Skip to main content
NCBI home page
The Journals of Gerontology Series B: Psychological Sciences and Social Sciences logoLink to The Journals of Gerontology Series B: Psychological Sciences and Social Sciences
. 2018 Apr 6;74(4):585–594. doi: 10.1093/geronb/gby026

Stereotype Threat Reduces the Positivity of Older Adults’ Recall

Sarah J Barber 1,✉, Jordan Seliger 1, Nicholas Yeh 1, Shyuan Ching Tan 1
PMCID: PMC6460340  PMID: 29660076

Abstract

Objectives

As people get older, they show a relative preference to remember positive information over negative information. In two experiments, we tested whether the positivity of older adults’ memory is affected by stereotype threat about age-related cognitive declines. We also tested whether highlighting a positive aging stereotype (older adults are wise) would inoculate older adults from stereotype threat’s adverse effects.

Method

In Experiments 1 and 2, we manipulated whether stereotypes about age-related cognitive decline were highlighted (stereotype threat) or mitigated (stereotype alleviation). In Experiment 2, we included a third condition (intervention + stereotype threat), which highlighted positive and negative aging stereotypes. Participants then saw emotionally evocative pictures and completed a memory test.

Results

In both experiments, stereotype threat selectively reduced older adults’ memory for positive pictures but did not affect their memory for negative pictures. This eliminated the positivity effect (i.e., the Age × Valence interaction; Experiment 1). Our positive stereotype intervention did not reduce stereotype threat’s adverse effect (Experiment 2).

Discussion

Our findings show that the positivity effect is more robust when testing situations minimize stereotype threat. They also suggest that health interventions designed to capitalize on the positivity effect should ensure that ageist stereotypes are mitigated in the environment.

Keywords: Ageism, Emotional memory, Positivity effect, Socioemotional selectivity theory, Stereotype threat

According to socioemotional selectivity theory (Carstensen, 1995, 2006), perceptions of time play a fundamental role in how people prioritize instrumental and emotional goals. Younger adults typically view their future time as expansive. Because of this, younger adults prioritize future-oriented goals such as expanding their knowledge and experiencing novel situations. In contrast, older adults typically view their future time as finite. Because of this, older adults shift their goal hierarchies to prioritize present-oriented goals such as maximizing their current emotional well-being. One way that emotional satisfaction can be maximized is by preferentially processing positive information (see Reed & Carstensen, 2012). This age-related trend, in which older adults’ preferentially favor positive over negative information in attention and memory, is known as the positivity effect (see Mather, 2016; Mather & Carstensen, 2005).

Although a recent meta-analysis of 100 empirical studies confirmed that the positivity effect is robust and reliable (Reed, Chan, & Mikels, 2014), the goal of this current research was to examine how it is affected by stereotype threat about age-based cognitive decline. Stereotype threat occurs in situations where people are concerned that poor performance will confirm—either to themselves and/or to others—that a negative self-relevant stereotype is true. In response to stereotype threat, people often underperform compared with their potential (Steele & Aronson, 1995; for a recent review, see Spencer, Logel, & Davies, 2016). For example, older adults are characterized as being forgetful and senile. When reminded of this stereotype, older adults can experience stereotype threat and underperform on memory and cognitive tasks (e.g., Hess, Auman, Colcombe, & Rahhal, 2003; Hess, Emery, & Queen, 2009; Mazerolle et al., 2017; for reviews, see Barber & Mather, 2014; Chasteen, Kang, & Remedios, 2011).

