States, for selected years from 1970–2004
Year	Number of divorces	Rate per 1,000 population
1970	708,000	3.5
1975	1,036,000	4.8
1980	1,189,000	5.2
1985	1,190,000	5.0
1990	1,182,000	4.7
1995	973,000	4.1
2000	Not available (NA)	4.1
2004	NA	3.7
2006	NA	3.6

Did you find these statistics surprising, given all the attention to divorce in recent years? As you can see from Table 1, when we calculate the actual rate, the number is much lower than the projection of one-half of all marriages ending in divorce.

We can also calculate what percentage of the population is divorced or separated. Table 2 provides information from the U.S. Census taken from annual Current Population Surveys between the years 1970 and 2009. In these surveys, samples of the American population were asked about their current marital status.

Table 2: The Percentage of the U.S. Population Aged 15 and Over That Is Separated or Divorced, for Selected Years from 1970 to 2009
	Males	Females
1970	3.5	5.7
1980	6.6	9.4
1990	8.8	11.7
1995	9.4	12.7
2000	10.1	12.6
2005	10.0	13.4
2009	10.4	13.4

Writing Assignment 1

Think of a sociological explanation for why the percentage of divorced or separated males was smaller than that of females.

Now, let’s see how the respondents who participated in the GSS compared with these national surveys by the U.S. Census Bureau. Regularly, between 1972 and 2002, married and widowed respondents to the GSS were asked if they had ever been divorced. We’ll start by going to the Survey Documentation and Analysis site at the University of California, Berkeley, and following these steps:

• 1. Enter DIVORCE as the Row variable.
• 2. Next, we will use YEAR as the Column variable choice; however, we want to collapse the long list of yearly surveys by applying a command to code the data into a shorter list of combined categories. Therefore, enter the following syntax as the Column variable, which will recode the YEAR variable: YEAR (r:1972-1974; 1975-1977; 1978-1982; 1983-1985; 1986-1988; 1989-1991; 1993-1996; 1998-2002).
• 3. Last, mark the boxes to request “Column” percentages, “Color coding,” and “Question text”; then press “Run the Table.”

Writing Assignment 2

Examine the resulting table carefully. How would you describe the pattern of divorce since the early 1970s? How do the patterns you see in the GSS data compare with those published by the U.S. Census? Using your sociological imagination, identify social factors that may have contributed to the particular outcome you obtained.

Part II

Next, let’s investigate how marriage rates have also changed over time. Once again, begin at the Survey Documentation and Analysis site at the University of California, Berkeley, and following these steps:

• 1. We will use MARITAL as the row variable; however, you will need to recode this variable, and enter the following syntax: MARITAL (r:1-4 "Ever Married"; 5 "Never Married").
• 2. For the column variable, apply the same year code command with the same set of additional recode commands as you did in the previous part of the exercise—YEAR (r:1972-1974; 1975-1977; 1978-1982; 1983-1985; 1986-1988; 1989-1991; 1993-1996; 1998-2004) in order to see the pattern of people marrying over the 30-year period of the surveys.
• 3. Last, mark the boxes to request “Column” percentages, “Color coding,” and  “Question text”; then press “Run the Table.”

Writing Assignment 3

Interpret this data and determine what has been happening to the marriage rate over the last thirty years. Are there any points of sharp rises or falls? Discuss these results in depth. Do you think the factors that influenced the patterns of divorce have also contributed to the patterns of marriage and non-marriage?

Part III

Another issue to explore with the GSS data is acceptance of cohabitation before marriage. The GSS has included three different questions on cohabitation; the years in which the question was included in the survey are noted in parentheses.

• Did you live with your husband/wife before you got married? (1988, 1994)
• Do you agree or disagree that it’s a good idea for a couple who intend to get married to live together first? (1994, 1998, 2002)
• Do you agree or disagree that it is alright for a couple to live together without intending to get married? (1994, 1998, 2002)

Each question reflects a different aspect of and approval for cohabitation. In this part of the exercise we will examine how behavior, as measured by the first question, and attitudes, as measured by the second and third questions, may have changed across time.

