The effect of stereotype threat on older people’s clinical cognitive outcomes: investigating the moderating role of dementia worry

Abstract

Objective: Numerous studies have shown that stereotype threat (ST) reduces older people’s cognitive performance, but few have studied its impact on clinical cognitive outcomes. Our study was designed to further examine the impact of ST on the clinical assessment of older subjects’ cognitive functioning, as well as the moderating role of fear of Alzheimer’s Disease (AD) (or ‘dementia worry’). Method: Seventy-two neurologically normal (MMSE > 26) participants aged between 59 and 70 completed a set of neuropsychological tasks in either an ST or a positive condition (condition in which negative stereotypes were invalidated). Results: Regression-based path analyses showed that only participants who expressed moderate or high fear of AD underperformed on executive tasks in the ST condition compared to their counterparts in the positive condition. Moreover, in the ST condition, participants’ performance on executive tasks was more impaired (relative to normative data) than in the positive condition. However, ST had no effect on memory and attention performance. Discussion: Our results showed that ST can cause older people to perform at pathological levels on executive tasks. Results highlight the need for clinicians to be cautious when conducting neuropsychological assessments of older people who express high levels of dementia worry.

Keywords:

• Cognitive aging
• stereotype threat
• dementia worry
• neuropsychological assessment

Acknowledgments

We thank Professor Thomas M. Hess for providing us with the news articles he used in his stereotype threat studies with older people.

Notes

1. It is well known that fluency tasks are multi-determined. Indeed, it is recognized that they recruit memory processes, executive functioning, and vocabulary, among other things. Because fluency tasks recruit memory processes (Rende et al., 2002; Unsworth et al., 2012) and since our fluency measure was correlated with all our memory measures, but with only one of the three executive measures (r = −.28, p = .018 for the TMT errors; ps > .132 for Stroop errors and divided attention errors), we chose to include fluency measures in the composite memory score.

2. The demographic questions included a ‘yes or no’ question regarding a past medical consultation for memory and attention concerns. If they responded negatively to this question, participants rated on a 7-point Likert scale ranging from ‘not at all’ to ‘totally’ to what extent they intended to have a cognitive assessment. These items were administered to assess the intention to undertake a cognitive assessment. Indeed, we wanted to assess not only attitudes (dementia worry) but also intentions, with the hypothesis that the interaction between condition and dementia (moderation effect) would be even stronger for older persons who already intended to undertake a cognitive assessment. However, because too many participants responded ‘not at all,’ we could not include this variable in our statistical analyses. Note that controlling for this variable does not change the pattern of results presented below.

3. When referring to the normative mean, we refer to the published normative data on the clinical cognitive tasks we used. Hence, these normative data were collected on other samples of healthy older persons.