Bias versus Bias: Harnessing Hindsight to Reveal Paranormal Belief Change beyond Demand Characteristics
Kane, Michael J., Core, Tammy J., Hunt, R. Reed, Psychonomic Bulletin & Review
Psychological change is difficult to assess, in part because self-reported beliefs and attitudes may be biased or distorted. The present study probed belief change, in an educational context, by using the hindsight bias to counter another bias that generally plagues assessment of subjective change. Although research has indicated that skepticism courses reduce paranormal beliefs, those findings may reflect demand characteristics (biases toward desired, skeptical responses). Our hindsight-bias procedure circumvented demand by asking students, following semester-long skepticism (and control) courses, to recall their precourse levels of paranormal belief. People typically remember themselves as previously thinking, believing, and acting as they do now, so current skepticism should provoke false recollections of previous skepticism. Given true belief change, therefore, skepticism students should have remembered themselves as having been more skeptical than they were. They did, at least about paranormal topics that were covered most extensively in the course. Our findings thus show hindsight to be useful in evaluating cognitive change beyond demand characteristics.
Psychology and its allied disciplines have long struggled to accurately assess change, whether that ostensible change results from maturation, senescence, laboratory experimental manipulations, psychotherapeutic techniques, community interventions, or educational programs (see, e.g., Cronbach & Furby, 1970; Hertzog & Nesselroade, 2003; Lord, 1956, 1967; Nesselroade, Stigler, & Baltes, 1980; Rubin, 1974). Of course, in contexts in which the desired change is entirely subjective-as is the case with attitudes, beliefs, cognitions, evaluations, or emotional states-the risks of misidentifying or misinterpreting change will only increase, since subjects' self-reports may be biased, distorted, or erroneous (see, e.g., Conway & Ross, 1984; Festinger, 1957; Greenwald, Spangenberg, Pratkanis, & Eskenazi, 1991; Hoogstraten, 1979; Kirsch, 1985; Lewinsohn & Rosenbaum, 1987; Loftus, 1979; H. Markus & Kunda, 1986; Orne, 1962; Wilson & Brekke, 1994). Researchers must therefore develop statistical and methodological tools to help discriminate real from illusory change. The present study demonstrated a seemingly paradoxical approach, whereby a powerful cognitive bias was strategically deployed as a means to counter another, especially formidable bias that plagues assessment of subjective change-here, in the context of an educational intervention designed to affect undergraduates' beliefs.
Education and Paranormal Belief
Most Americans, even many with advanced educational degrees, hold paranormal, superstitious, or pseudoscientific beliefs, such as belief in extrasensory perception (ESP), alien abduction, or creationism (Moore, 2005; Newport & Strausberg, 2001; Rice, 2003). Indeed, neither general science knowledge nor a scientific major consistently hinders such beliefs (Aarnio & Lindeman, 2005; Goode, 2002). Limited research suggests, however, that university courses that directly and skeptically examine paranormal and pseudoscientific phenomena may reduce students' beliefs in them, at least in the short term.
Unfortunately, this literature is limited in both size and methodological rigor. Of the dozen or so published studies on educational interventions and paranormal belief, only three included control groups (students in unrelated courses; see Dougherty, 2004; Gray, 1985; Morier & Keeports, 1994);1 furthermore, some studies included only postcourse evaluations with no pre-to-post comparisons (Broch, 2000; Calvin, 2009), and most studies asked students to report their beliefs without anonymity (Banziger, 1983; Emme, 1940; Gilliland, 1930; Jones & Zusne, 1981; McBurney, 1976; Swords, 1990; Tobacyk, 1983). Most reports of education-induced paranormal belief change may thus have derived simply from passing time (or other external influences) or from students' reaction to their identifiabilty. …