Let’s begin by looking at changes in behavior. Once again, follow these steps:

• 1. Again, at the Survey Documentation and Analysis site, type in COHABIT for the row variable. 
• 2. For the column variable enter YEAR. 
• 3. Last, mark the boxes to request Column percentages, Color coding, and Question text, as well as the one for Statistics, and press “Run the Table.”

Writing Assignment 4

Examine whether any changes in cohabitation occurred over this six-year period. How do sociologists interpret such findings? 

An additional question is whether it is a class-related behavior. To answer this question, follow these steps:

• 1. At the Survey Documentation and Analysis site, keep COHABIT as the Row variable.
• 2.  Put CLASS in as the Column variable. 
• 3. This time, enter YEAR as a Control variable.
• 4. Mark the boxes to request Column percentages, Color coding, and Question text, as well as Statistics, before pressing the “Run the Table” button. 

Writing Assignment 5

The output will be organized into three tables: one for each of the two years in which this question was asked (1988 and 1994) and a third one for both years combined. Begin by examining the third table, for the two years combined. What are the differences between social classes in terms of cohabitation? Which class was most likely to cohabit prior to marriage and which was least? What do the summary statistics suggest? Now, look at the relationship between cohabitation and social class for the two years (the first and second tables). Can you observe any changes in behavior among the classes between 1988 and 1994? Can you develop a sociological explanation for these patterns?

Before we leave the question of cohabitation, let us look at people’s attitudes. We can use the GSS questions from 1994, 1998, and 2002 in our analysis of changing attitudes. Follow these steps:

• 1. From the Survey Documentation and Analysis, use the row variables COHABFST and COHABOK.  The first variable is tied to the question about cohabitation before marriage and the second, to cohabitation without plans for marriage. 
• 2. For the column variable enter YEAR.
• 3. Last, mark the boxes to request Column percentages, Color coding, and Question text, as well as Statistics, then press “Run the Table.”

Writing Assignment 6

Have attitudes about cohabitation changed since the mid-1990s? Did you find any differences between the two questions? What factors do you think contribute to any observed changes?

Part IV

Next, let us examine the matter of having children. Are people having bigger or smaller families today than they did years ago? GSS respondents have regularly been asked, How many children have you ever had? Please count all that were born alive at any time (including any you had from a previous marriage). We will be using the responses to this question to determine if there have been any changes in family size. Follow these steps:

• 1. Again, at the Survey Documentation and Analysis site, we will use the row variable CHILDS with the following recode: CHILDS (r:0 “No children”; 1 “1 Child”; 2 “2 Children”; 3-8 “3-8 Children”).
• 2. To determine the pattern of child bearing over time apply the following YEAR code command: YEAR (r:1972-1974; 1975-1977; 1978-1982; 1983-1985; 1986-1988; 1989-1991; 1993-1996; 1998-2004). Last, mark the boxes to request Column percentages, Color coding, and Question text and press “Run the Table.”

Writing Assignment 7

Based on this data, what conclusions can you draw about changes in the number of children people are having now? Are more people today having only one child or even remaining childless compared to thirty years ago? Do you think this has become more acceptable? What broader social factors may be contributing to any observed changes? 

If you wish, it might be interesting to explore this question by examining other possible differences among respondents. For instance, by entering CLASS as the Column variable, you can look for any differences by social class. If you wanted to find out whether a woman’s employment is related to the number of children she has, you could enter the variable for work status, WRKSTAT, as the Column variable and then enter the recoded syntax SEX (2) as the Control variable.

Part V

Writing Assignment 8/Essay

Compare and contrast these GSS-based findings with what your textbook authors have presented on recent American family changes. How does the textbook discussion fit with the data that you interpreted here? What do you think the trends will be in the next five, ten, or even twenty years?

This page is printer-friendly: