For example, a natural experiment at Harvard University assigned first-year students from different races as roommates. Among the findings was that breakups among the roommates were more likely when an East Asian student lived with two white students (Chakravarti, Menon, and Winship 2014).
Field experiments. In some natural situations, researchers are able to exert at least some control over who participates and what happens during experiments (Bertrand and Mullainathan 2004; Pager and Western 2012). These are called field experiments. One of the most famous studies in the history of sociology is the “Robbers Cave” field experiment (Sherif et al. [1954] 1961), so called because it took place in Robbers Cave State Park in Oklahoma. The researchers controlled important aspects of what took place at the site. For example, they were able to assign the 22 boys in the study into two groups, called the Rattlers and the Eagles. The researchers were also able to create various situations that led to rivalry, bickering, and hostility between the groups. At the end of the experiment, they had each group rate the other: 53 percent of ratings of the Eagles were unfavorable, while nearly 77 percent of ratings of the Rattlers were unfavorable. Later, the researchers introduced conditions they hoped would reduce bad feelings and friction between the groups. In fact, greater harmony between the groups was created by having them work together on tasks such as securing needed water and paying collectively and equitably for a movie that everyone wanted to see. By the end of the latter part of the experiment, just 5 percent of the ratings of the Eagles were unfavorable, and unfavorable ratings of the Rattlers had dropped to 23 percent.
A more recent field experiment in Sweden dealt, in part, with hiring discrimination against ethnic groups (Bursell 2014). Pairs of equally qualified applicants were sent to interview for open jobs. Those with Arabic or North African names were less likely to be called back for additional consideration.
Secondary Data Analysis
All of the methods discussed thus far involve the collection of new and original data, but many sociologists engage in secondary data analysis, in which they reanalyze data collected by others. Secondary analysis can involve a wide variety of different types of data, from censuses and other surveys to historical records and old transcripts of interviews and focus groups. Until recently, obtaining and using some of these secondary data sets was laborious and time-consuming. Today, however, thousands of data sets are available online, and they can be accessed with a few keystrokes. A number of websites provide both the data sets and statistical software for looking at them in different ways (Quinton 2018).
Secondary data analysis very often involves statistical analysis of government surveys and census data. For example, researchers studying multiracial identity used data from the 2000 and 2010 censuses and found that individuals who identified as two or more races increased from 6.8 million in 2000 to 9 million in 2010 (Jones and Bullock 2013). It is not unusual for one body of data to lead to hundreds of secondary analyses. For example, the World Values Survey (WVS; www.worldvaluessurvey.org/wvs.jsp) is conducted in nearly 100 countries containing almost 90 percent of the world’s population. Its seven waves have been used to produce more than 1,000 research publications in more than 20 languages. Some of this research has used the WVS to examine what social, cultural, and economic factors contribute to an individual’s happiness. Figure 2.5 shows a “happiness map” depicting levels of happiness among the citizens of various countries included in the surveys from 2010 to 2018.
Historical-Comparative Method
The goal of historical-comparative research is to contrast how different historical events and conditions in various societies have led to different societal outcomes. The historical component involves the study of the history of societies as well as of their major components, such as the state, religious system, and economy. The addition of the comparative element, comparing the histories of two or more societies, or of components of societies, makes this method more distinctively sociological.
Figure 2.5 World Happiness Ranking of 158 Countries
Source: John F. Helliwell, Richard Layard, and Jeffrey Sachs, eds. 2018. World Happiness Report 2018. New York: Sustainable Development Solutions Network.
Historians go into much more detail, and collect far more original historical data, than do sociologists. In contrast, sociologists are much more interested in generalizing about society than are historians. Weber did historical comparative research in analyzing the relationship between world religions and the development of capitalism. More recent instances of historical-comparative research have covered a wide range of issues, but one of the most popular topics has been examining the relationship between the state and war (Skocpol 1979; Rodriguez-Franco 2016).
Some scholars have combined the use of other methods with historical-comparative analysis to generate important theoretical insights about more contemporary issues. For example, Piketty (2014) examined a variety of statistical data to uncover the historical changes of income and wealth inequality in Europe and the United States. He found that inequality has been produced and reproduced over time due to the concentration of wealth in the upper class. The concentration of wealth was quite high in Europe during the eighteenth and nineteenth centuries and in the United States in the decades before the Great Depression but leveled off during and after World War II due to inflation, higher taxes, and the policies of the modern welfare state. However, near the end of the twentieth century the concentration of wealth (and income) started to increase dramatically when inherited wealth, in particular, began to grow faster than economic output and income (Piketty 2014). If we aspire to a more egalitarian society, Piketty suggests that states implement a global tax on wealth.
Content Analysis
Another type of secondary analysis, called content analysis, relies on the systematic and objective analysis of the content of cultural artifacts in print, visual, audio, and digital media, including photographs, movies, advertisements, speeches, and newspaper articles (Wolff 2007). The goal is to use qualitative and especially quantitative methods to understand the content of messages. In one well-known study, Herbert Gans (1979) did a quantitative and qualitative content analysis of news on television and in newsmagazines to identify patterns in the reporting of news. For example, he found that well-known people were dealt with much more frequently than were unknowns. Among non-war-related stories, government conflicts and disagreements were more likely to be dealt with than were government decisions.
Gans’s content analysis took as its focus the overt content of the news, but it is also possible to use content analysis to analyze other issues, such as gender inequality. For example, researchers performed a content analysis of 1,245 characters on 89 prime-time television programs, finding that though some gender stereotypes have declined, others, such as dominant men and sexually provocative women, persist (Sink and Mastro 2016). Content analysis is moving beyond traditional media and is now also being conducted on social media. For example, Pilkington and Rominov (2017) researched fathers’ worries during their partners’ pregnancies by analyzing the content of their posts on Reddit. They found that most fathers-to-be are worried about infant well-being and the potential for prenatal loss.
Issues in Social Research
The research conducted by sociologists raises a number of issues of great importance. These include the reliability and validity of findings and the ethics involved in the research process.
Reliability and Validity
A key issue with sociological data relates to one’s ability to trust the findings. Reliability involves the degree to which a given question, or another kind of measure, produces the same results time after time. In other words, would the same question asked one day get the same response from the participants or the same measurement on the