Validation of self-reports for use in contact research

The aim of this thesis was to investigate whether self-report measures of contact are valid for use in research testing the ‘contact hypothesis’. The vast majority of contact research has relied on the assumed validity of self-report methods of data collection (Pettigrew & Tropp, 2006), even though the potential weaknesses of self-report methodology generally have been well documented. This reliance is necessary, as self-reports remain the only practical method so far developed of measuring certain of the facilitating conditions developed by Allport (1954/1979), and particularly of direct and indirect cross-group friendship (Pettigrew, 1998; Wright, Aron, McLaughlin-Volpe, & Ropp, 1997). However, if self-reports are not a valid method for measuring contact, the derived implications of a large portion of the research effort are potentially flawed. This thesis attempted to address this important oversight, using a variety of methods to investigate whether the use of self-reports in future research on intergroup contact is appropriate. Studies 1 and 2 demonstrated that self-reports of contact show considerable resistance to context effects, particularly in comparison with self-reports of the more subjective construct of attitudes. Studies 3-5 demonstrated that self-reports of contact agree with the observer-reports of a single observer who knows the target intimately – the spouse or parent. Studies 6 and 7 replicate this agreement through the consensually supported observer-reports of three close friends of the target, thereby reducing any variance due to individual response biases. Finally, studies 8 and 9 demonstrate the concurrent criterion-related validity of self-reports of contact, in that they are able to predict contact on a very large online network called Facebook, on which real-world rather than purely online friendships are primarily represented. These findings offer considerable support for the validity of self-reports as a suitable method for measuring contact. As self-reports remain the only method which has thus far proven suitable for the measurement of those aspects of contact which are essential for exploration of the contact hypothesis, this thesis presents a very heartening and optimistic conclusion and supports the continued use of self-reports in contact research.