Desktop version

Home arrow Mathematics

  • Increase font
  • Decrease font

<<   CONTENTS   >>

Social Desirability

A major concern for researchers who hire local interviewers is social desirability effects. Error from social desirability occurs when a participant alters their true response in order to present a more desirable view of themselves to those collecting data. An early example comes from W. E. B. Du Bois' study of the seventh ward of Philadelphia in the late 19th century (Du Bois 1899). The ability to read was a point of pride among residents and Du Bois was concerned that illiterate participants might say they are literate in order to appear more socially desirable.

Participants from Du Bois' study and others who engage in socially desirable responses are often thought to edit their answer in the latter phase of the cognitive response process (Sudman, Bradburn, and Schwarz 1996; Tourangeau, Rips, and Rasinski 2000). In this last phase, a participant translates the response they have in mind to fit the question and decides if they need to make any edits to their response before they finally answer. Here, participants think about how their answer will be perceived by the interviewer and whether they should change their answer to avoid a negative response from the interviewer (real or perceived). A major focus of social desirability research is therefore on circumstances where a participant decides to edit their answer to be more socially desirable.

There are two general views on social desirability. The first is that social desirability may be a characteristic of a person. This view proposes that there are some people who wish to be viewed by others in a favorable light in almost all of their actions with others (Johnson and Van de Vijver 2003; Tourangeau and Yan 2007). Assuming the participant has an idea (real or perceived) about what the other wishes to hear, they will shift their responses accordingly. A different perspective suggests that social desirability is specific to a given question (Johnson and Van de Vijver 2003). This view suggests that a socially desirable response is provoked by the property of the question itself, triggering a socially desirable response, and not a personality trait.

Socially desirable answers are more likely when questions focus upon sensitive issues and may result in under- or overreporting. Underreported responses have been detected when questions have been asked about illegal and deviant substance use (Davis, Thake, and Vilhena 2010; Johnson and Van de Vijver 2003; Krumpal 2013; van de Mortel 2008) and questions about victimization (Bell and Naugle 2007; Krumpal 2013; Sugarman and Hotaling 1997; van de Mortel 2008). Overreporting is more likely with desirable responses such as voter participation or seat belt use (Johnson and Van de Vijver 2003; Krumpal 2013). However, the sensitivity of a question may also vary between participants depending upon their history and current context (Kreuter, Presser, and Tourangeau 2008).

A major potential source of social desirability in an interview is a lack of privacy. Whenever a participant perceives that an interview is not private, there may be higher potential for an edited response, particularly with sensitive questions. A common threat to privacy is the presence of a third party during an interview. In these settings participants have often been found to edit their responses when asked about sensitive questions, but less so when asked about neutral questions (Aquilino 1993; Gfroerer, Wright, and Kopstein 1997; Hartmann 1995; Mneimneh, et al. 2015; Turner and Martin 1984). Where the interview takes place also affects privacy. Gfroerer, Wright, and Kopstein (1997) speculate that differences between national surveys on drug use among minors may in part be due to where the interviews take place, one in the home and one in the classroom. In this example, the classroom is presumed to be more private than the home as there may be parents present or nearby during an at-home interview.

A third area where privacy may affect social desirability in an interview may occur when the interviewer and the participant have a pre-existing relationship that is external to the interview context. This situation is exceptionally rare in standardized interviewing and most interview protocols assume that the interviewer and the participant are complete strangers (Fowler and Mangione 1990; Hyman 1954). In a CBPR context, with local interviewers hired from small communities, a participant knowing the interviewer becomes more likely. In such an interview, the participant has to consider what they are comfortable telling an interviewer who is also a member of their own community and with whom they will continue to have contact after the interview is complete. This goes beyond considering risk of divulgence to an outside entity and to the real consideration that a question response may affect their day to day life.

Research Questions

In this chapter, we test if the primary outcomes measured through interviewer-administered surveys are associated with three measurements of privacy in a family-based substance use prevention program for American Indian youths. We measure the presence of a third party during the interview, whether the interview took place at home or somewhere else, and if the interviewer knew the participant before the interview. We use data from the baseline wave of the program to test for associations between interview privacy and differences in mental health, substance use, cultural participation, and cultural discrimination among 8- to 10-year-olds. We expect that social desirability effects might decrease reports of externalizing and internalizing behaviors, reduce reports of substance use, and increase reports of cultural participation.

<<   CONTENTS   >>

Related topics