Stereotype Threat As a Trigger of Mind-Wandering in Older Adults

Supplemental Materials

Stereotype Threat as a Trigger of Mind-Wandering in Older Adults

By Megan L. Jordano and Dayna R. Touron, 2015, Psychology and Aging

Additional measures. Several single-item, Likert-scale post-task questions were included to measure factors that might influence mind-wandering and Operation Span (OSPAN) letter recall performance. We found that perceived overall task difficulty (F (3,110) = 1.992, p = .119) and perceived difficulty of the letter recall portion of the OSPAN (F (3,110) = 1.106, p = .350) did not differ by condition. Although older adults (OAs) did worse on average than younger adults (YAs) on the OSPAN, they did not report finding the OSPAN or its subcomponents more difficult than did YAs. Self-rated effort in completing the OSPAN likewise did not differ by condition (F (3,110) = 2.231, p = .089). As shown in Table 2, greater recall and overall difficulty were related to higher proportions of task-related interference (TRI), whereas greater effort was related to lower proportions of task-unrelated thoughts (TUTs).

We also included a single-item Likert-scale question asking participants if they believed there was a stereotype that OAs have worse memory ability than do YAs, and found that there was an effect of condition of this response, F (3,110) = 5.186, p = .002. ST activated OAs had higher scores for this question than did YAs and OA controls (all p values < .05 for t-test focused comparisons), meaning that they were more likely to believe that the negative stereotype existed. ST relief OAs were also more likely to believe the negative stereotype existed than did YAs and OA controls (all p values < .05 for t-test focused comparisons). As shown in Table 2, greater stereotype awareness was related to lower proportions of TUTs and to higher proportions of TRI.

Because the effect of stereotype threat (ST) on task performance has been proposed to be mediated by the importance of good performance in the stereotype domain (Hess et al., 2013) and anxiety during completion of the stereotype task (Chasteen, 2005; Abrams et al., 2005), we included the Anxiety and Achievement subscales of the Metamemory in Adulthood (MIA) Questionnaire (Dixon & Hultsch, 1984). There was no effect of condition on MIA Achievement scores (F (3,113) = .212, p =.888) but there was on effect of condition on MIA Anxiety scores, F (3,113) = 3.716, p = .014. As expected, OAs activated for stereotype threat had higher scores on the anxiety scale of the MIA than did OA controls and OAs relieved of stereotype threat (all p values < .05 for t-test focused comparisons), indicating that activated OAs reported experiencing greater anxiety while completing everyday memory tasks. YAs also reported significantly higher scores on the anxiety scale of the MIA than did OA controls and ST relief OAs relieved stereotype threat (all p values < .05 for t-test focused comparisons). As shown in Table 2, higher MIA achievement scores were related to lower proportions of TUTs and to higher proportions of TRI, and higher MIA Anxiety scores were related to lower proportions of TUTs.

The Influence of Age, Task difficulty, and Importance of Task Performance on TRI. Different mechanisms have been proposed to explain how stereotype threat affects task performance. In the current study, we conducted exploratory mediational analyses to gain a better understanding of which factors influence proportion of probe-caught TRI. We also wished to gain a better understanding of how mind-wandering (particularly TRI) may mediate the relationship between factors such as task engagement, perceived task difficulty, and mood, and performance on the stereotyped task. To do this, we conducted a regression predicting letter recall accuracy, as well as an additional mediational analyses using bootstrapping.

When looking at the entire sample of participants, the following factors were significantly correlated with TRI and with letter recall accuracy during the OSPAN: chronological age, perceived overall OSPAN difficulty, perceived letter recall difficulty, self-rated task effort, self-rated stress, self-rated focus on the letter recall portion of the OSPAN, importance of good memory performance (as measured by the Achievement scale of the MIA), memory anxiety (as measured by the Anxiety scale of the MIA), and belief in the existence of a negative aging and memory stereotype (see Table 2). When excluding YA participants and only looking at OA participants, the only factor significantly correlated with TRI variables and with letter recall accuracy during the OSPAN was perceived letter recall difficulty (see Table 3).

