LITMIR.BIZ

      A Chinese census taker collecting personal information. Self-reports are often accurate and can be very useful, particularly when the information sought is noncontroversial and the respondent is unlikely to wish to hide his answers from the researcher.

      AP Photo/Liu mingxiang - Imaginechina.

      As surprising and counterintuitive as self-perception theory may be, it has wide-ranging application and has received a great deal of empirical support. Note though that Bem (1967) did not argue that we rely exclusively on observing our own behavior when determining our attitudes, feelings, and personality traits, but that we do so only when are uncertain about them. Earlier in this chapter, we asked how it is that we come to know ourselves. So far we have discussed two ways in which people can learn about their self-concept: introspection and self-perception. But there is alternative approach: looking to other people for feedback about how we are doing. We will elaborate on this method in the section on self-evaluation and self-enhancement below.

      Self-Perception Theory: Idea that peoples sometimes infer their own attitudes in the same way that a third party might infer their attitudes: by watching their behavior

      Intrinsic Motivation: Desire to engage in a behavior simply because it is interesting or enjoyable

      Extrinsic Motivation: Desire to perform a behavior as a result of external rewards or pressures

      Overjustification Effect: When one’s intrinsic motivation—such as enjoyment experienced by simply enacting the behavior—is weakened by the presence of extrinsic motivation

      Facial Feedback Hypothesis: Idea that people infer their feelings from their facial expressions

      Think Again!

      Doing Research: Questioning Self-Reports And Surveys

      The most commonly used method for obtaining data in social psychology is the self-report, which is an individual’s conscious response to a question or situation. A direct question asking about your attitude toward your college is a self-report. Self-reports are often accurate and can be very useful, particularly when the information sought is noncontroversial and the respondent is unlikely to wish to hide his answers from the researcher. For example, self-reported gender is likely to be accurate, whereas attitudes toward members of another race is less likely to be. Researchers use self-reports to obtain a variety of data, including people’s opinions, feelings, behaviors, and physiological experiences (e.g., hunger or pain). Three advantages of self-report measures are that they are relatively easy to construct, are inexpensive, and can be utilized in a variety of research methods, including surveys, interviews, and many experiments.

      There are also several disadvantages to using self-reports, whether in surveys, interviews, or experiments. One is that self-reports may not always provide accurate information, either as a result of participant psychology or the construction of the questions (Krosnick, Lavarakas, & Kim, 2014; Schwarz, 2007b). Researchers have extensively investigated the psychology of self-reports and have identified a number of undesirable response effects that can undermine the accuracy of the answers. Response effects are unintended variations in question responses that stem from procedural aspects or features of the survey instrument, such as the wording of a question or the order of the questions (Heintzelman, Trent, & King, 2015; Helmes, Holden, & Ziegler, 2015; Schwarz, 1999; Tourangeau, Rips, & Rasinski, 2000). These survey features may affect how participants understand questions, the role of memory and judgment in generating potential responses, and how participants report their answers (Schwarz, 1999, 2007a; Tourangeau et al., 2000).

      One response effect is the acquiescence bias, which is the tendency to agree with or say “yes” to questions (Savalei & Falk, 2014). This is of particular concern when conducting cross-cultural research, because clear culture-based differences have been found (Riemer & Shavitt, 2011). For instance, East Asians are more likely to agree with questions than are certain other groups (Grimm & Church, 1999). Another type of response effect is the extremity bias, wherein respondents provide answers that are at the extremes of the response options (Levashina, Weekley, Roulin, & Hauck, 2014). For instance, on a scale ranging from “very unlikely” to “very likely,” a person exhibiting this bias will tend to chose one of the endpoints of the scale rather than the more moderate options, such as “likely” or “unlikely.” As with the acquiescence bias, the extremity bias varies across cultures: African Americans and Hispanics are more likely to demonstrate extreme responding than European Americans, whereas East Asians are less likely (Bachman & O’Malley, 1984; Chen, Lee, & Stevenson, 1995; Hui & Triandis, 1989). Question wording can also bias or distort the answers (Schwarz, 2007a). A recent example is a survey reported in the New York Times in which 20% of respondents said that the U.S. government spent too little on “welfare,” but 65% indicated that it spent too little on “assistance to the poor” (See Figure 4.3) (Schneiderman, 2008).

      Another response bias may result from the context in which the question is asked. For instance, participants in one study provided different explanations for a mass murder depending on whether the letterhead at the top of the survey was a fictional “Institute for Social Research” or “Institute for Personality Research.” In the “social” condition respondents tended to focus on external or environmental causes, in contrast to the internal or personality factors emphasized in the “personality” condition (Norenzayan & Schwarz, 1999). Such context effects—variations in responding because of survey features encountered prior to answering a question—are also seen when the presence or wording of earlier questions alters responses to later ones (Schwarz, 1999; Toepoel & Couper, 2011; Weijters, Geuens, & Baumgartner, 2013). Researchers need to consider these biases when designing surveys and other self-reports (Schwarz, 1999, 2007a). One additional weakness in self-reports was discussed above: We often do not know how we know what we know; that is, we cannot report on many of our mental processes, even if we can report on the outcome (Nisbett & Wilson, 1977).