Although the processes underlying stereotype threat–related performance deficits for older adults are still debated (see Barber, 2017; Popham & Hess, 2016), multiple accounts of stereotype threat converge in predicting that that positive information, relative to negative information, will be particularly affected by stereotype threat. For example, one commonly cited theory is Schmader, Johns, and Forbes’ (2008) executive control interference account. According to this framework, stereotype threat leads to evaluative concerns, stress, and negative affect. This sets in motion self-monitoring and emotion-regulation processes. However, because these processes require cognitive control resources, there are fewer resources available to devote to the critical task, and thus performance suffers. Of relevance to these studies, some research has shown that the positivity effect requires cognitive control resources (Knight et al., 2007; Mantantzis, Schlaghecken, & Maylor, 2017; Mather & Knight, 2005; but see Allard & Isaacowitz, 2008; Bohn, Kwong See, & Fung, 2016; Leal, Noche, Murray, & Yassa, 2016 for contradictory accounts). For instance, dividing attention reduces the relative positivity of older adults’ recall (Mather & Knight, 2005). Thus, if stereotype threat reduces the availability of cognitive control resources, then it should preferentially reduce older adults’ memory for positive, compared with negative, information.

The hypothesis that stereotype threat should preferentially reduce older adults’ memory for positive information is also consistent with the regulatory focus account of stereotype threat. This account is based on Higgins’ (1997) notions of promotion and prevention regulatory focus. People with a promotion focus are concerned with achieving gains and making improvements. In contrast, people with a prevention focus are concerned with avoiding losses and meeting their obligations. It has been argued that under stereotype threat people adopt a short-term prevention focus. Rather than eagerly approaching the gains that will make them their best (i.e., a promotion focus), under stereotype threat people vigilantly avoid the losses that will confirm that they are their worst (i.e., a prevention focus; Seibt & Förster, 2004). Given that prevention focus increases sensitivity to losses and is associated with a deeper level of processing for negative information (Touryan et al., 2007), this account of stereotype threat again leads to the hypothesis that threat should preferentially reduce older adults’ memory for positive, compared with negative, information.

In summary, although stereotype threat is known to reduce older adults’ memory for neutral information, no previous study has examined the impact of explicit stereotype threat on older adults’ memory for emotional information. However, both the executive control interference and regulatory focus accounts of stereotype threat converge in their predictions. Thus, regardless of the exact processes underlying how stereotype threat affects performance, older adults’ memory impairments arising from age-based stereotype threat should be greater for positive than negative information. As a result of this, age-based stereotype threat should also reduce or eliminate the positivity effect (i.e., the age by valence interaction in which older adults’ memory is relatively more positive and/or less negative younger adults’ memory).

Experiment 1

Design

A 2 (Age group: Younger adult or Older adult) × 2 (Stereotype condition: Alleviation or Threat) × 2 (Picture valence: Positive or Negative) design was used. Age group and stereotype condition were between-subject factors, and picture valence was manipulated within-subject factors.

Participants

Participants in this study were 56 younger adults and 53 older adults. The older adult participants were recruited from local senior centers, from classes offered through the Osher Lifelong Learning Institute at San Francisco State University, and from a database of individuals who had previously expressed interest in volunteering for studies related to aging issues. The younger adult participants were recruited through San Francisco State University’s psychology SONA participant pool. On completion of the study, participants were compensated with either $5 in cash or with credits toward their psychology course requirements. From our initial sample, we excluded data from the four participants (one younger adult and three older adults) who scored less than 26 on the Mini–Mental State Exam (MMSE; Folstein, Folstein, & McHugh, 1975). On debriefing, one older adult participant from the stereotype threat group withdrew consent for her data to be used. This left a final sample size of 55 younger adults (aged 18–31, M age = 22.53, 47% female) and 49 older adults (aged 57–88, M age = 73.33, 33% female). Within this final sample, the older adults had completed significantly more years of education (M = 16.48) than the younger adults (M = 15.11), t(102) = 3.29, p = .001. Participants in the two age groups did not significantly differ in their self-rated health, t(102) = 0.92, p = .36, or quality of life, t(102) = 1.91, p = .059. For more demographic details as a function of age group and stereotype condition, see Table 1.

Table 1.

Experiment 1 Participant Demographic Characteristics as a Function of Age Group and Stereotype Group

Younger adults Older adults
Stereotype alleviation (n = 28) Stereotype threat (n = 27) Stereotype alleviation (n = 24) Stereotype threat (n = 25)
Age (years) 22.18 (2.87) 22.89 (2.89) 72.42 (6.72) 72.24 (7.10)
Gender 24 women 23 women 18 women 15 women
4 men 4 men 6 men 10 men
Ethnicity 12 Hispanic 8 Hispanic 2 Hispanic 1 Hispanic
13 Non-Hispanic 18 Non-Hispanic 19 Non-Hispanic 22 Non-Hispanic
1 declined to state 1 declined to state 3 declined to state 2 declined to state
Race 7 Caucasian 7 Caucasian 18 Caucasian 18 Caucasian
5 Asian 6 Asian 3 Asian 4 Asian
2 African American 3 African American 0 African American 0 African American
2 Biracial/Multiracial 6 Biracial/Multiracial 1 Biracial/Multiracial 1 Biracial/Multiracial
7 “Other” 5 “Other” 2 “Other” 0 “Other”
5 declined to state 0 declined to state 0 declined to state 2 declined to state
Education (years) 14.93 (1.12) 15.30 (0.72) 16.40 (2.49) 16.56 (3.33)
Self-rated health 3.64 (0.68) 3.81 (0.88) 3.88 (0.95) 3.88 (0.83)
Self-rated quality of life 5.18 (1.06) 5.41 (1.12) 5.54 (1.35) 5.92 (1.22)
MMSE score 28.96 (1.00) 28.89 (0.89) 28.54 (1.14) 28.40 (1.26)
Open in a new tab

Note: MMSE = Mini–Mental State Exam. Numbers in parentheses represent SD. Self-rated health was answered on a 1 (poor) to 5 (excellent) scale. Self-rated quality of life was answered on a 1 (very poor) to 7 (excellent) scale.

Materials

A total of 30 pictures (15 positive and 15 negative) from the International Affective Picture System (IAPS; Lang, Bradley, & Cuthbert, 2008) served as the critical stimuli. (IAPS numbers for the positive critical pictures: 1710, 1750, 2260, 2340, 2360, 2370, 2550, 5001, 5260, 5470, 5621, 5660, 5831, 7480, 8370. IAPS numbers for the negative critical pictures: 2205, 2490, 2590, 2691, 2750, 6560, 7359, 9050, 9220, 9340, 9415, 9470, 9560, 9630, 9810.) Based on the published normative data, the selected positive and negative pictures differed in their IAPS scores of emotional valence, t(28) = 31.13, p < .001, d = 11.37, but not in their IAPS scores of emotional arousal, t(28) = −0.86, p = .40, d = −0.31. An additional four neutrally valenced pictures (IAPS numbers: 7090, 7100, 7130, and 7150) were used to buffer against primacy and recency effects.

Procedure

At the beginning of the study, all participants were told that they would be completing an emotional memory test. However, prior to encoding, we manipulated whether stereotypes about age-based cognitive decline were disputed or confirmed. Participants were randomly assigned to receive either stereotype alleviation or stereotype threat instructions. More specifically, participants assigned to the stereotype alleviation condition read the following: “In general, older adults do not tend to do as well as younger adults on memory tests. However, that is not true for this particular memory test. On this picture memory test, older adults always do as well as younger adults.” In contrast, participants in the stereotype threat condition read: “In general, older adults do not tend to do as well as younger adults on memory tests. In the current study, we are testing whether this age difference occurs for this particular memory test.” On the computer screen, all participants then saw a series of 34 pictures (2 primacy buffers, 30 critical, 2 recency buffers) and were asked to study them in preparation for a memory test. Each picture was shown for 4 s. Order of the positive and negative critical pictures was randomized, but the same random order was used for all participants.

Immediately after viewing the pictures, participants completed a self-paced free recall test. This test began by restating the instructions just described (i.e., that the test was age-fair or that it was designed to examine age differences in memory). In the stereotype threat condition, participants were also asked to list their age prior to beginning the free recall test. However, due to experimenter error, one older adult participant in the stereotype threat condition was not asked to list her age prior to starting the memory test. Participants then wrote descriptions of each picture that they recalled. Participants were instructed that their descriptions could be short but must contain enough details for an independent reader to match it to one of the studied pictures (for similar procedures, see Barber, Castrellon, Opitz, & Mather, 2017; Barber, Opitz, Martins, Sakaki, & Mather, 2016; Charles, Mather, & Carstensen, 2003). Following picture recall, all participants completed a demographic questionnaire and we administered the MMSE (Folstein et al., 1975).

Recall Test Scoring

A primary coder (S.B.) scored whether or not each free recall written response matched one of the studied pictures. A second coder (J.S.) independently scored all of the older adult participants’ responses; there was 98.86% agreement with the primary coder. A third coder (K.T.) independently scored all of the younger adult participants’ responses; there was 97.64% agreement with the primary coder. Scores from the primary coder were used in all subsequent analyses. Only 37 of the 1,243 written responses could not be matched by the primary coder to a specific picture; the majority of these were descriptions that were too general to match to any one picture (e.g., “water” or “friends”). The distribution of unmatched responses did not vary as a function of age group or stereotype group: 8 responses came from older adults in the stereotype alleviation group, 9 responses from older adults in the stereotype threat group, 6 responses from younger adults in the stereotype alleviation group, and 13 responses from younger adults in the stereotype threat group.

Results and Discussion

We hypothesized that older adults’ memory impairments arising from age-based stereotype threat would be greater for positive information than negative information and that this in turn would reduce or eliminate the positivity effect in recall (i.e., the Age × Valence interaction in which older adults’ memory is relatively more positive and/or less negative as compared to younger adults’ memory). To evaluate these predictions, we first conducted a 2 (Age group) × 2 (Stereotype condition) × 2 (Picture valence) analysis of variance (ANOVA) on the proportion of critical pictures recalled (buffer pictures were excluded from all analyses). Within this analysis, there were only three significant effects. First, there was a main effect of age group, F(1, 100) = 5.41, MSE = 0.03, p = .022, ηp2 = 0.051, which reflected the fact that younger adults had higher recall than older adults. Second, there was an interaction between age group and picture valence, F(1, 100) = 5.25, MSE = 0.01, p = .024, ηp2 = 0.050. This replicates the standard positivity effect. However, these previous effects were qualified by a significant three-way interaction between age group, stereotype condition, and picture valence, F(1, 100) = 4.33, MSE = 0.01, p = .040, ηp2 = 0.041 (see Figure 1).

Figure 1.

Figure 1.

Open in a new tab

Mean proportion of pictures recalled as a function of age group, stereotype group, and picture valence. Error bars represent SEM.

To decompose this three-way interaction, we next examined how our stereotype condition manipulation affected both the younger and the older adults. Looking first at the younger adults, we conducted a 2 (Stereotype condition) × 2 (Picture valence) ANOVA on the proportion of critical pictures recalled. Within this analysis, there was a significant main effect of picture valence, F(1, 53) = 5.97, MSE = 0.01, p = .018, ηp2 = 0.101. This reflected the fact that younger adults had higher recall of the negative pictures (stereotype threat: M = 0.40, SD = 0.18; stereotype alleviation: M = 0.39, SD = 0.13) than of the positive pictures (stereotype threat: M = 0.37, SD = 0.14; stereotype alleviation: M = 0.32, SD = 0.12). However, younger adults’ memory was not affected by our stereotype manipulation. There was no significant main effect of stereotype condition, F(1, 53) = 0.60, MSE = 0.03, p = .442, ηp2 = 0.011, and no significant interaction between stereotype condition and picture valence, F(1, 53) = 0.88, MSE = 0.01, p = .352, ηp2 = 0.016.

Having established that the younger adults’ memory was not affected by our stereotype manipulation, we next examined its impact on the older adult participants’ memory performance. Novel to this study, we found that the detrimental effect of stereotype threat on older adults’ memory was valence specific. Within a 2 (Stereotype condition) × 2 (Picture valence) ANOVA on the proportion of critical pictures recalled by the older adult participants, there was no significant main effect of stereotype condition, F(1, 47) = 1.95, MSE = 0.03, p = .169, ηp2 = 0.040, or picture valence, F(1, 47) = 0.55, MSE = 0.01, p = .464, ηp2 = 0.011. However, consistent with our predictions, there was a significant interaction between stereotype condition and picture valence, F(1, 47) = 4.43, MSE = 0.01, p = .041, ηp2 = 0.086. Follow-up independent-sample t tests showed that older adults in the stereotype threat condition recalled significantly fewer positive pictures (M = 0.28, SD = 0.13) than older adults in the stereotype alleviation condition (M = 0.36, SD = 0.14), t(47) = −2.20, p = .032, d = −0.375. However, older adults in the stereotype threat and stereotype alleviation conditions did not significantly differ in their recall of the negative pictures (stereotype threat: M = 0.30, SD = 0.15; stereotype alleviation: M = 0.31, SD = 0.11), t(47) = −0.26, p = .796.

Because the effects of our stereotype manipulation were age and valence specific, it also affected the likelihood that a positivity effect was observed. Looking first at participants in the stereotype alleviation condition, we conducted a 2 (Age group) × 2 (Picture valence) ANOVA on the proportion of critical pictures recalled. Within this analysis, there was no significant main effect of either age group, F(1, 50) = 0.37, MSE = 0.02, p = .546, ηp2 = 0.007, or picture valence, F(1, 50) = 0.27, MSE = 0.01, p = .607, ηp2 = 0.005. However, there was a significant interaction between age group and picture valence, F(1, 50) = 13.15, MSE = 0.01, p = .001, ηp2 = 0.208. Follow-up independent t tests showed that after receiving stereotype alleviation instructions, older adults recalled significantly fewer negative pictures (M = 0.31) than the younger adults (M = 0.39), t(50) = −2.34, p = .023, d = −0.655. However, recall of the positive pictures did not significantly differ between the older and younger adults (M = 0.36 and M = 0.32, respectively), t(50) = 1.09, p = .280. Thus, a positivity effect emerged when participants received the stereotype alleviation test instructions.

However, a different pattern of results was obtained when examining participants who received the stereotype threat test instructions. In a 2 (Age group) × 2 (Picture valence) ANOVA on the proportion of critical pictures recalled by participants in the stereotype threat conditions, we observed only a significant main effect of age group, F(1, 50) = 6.49, MSE = 0.03, p = .014, ηp2 = 0.115. Within this analysis, there was no significant effect of picture valence, F(1, 50) = 1.59, MSE = 0.01, p = .213, ηp2 = 0.031, and no significant interaction between age group and picture valence, F(1, 50) = 0.02, MSE = 0.01, p = .895, ηp2 < .001. Follow-up independent t tests showed that after receiving stereotype threat instructions, older adults recalled significantly fewer negative pictures (M = 0.30) than the younger adults (M = 0.40), t(50) = −2.03, p = .048, d = −0.565. They also recalled significantly fewer positive pictures (M = 0.28) than the younger adults (M = 0.37), t(50) = −2.43, p = .019, d = −0.675.

In summary, the results of Experiment 1 supported our hypotheses. Age-based stereotype threat significantly impaired older adults’ memory for the positive pictures but did not significantly affect their memory for the negative pictures. In contrast, there was no effect of the stereotype manipulation on younger adults’ memory. This was expected because the stereotype was not relevant to them. Because the effects of age-based stereotype threat were both age and valence specific, it also affected the likelihood of observing a positivity effect (i.e., an interaction between age and valence). Although a positivity effect was observed when participants received the stereotype alleviation instructions, it was eliminated when participants received the stereotype threat instructions. This result adds to our understanding about when the positivity effect is most likely to be observed. In a prior meta-analysis, it was reported that the positivity effect is most likely to emerge when processing is unconstrained and when there is a large age gap between the younger and older adult participants (Reed et al., 2014). Adding to this, our study suggests that the positivity effect is most likely to be observed when stereotype threat is reduced.

Experiment 2

In Experiment 2, we had two goals. The first was to demonstrate the replicability of the key Experiment 1 finding (i.e., that age-based stereotype threat selectively reduces older adults’ memory for positive pictures as compared to negative pictures). Although replications have traditionally been rare occurrences (Makel, Plucker, & Hegarty, 2012), there is an increased acknowledgement of their importance (see Asendorpf et al., 2013; Earp & Trafimow, 2015; Schmidt, 2009). To this end, we undertook a direct replication of our key Experiment 1 effect.

