Abstract
Perceived support availability (PSA), a general belief about the likelihood that social support will be available when needed, is associated with numerous processes and outcomes of supportive communication. Currently, however, there is little understanding of the factors that contribute to this belief. Numerous studies have reported gender differences in PSA, with women generally indicating that they see support as more available than do men; in turn, gender differences in PSA have been cited to explain gender differences in the production and interpretation of supportive messages. In an effort to explain gender differences in PSA and, more broadly, understand the social-cognitive factors that contribute to individual differences in PSA, this article proposes and reports a test of a theoretical model that treats PSA as the outcome of the availability and accessibility of cognitive schemata for construing social situations. Participants (150 men and 271 women) completed instruments providing assessments of PSA, construct availability (cognitive complexity), and construct accessibility (expressive and instrumental orientations). Bootstrap procedures for the simultaneous assessment of multiple mediators found that construct accessibility generally was a stronger mediator of gender differences in PSA than construct availability. However, a cooperative suppression effect was found for 1 index of construct accessibility, complicating the interpretation of mediation effects.
Acknowledgments
All authors were members of a special research team that focused on social support processes and gender differences.
Notes
Note. N = 421. For participant gender, men = 0 and women = 1. PSA = perceived support availability.
*p < .001.
Note. Gender of participant was coded as men = 0, women = 1. Percentages refer to the percentage of the total effect explained by a particular variable or effect.
*p < .05. **p < .01. ***p < .001. †p < .10.
The bootstrapping procedure simultaneously estimates the effect for multiple mediators, is not shackled by assumptions of normal sampling distributions, and tends to provide more accurate Type 1 error rates than more traditional procedures (MacKinnon, Lockwood, & Williams, Citation2004).
Because of their negative association, each of the independent variables removes or “suppresses” the variance it shares with the other independent variable, and this has the effect of increasing the magnitude of its association with the dependent variable above its validity coefficient (i.e., its zero-order association) with the dependent variable (see Cohen & Cohen, Citation1983). Moreover, although instrumentality was positively associated with perceived support availability (PSA), the indirect effect for instrumentality (i.e., its mediation of the effect of gender on PSA) had a negative sign, in effect subtracting variance from the collective indirect effect attributable to the three mediators (see Table ).