Misrepresenting Health Conditions Through Fabrication and Exaggeration: An Adaptation and Replication of the False Alarm Effect

ABSTRACT

This article reports on a series of studies of the false alarm effect (FAE), suggesting that individuals’ perceptions that relational partners are fabricating and exaggerating their health conditions are negatively associated with perceptions of health condition credibility, which in turn are associated with decreases in individuals’ protective behaviors and attitudes. In Study One (N = 216), we took a mixed-methods approach to test an initial model predicting that health condition credibility mediates associations between individuals’ perceptions that partners are fabricating and exaggerating and the extent to which individuals provide support, seek information about the condition, feel efficacious in their ability to assist partners, and believe that the condition is serious. We also analyzed open-ended responses to parse the source(s) of credibility lost when individuals believe partners are fabricating and exaggerating their health conditions. We found that they express doubt not only about the credibility of the health condition itself, but also about their partner’s credibility in terms of trustworthiness. We then refined our conceptual model to account for these two sources of credibility and tested it with a path model in a second study utilizing a nationally representative sample (N = 508). Results supported our hypotheses. We discuss the implications of this research for how people present themselves as ill in personal relationships, and what happens when these presentations are unconvincing.

Notes

¹ We interpret this modification via Kline (2011) to mean that our model would be better specified as partially recursive because it has unidirectional effects but correlated disturbances that act as feedback loops among the outcome variables. For example, it is likely that information seeking influences providing support and vice versa.

² A chi-square difference test comparing the saturated (with values of zero) and nested models would yield the same results as the nested model, indicating that removing paths did not significantly worsen the fit of the model.

³ In particular, there is an inconsistent mediation (see MacKinnon et al., 2000) in which health condition credibility acted as a suppressor variable. The introduction of the mediated effect of health condition credibility may have slightly magnified the association between fabricating and exaggerating and information seeking. It could be that health condition credibility is related to both fabricating and exaggerating and information seeking outside the model proposed. It may also be that the association between these two variables is actually higher than was observed.

⁴ It is possible that trustworthiness is driving how individuals interpret partners’ health condition presentations. We tested an alternative model in which trustworthiness predicts perceptions of fabricating and exaggerating rather than serving as an outcome variable. We correlated the error terms for the outcome variables as we did for the other models to make the models as consistent as possible for comparison. The initial model had poor fit, and modification indices suggested adding paths between trustworthiness and several variables in the model. Though this model did have better fit when these paths were added, there was evidence of suppression, which makes some of the effects in the model uninterpretable. Therefore, we elected to retain the model that was more parsimonious and better reflected the FAE framework of the study.