Our second goal was to test whether age-based stereotype threat can be removed by reminding older adults of a positive aging stereotype, namely that aging brings wisdom (e.g., Hummert, 1990). Although wisdom is multifaceted (e.g., Baltes & Staudinger, 2000), cognitive expertise is a necessary (but not sufficient) quality for obtaining it. Likewise, declines in working memory or senility may lead to the loss of wisdom (see Baltes & Staudinger, 2000; Jeste et al., 2010). Thus, if an older adult is respected for their wisdom, it implies an absence of senility or severe memory declines. We reasoned that highlighting this positive stereotype (older adults are wise) may attenuate concerns about a negative age stereotype (older adults are senile). Supporting this, some prior research with younger adults has shown that activating a positive stereotype (which is both self-relevant and domain-relevant) can eliminate the adverse effects of stereotype threat (Rydell & Boucher, 2010; Rydell, McConnell, & Beilock, 2009). These prior studies have focused on the adverse effects of gender-based stereotypes (i.e., women are bad at math) in affecting undergraduate women’s math performance. When this negative gender stereotype is made salient, women tend to experience stereotype threat and underperform on math tests. However, this adverse effect is eliminated when a positive stereotype is simultaneously activated (i.e., that college students are good at math).

Design

A 3 (Stereotype condition: Alleviation or Threat or Intervention + Threat) × 2 (Picture valence: Positive or Negative) design was used. Stereotype condition was manipulated between subjects, whereas picture valence was manipulated within subjects.

Participants

Participants were 90 older adults who ranged in age from 60 to 79 (M age = 70.07). Within this sample, there were 50 women and 40 men. They had completed an average of 15.61 years of education and all scored 26 or higher on the MMSE. For more demographic details as a function of stereotype condition, see Table 2. The participants were recruited from a list of volunteers from the San Francisco area who had previously expressed interest in participating in research studies related to aging issues.

Table 2.

Experiment 2 Participant Demographic Characteristics as a Function of Stereotype Group

Stereotype alleviation (n = 30) Stereotype threat (n = 30) Intervention + stereotype threat (n = 30)
Age (years) 70.17 (4.62) 70.27 (4.41) 69.78 (5.08)
Gender 14 women, 16 men 17 women, 13 men 19 women, 11 men
Ethnicity 1 Hispanic 0 Hispanic 0 Hispanic
29 Non-Hispanic 30 Non-Hispanic 29 Non-Hispanic
0 declined to state 0 declined to state 1 declined to state
Race 23 Caucasian 22 Caucasian 23 Caucasian
5 Asian 5 Asian 6 Asian
0 African American 1 African American 0 African American
0 Biracial/Multiracial 1 Biracial/Multiracial 1 Biracial/Multiracial
1 “Other” 1 “Other” 0 “Other”
1 declined to state 0 declined to state 0 declined to state
Education (years) 15.40 (5.95) 17.20 (2.04) 14.36 (7.16)
Self-rated health 7.43 (0.97) 7.98 (0.95) 7.72 (0.85)
Fear of AD: General fear 38.47 (11.15) 36.43 (10.97) 37.80 (13.55)
Fear of AD: Physical symptoms 8.90 (2.01) 9.53 (2.60) 9.13 (2.96)
Fear of AD: Catastrophic attitude 11.77 (4.55) 14.07 (5.39) 13.03 (5.46)
MMSE score 29.20 (1.54) 28.97 (0.81) 28.37 (1.03)
Open in a new tab

Note: MMSE = Mini–Mental State Exam. Numbers in parentheses represent SD. Self-rated health was answered on a 1 (poor) to 9 (excellent) scale. One participant in the stereotype threat condition did not provide self-rated health.

