Social Capital Over Time and Across Generations: A Data-Driven Learning Guide
Goal & Concept
Following from Robert Putnam's work, this exercise explores time and generational trends in measures of social capital from 1972 to 2004. Crosstabulation and comparison of means are used.
The term social capital is used to refer to connections among people and organizations. These social networks have important implications for social identity, emotional support, as well as the exchange of goods, services, and information.
In his influential book, Bowling Alone: The Collapse and Revival of American Community(2000), Robert Putnam identified seven dimensions of social capital (political participation, civic participation, religious participation, workplace networks, informal networks, mutual trust, and altruism), and showed that there are two distinct, though not necessarily mutually exclusive, types of social capital: bridging social capital (inclusive, outward-looking social networks), and bonding social capital (exclusive, inward-looking social networks). In addition Putnam explored changes in social capital over the 20th century, demonstrating that social capital increased between 1900 and the 1960s, then decreased dramatically as a result of pressures of time and money, mobility and sprawl, television, and generational differences.
This exercise uses several dimensions of social capital (political, civic and religious participation; informal networks; mutual trust; and altruism) to explore the time and generational trends that Putnam described in his book.
Examples of research questions about social capital:
- What factors account for changes in social capital over time?
- What are the consequences of high/low levels of social capital?
- Under what conditions might high levels of social capital be detrimental, and low levels of social capital beneficial?
- Have all dimensions of social capital evolved similarly over time?
- Do factors such as race, gender, and social class influence how much and what kind of social capital groups have?
This exercise will use the General Social Surveys, 1972-2006 [Cumulative File] . The General Social Survey (GSS) is one of the longest running surveys of social, political and cultural conditions in American society. Since its inception in 1972 it has been monitoring social change and tracking trends in attitudes, behaviors and attributes of the United States adult population. A nationally representative personal interview survey of the United States adult population, it collects data on a wide range of topics: behavioral items such as group membership and participation; personal psychological evaluations including measures of well-being, misanthropy, and life satisfaction; attitudinal questions on such public issues as crime and punishment, race relations, gender roles, and spending priorities; and demographic characteristics of respondents and their parents.
The survey is currently administered biennially. The GSS contains a core of demographic and attitudinal questions, many of which have remained unchanged to facilitate time trend studies. The survey also contains topical modules about topics of special interest. The cumulative dataset merges all previous years of the GSS into a single file, with each year or survey constituting a subfile.
This exercise will use the following variables:
- Age (AGE)
- Year (YEAR)
- Voted in the 1972 election (VOTE72)
- Vote in the 2004 election (VOTE04)
- How often attend religious services (ATTEND)
- Number of memberships to groups/organizations (MEMNUM)
- How often spend a social evening with relatives (SOCREL)
- How often spend a social evening with neighbors (SOCOMMUN)
- How often spend a social evening with friends (SOCFREND)
- Most people can be trusted (TRUST)
- How often done volunteer work for a charity in the past year (VOLCHRTY)
In this exercise we will explore time and generational trends for several measures of social capital over the 1972-2004 period. Note that some of the questions used in the GSS vary from year to year and not all questions are asked in every survey. The variables needed for this exercise come from different years and therefore different samples. While this is not ideal, each year's subfile should be fairly representative of the general population because the GSS uses a national probability sample.
For this exercise, we recoded the ordinal variable AGE into a more useful categorical variable, AGECAT, with 4 categories: 18-34 (coded as 1), 35-49 (2), 50-64 (3), and 65-89 (4).
Political, Civic, and Religious Participation
Voting behavior offers one measure of political and civic participation. The variables VOTE72 and VOTE04 indicate whether respondents voted in the 1972 and 2004 presidential elections. Examine the crosstabs of each of VOTE72 and VOTE04 with AGECAT. What percentage of 18-34 year olds voted in 1972? In 2004? Which age group was most likely to vote in 1972? In 2004?
The variable ATTEND measures how often respondents attend religious services. It contains 8 categories--too many for useful analysis. We collapsed the categories from 8 to 4, where 0 = never, 1 = several times a year, 2 = several times a month, and 3 = once a week or more. We named the new variable CHURCH.
Consider the crosstab between CHURCH and AGECAT, using YEAR (1972, 2004) as control. First take a look at the "row total" column in each table. Has the percentage of respondents who said "never" changed between 1972 and 2004? What percentage of respondents answered "once a week or more" in 2004 compared to 1972? Are these results consistent across age groups?
Examine participation in groups and organizations by running a comparison of means analysis between MEMNUM (number of memberships) and AGECAT, with YEAR (1974, 2004) in the column variable. Has the average number of memberships increased or decreased between 1974 and 2004? Which age group shows the most change?
The variables SOCREL, SOCOMMUN, and SOCFREND indicate how often respondents spend a social evening with relatives, neighbors, and friends respectively. We recoded each variable, collapsing the categories from 7 to 5 and reversing the coding so that higher numbers meant higher levels of sociability (from 0="Never" to 4="Several times a week"). We named the new variables SOCRELS, SOCNEIGH, and SOCFRDS.
Consider the results of the crosstabs between SOCRELS and YEAR, SOCNEIGH and YEAR, and SOCFRDS and YEAR, limiting the analysis to the years 1974 and 2004. Are there significant differences in how much people socialize with relatives, neighbors and friends in 2004 compared to 1974?
To assess whether the level of mutual trust has changed over the past 30 years, we ran a crosstab between the variables TRUST and AGECAT, using YEAR (1972, 2004) as control. First, take a look at the "Row Total" column in the first two tables. In 1972, what percentage of respondents said that most people can be trusted? In 2004? Which age group seems to have experienced the greatest change in the level of trust between 1972 and 2004?