To better understand which factors are most important in predicting proportion of probe-caught TRI, we conducted three regression analyses using our entire sample of participants in which chronological age, OSPAN difficulty, letter recall difficulty, task effort, stress, focus on the letter recall portion of the OSPAN, MIA Achievement scores, and MIA anxiety scores were simultaneously entered to predict proportion of probe-caught TRI, proportion of probe-caught proactive TRI, and proportion of probe-caught reactive TRI (see Table 4). Chronological age ( = .033, p = .012) and participants’ self-rated importance of good memory performance as measured by MIA Achievement scores ( = .009, p = .031) both emerged as significant predictors of overall proportion of probe-caught TRI. Chronological age also emerged as a significant predictor of reactive TRI about task performance evaluation ( = .001, p = .002). Conversely, for proportion of proactive TRI about task-strategy or task approach, participants’ self-rated letter recall difficulty was the only factor that emerged as a significant predictor ( = .002, p = .011). These regression analyses indicate that different participant characteristics predict different types of TRI experiences. Wihtin our older adult sample, self-rated letter recall difficulty was a significant predictor of mean proportion of overall probe-caught TRI ( = .051, p = .002).

TRI as a Mediator in the Relationships between Stress, Task Difficulty, and Task Performance. In addition to the regression analyses, we conducted exploratory mediational analyses to determine whether probe-caught TRI mediates the relationship between the aforementioned factors and letter recall accuracy. Using our entire sample of participants, we examined overall TRI as a mediator for the stress, OSPAN difficulty, and recall difficulty factors, while the fourth examined proactive TRI as a mediator for recall difficulty. The results of these analyses are displayed in Figure 1.

First, using 6000 resamples, we tested for indirect effects of overall probe-caught TRI on letter recall accuracy. Bootstrapping was used to calculate the confidence interval around the indirect effect. The 95% confidence interval (CI) was [-2.236, -.167] with zero falling outside of this confidence interval, indicating that probe-caught TRI mediates the relationship between stress and letter recall accuracy, p = .03. Probe-caught TRI was also found to mediate the relationship between perceived OSPAN task difficulty and letter recall accuracy and between perceived letter recall difficulty and letter recall accuracy. The 95% CIs were [-3.20, -.233] and [-3.212, -.326], respectively, with zero falling outside of these intervals (p = .021 and p = .047, respectively). Finally, we tested for an indirect effect of proactive TRI on recall accuracy, finding that proactive TRI likewise mediated the relationship between perceived letter recall difficulty and recall accuracy (95% CI [-2.367, -.071], p = .025).

We were interested in whether these patterns would hold for older adults. Accordingly, excluding the YA sample and only looking at our OA sample, we conducted an additional mediational analysis to determine whether probe-caught TRI mediates the relationship between older adults’ self-rated letter recall difficulty and letter recall accuracy. We did not test the mediation models for the relationships between stress and overall difficulty ratings with TRI and recall accuracy, due to having obtained nonsignificant correlations between these variables in the OA sample. The result of our analysis is displayed in Figure 2. Using 6000 resamples, we tested for indirect effects of overall probe-caught TRI on letter recall accuracy. Bootstrapping was used to calculate the confidence interval around the indirect effect. The 95% confidence interval (CI) was [-.177, .002] with zero falling inside of this confidence interval, indicating that probe-caught TRI did not mediate the relationship between self-rated letter recall difficulty and letter recall accuracy, p = .06 in older adults.

The difference in correlation and mediation model patterns when we restrict analysis to the OA sample likely reflects the reduced statistical power of these tests. However, it is also possible that the effects obtained for the full sample are inflated by the low levels of TRI and higher recall accuracy by the young adult group. Given the exploratory nature of these models, the outcomes should be considered tentative and should be further investigated in future research.

Table 1. Experimental task verbal instructions.

Table 2

Overall Correlations for Mind-Wandering Variables and Letter Recall

Note. OT = proportion of probe-caught on-task thoughts. TUT = proportion of probe-caught task-unrelated thoughts. TRI =proportion of probe-caught task-related interference. Proactive TRI = proportion of probe-caught TRI regarding task approach or strategy. Reactive TRI = proportion of probe-caught TRI regarding evaluation of task performance. Recall accuracy = accuracy on the letter recall portion of the OSPAN. Age = Chronological age in years. DSSQ TUT= score on the TUT subscale of the DSSQ (out of 40), with higher scores indicating more TUTs. DSSQ TRI= score on the TRI subscale of the DSSQ (out of 40), with higher scores indicating more TRI. Recall difficulty = self-rated perceived difficulty of the recall portion of the OSPAN (out of 5, where 1=not at all difficulty, 5= very difficult). Overall difficulty = self-rated perceived difficulty of the OSPAN task overall (out of 5, where 1= not at all difficult, 5 = very difficult. Stress = self-rated perceived stress or anxiety during the OSPAN (out of 5, where 1= not at all stressed, 5 = very stressed). Fatigue = how fatiguing the OSPAN was perceived to be (out of 5, where 1= not at all fatiguing, 5 = very fatiguing). Effort = self-rated effort during the OSPAN task (out of 5, where 1= no effort at all, 5 = a lot of effort). Stereotype awareness = self-rated belief that a negative stereotype regarding aging and cognitive decline exists (out of 5, where 1= strongly disagree that there is a stereotype that OAs are inferior to YAs in terms of cognitive ability, 5 = strongly agree that there is a stereotype that OAs are inferior to YAs in terms of cognitive ability). MIA Achievement = Score on the memory achievement scale of the MIA (out of 75), with higher scores indicating more self-rated importance of good memory performance. MIA Anxiety = Score on the memory anxiety scale of the MIA (out of 65) , with higher scores indicating more self-rated anxiety while completing daily task that have a memory component. Positive mood = score on the positive affect scale of the PANAS (out of 50), with higher scores indicating more positive mood. Negative mood = score on the negative affect scale of the PANAS (out of 50), with higher scores indicating more negative mood.

