Open Collections

The nature of constructs in social psychological research

University of British Columbia

A fundamental assumption in social psychological research, that the construct underlying conditions in an experiment is stable and varies only in quantity across conditions, is explored. It is argued that instead the result of the manipulation of the independent variable is a qualitative difference between the relevant psychological constructs. A distinction is made between theories of empirical constructs and those which posit unobservable, hypothetical constructs. The assumption of construct stability relates only to theories of relationships between hypothetical constructs. The lack of direct access to the identity of geno-typic processes underlying observable behaviour or manipulations allows the possibility of construct instability. Methods of improving access discussed centre on increasing the number and diversity of observations by repeating the operationalization and measurement of variables in ways different but theoretically tied to the construct of interest. The Q-sort method of correlating behaviour with items descriptive of personality is proposed as a method of examining the identity of constructs and thus of assessing the hypothesis that conditions are qualitatively different. This method is applied to an examination of the relationship between incentive and choice and latency behaviour in the Prisoner's Dilemma game. A significant difference between conditions was found for choice behaviour but not for latency. Q-item—behaviour correlates, however, were different in each condition and for each behaviour, supporting the hypothesis that conditions differ in some meaningful way. A stronger test of the hypothesis involved the construction of templates describing the ideal-personality most likely to exhibit the designated behaviour. These templates were able to predict behaviour within conditions and, in agreement with the hypothesis, no predictability was found across conditions. However, within-condition validation coefficients were very low, such that the across conditions results were uninterpretable. It was concluded that the Q-sort method, while attractive, is constrained in its usefulness and ability to test the hypothesis. Discussion included examination of the link between qualitative differences and person by situation interaction. It was recommended that social psychology improve its recognition of methodological assumptions made and of how these and theories of social behaviour are interrelated.

Text

2010-03-29

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/23012

Psychology

University of British Columbia

Graduate