Altruism will be assessed using the variable VOLCHRTY, which measures how often respondents have done volunteer work for a charity in the past year. This variable contains 6 categories: "More than once a week" (1), "Once a week" (2), "Once a month" (3), "At least two or three times in the past year" (4), "Once in the past year" (5), and "Not at all in the past year" (6). We recoded it to exclude those who responded "Don't know" or did not answer. We also collapsed the categories from 6 to 4 and reversed the coding so that higher numbers meant higher levels of altruism (3="Once a week or more" 2="Once a month" 1="One to three times per year" and 0="Not at all in the past year"). We named the new variable CHARITY. 2004 is the only year for which there is data for this variable.
Examine the results of the crosstab between CHARITY and AGECAT. Do there appear to be generational differences in volunteerism?
Note: The online data analysis system (DAS) used on this site uses a system called Survey Documentation and Analysis (SDA), developed and maintained by the Computer-assisted Survey Methods Program (CSM) at the University of California, Berkeley. Documentation for DAS/SDA can be found on their Web site.
Interpretation & Summary
Think about your answers to the application questions before you click through to the interpretation guide for help in answering them.
Weights (mathematical formulas) are often used to adjust the sample proportions, usually by race, sex, or age, to more closely match those of the general population. The analyses in this guide used weights to increase the generalizability of the findings, so the resulting tables are meant to reflect the relationships we would expect to see in the general population.
What percentage of 18-34 year olds voted in 1972? In 2004? Which age group was most likely to vote in 1972? In 2004?
Has the percentage of respondents who said "never" changed between 1972 and 2004? What percentage of respondents answered "once a week or more" in 2004 compared to 1972? Are these results consistent across age groups?
Has the average number of memberships in groups or organizations increased or decreased between 1974 and 2004? Which age group shows the most change?
Are there significant differences in how much people socialize with relatives, neighbors and friends in 2004 compared to 1974?
What percentage of respondents said that most people can be trusted in 1972? In 2004? Which age group seems to have experienced the greatest change in the level of trust between 1972 and 2004?
Do there appear to be generational differences in volunteerism?
Reading the Results:
- The numbers in each cell of the crosstabulation tables show the percent of the people who fall into the overlapping categories, followed by the actual number of people it represents in this sample. The coloring in the tables demonstrates how the observed numbers in each cell compares to the expected number if there were no association between the two variables. The accompanying bar charts display the patterns visually as well.
- In the comparison of means tables, the top number in each cell represents the mean of the dependent variable for each category of the independent variable(s) and the bottom number is the actual number of people in that category in the sample.
- Weights (mathematical formulas) are often used to adjust the sample proportions, usually by race, sex, or age, to more closely match those of the general population. The analyses in this guide used weights to increase the generalizability of the findings, so the resulting tables are meant to reflect the relationships we would expect to see in the general population.
- The results show the following:
- Among 18-34 year-olds, 57.5% voted in 1972. This number went up to 61.9% in 2004. With 80.4%, 50-64 year-olds had the highest rate of voting in 1972. Respondents 65 and older had the highest rate of voting in 2004 (86.3%).
- The percentage of respondents who never attend religious services increased from 9.3% in 1972 to 15.4% in 2004, while the percentage of respondents who attend once a week or more decreased from 35.3% to 27.5%. Religious participation remains highest among older respondents but appears to have declined for all other age groups.
- The average number of memberships in groups or organizations seems to have dropped for all age groups between 1974 and 2004. The decline is the greatest for 35-49 year-olds (from 2.37 in 1974, to 1.78 in 2004).
- There is very little difference in how much people socialized with relatives and friends in 2004 compared to 1974. However 1974 respondents were more likely to socialize with their neighbors than their 2004 counterparts. 30.9% said "once a month" (26.8% in 2004), another 29.9% said "once a week" (20.9% in 2004), and 21.3 % said "never" (compared to 27.6% in 2004).
- Overall, the level of trust dropped from 46.3% in 1972 to 35.7% in 2004. The change is most pronounced among 18-34 year-olds, where 41.4% in 1972, but only 25.6% in 2004 said that most people can be trusted.
- Volunteerism is lowest for the 18-34 age group, with 6.9% volunteering once a week, and 13.6% once a month, compared to 15.5% and about 24.2% for people age 50 and over.
The goal of this exercise was to explore time and generational trends for measures of social capital over the 1972-2004 period. Taken together, the results lend support to Robert Putnam's argument. With the exception of voting and some forms of sociability, social capital decreased between 1972 and 2004. Religious participation, group memberships, informal neighborhood networks, and mutual trust all indicate a decline in social capital over the period. The results also show generational differences, with older generations having higher levels of social capital.
Further research might explore each dimension of social capital in more detail (for example, have memberships in all groups/organizations declined to the same extent? Why might some show more decline than others?), as well as the consequences of these changes for both bonding and bridging social capital and their implications for society.
We have compiled a list of references that might be useful to instructors and students wishing to further explore this topic. All were chosen because they relate to the topic of study, whether or not they use the specific dataset that was used in this exercise. Some relate directly to the concepts as defined by the exercise, others explore the topic more broadly either conceptually or empirically. For even more resources, try a key word search in the ICPSR Bibliography!
CITATION: Inter-university Consortium for Political and Social Research. Social Capital Over Time and Across Generations: A Data-Driven Learning Guide. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2009-04-16. Doi:10.3886/socialcapital
This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.