* indicates p < .05. ** indicates p < .001.

Table 3

Correlations for Mind-Wandering Variables and Letter Recall for the Older Adult Sample

Note. OT = proportion of probe-caught on-task thoughts. TUT = proportion of probe-caught task-unrelated thoughts. TRI =proportion of probe-caught task-related interference. Proactive TRI = proportion of probe-caught TRI regarding task approach or strategy. Reactive TRI = proportion of probe-caught TRI regarding evaluation of task performance. Recall accuracy = accuracy on the letter recall portion of the OSPAN. Age = Chronological age in years. DSSQ TUT= score on the TUT subscale of the DSSQ (out of 40), with higher scores indicating more TUTs. DSSQ TRI= score on the TRI subscale of the DSSQ (out of 40), with higher scores indicating more TRI. Recall difficulty = self-rated perceived difficulty of the recall portion of the OSPAN (out of 5, where 1=not at all difficulty, 5= very difficult). Overall difficulty = self-rated perceived difficulty of the OSPAN task overall (out of 5, where 1= not at all difficult, 5 = very difficult. Stress = self-rated perceived stress or anxiety during the OSPAN (out of 5, where 1= not at all stressed, 5 = very stressed). Fatigue = how fatiguing the OSPAN was perceived to be (out of 5, where 1= not at all fatiguing, 5 = very fatiguing). Effort = self-rated effort during the OSPAN task (out of 5, where 1= no effort at all, 5 = a lot of effort). Stereotype awareness = self-rated belief that a negative stereotype regarding aging and cognitive decline exists (out of 5, where 1= strongly disagree that there is a stereotype that OAs are inferior to YAs in terms of cognitive ability, 5 = strongly agree that there is a stereotype that OAs are inferior to YAs in terms of cognitive ability). MIA Achievement = Score on the memory achievement scale of the MIA (out of 75), with higher scores indicating more self-rated importance of good memory performance. MIA Anxiety = Score on the memory anxiety scale of the MIA (out of 65) , with higher scores indicating more self-rated anxiety while completing daily task that have a memory component. Positive mood = score on the positive affect scale of the PANAS (out of 50), with higher scores indicating more positive mood. Negative mood = score on the negative affect scale of the PANAS (out of 50), with higher scores indicating more negative mood.

* indicates p < .05. ** indicates p < .001.

Table 4

Regression models for TRI, Proactive TRI, and Reactive TRI

Note. Age = Chronological age in years. Recall difficulty = self-rated perceived difficulty of the recall portion of the OSPAN (out of 5, where 1=not at all difficulty, 5= very difficult). Overall difficulty = self-rated perceived difficulty of the OSPAN task overall (out of 5, where 1=not at all difficulty, 5= very difficult). Recall focus = self-rated focus on the letter recall portion of the OSPAN (out of 5, where 1= not at all focused, 5 = very much focused). Stress = self-rated perceived stress or anxiety during the OSPAN (out of 5, where 1= not at all stressed, 5 = very stressed). Effort = self-rated effort during the OSPAN task (out of 5, where 1= no effort at all, 5 = a lot of effort). Stereotype awareness = self-rated belief that a negative stereotype regarding aging and cognitive decline exists (out of 5, where 1= strongly disagree that there is a stereotype that OAs are inferior to YAs in terms of cognitive ability, 5 = strongly agree that there is a stereotype that OAs are inferior to YAs in terms of cognitive ability). MIA Achievement = Score on the memory achievement scale of the MIA (out of 75), with higher scores indicating more self-rated importance of good memory performance. MIA Anxiety = Score on the memory anxiety scale of the MIA (out of 65), with higher scores indicating more self-rated anxiety while completing daily task that have a memory component.