Procedure

At the beginning of the study, all participants completed the emotional memory test that was used in Experiment 1. Instructions for this test varied as a function of stereotype group. For participants randomly assigned to the stereotype alleviation and stereotype threat groups, the instructions and procedures were identical to those used in Experiment 1. New to this study, participant in the intervention + stereotype threat first read the following prompt:

Old age can bring wisdom. As people get older, they accumulate a lifetime of experiences. This allows older people to make better practical decisions and increase their emotional well-being. By taking the time to reflect upon their knowledge, older people are often more successful in navigating the important challenges of social life. As “experts” on living, older adults are a valuable source of advice for the younger generations.

After reading this prompt, participants in the intervention + stereotype threat group received the same instructions as participants in the stereotype threat group; they were told that older adults do not tend to do as well as younger adults on memory tests and that the goal of the current study was to examine whether this age difference would occur in memory for pictures. As in the stereotype threat group, these participants were also asked to list their age prior to beginning the free recall test.

Following picture recall, participants next completed the 30-item Fear of Alzheimer’s Disease Scale, which consists of three subscales: General Fear, Physical Symptoms, and Catastrophic Attitudes (French, Floyd, Wilkins, & Osato, 2012). For demographic characteristics as a function of stereotype group, see Table 2. Participants also completed a demographics questionnaire and the MMSE (Folstein et al., 1975). Participants were also asked to provide a saliva sample, which was subsequently used in genotype analyses. However, these analyses will not be discussed further here. On completion of the study, participants were compensated at a rate of $15/hr.

Figure 2.

Figure 2.

Open in a new tab

Mean proportion of pictures recalled by older adult participants as a function of stereotype group and picture valence. Error bars represent SEM.

Recall Test Scoring

Two coders (S.B. and J.S.) independently scored whether or not each free recall written response matched one of the studied pictures. There was 98.18% agreement between the coders. Scores from the primary coder (S.B.) were used in all subsequent analyses. Only 32 of the 1,264 written responses were not matched by the primary coder to a specific picture. These tended to be general descriptions that could pertain to multiple pictures (e.g., “family”). The distribution of responses that could not be matched to specific pictures did not vary as a function of stereotype group: 12 came from participants in the stereotype alleviation group, 10 from the stereotype threat group, and 10 from the intervention + stereotype threat group.

Results and Discussion

There were two aims to this study. First, we aimed to replicate the Experiment 1 finding that older adults’ memory impairments arising from age-based stereotype threat would be greater for positive information than negative information. Second, we tested the efficacy of a wisdom intervention in ameliorating this deficit. To evaluate these aims, we conducted a 3 (Stereotype condition) × 2 (Picture valence) ANOVA on the proportion of critical pictures recalled (buffer pictures were excluded from all analyses). Within this analysis, there was no main effect of stereotype condition, F(1, 87) = 1.42, MSE = 0.03, p = .247, ηp2 = 0.032. However, there was a main effect of valence, F(1, 87) = 16.67, MSE = 0.01, p < .001, ηp2 = 0.161, which was driven by the fact that participants tended to recall more of the negative pictures than the positive pictures. Importantly, the main effect of picture valence was qualified by its significant interaction with stereotype condition, F(2, 87) = 3.95, MSE = 0.01, p = .023, ηp2 = 0.083.

To decompose this interaction, we next examined whether our three stereotype groups significantly differed in their recall of either negative or positive pictures. Looking first at memory for the negative pictures, a single-factor (Stereotype condition) ANOVA showed no significant group differences, F(2, 87) = 0.10, MSE = 0.02, p = .907, ηp2 = 0.002. As shown in Figure 2, participants in the three conditions performed equivalently in their memory for the negative pictures (stereotype alleviation: M = 0.44, SD = 0.11, stereotype threat: M = 0.43, SD = 0.18 intervention + stereotype threat: M = 0.45, SD = 0.11). However, a different pattern was observed for memory of the positive pictures. In a single-factor (Stereotype condition) ANOVA, there was a significant effect of stereotype condition, F(2, 87) = 4.48, MSE = 0.02, p = .014, ηp2 = 0.093. Follow-up independent-sample t tests showed that participants in the stereotype alleviation condition recalled significantly more positive pictures (M = 0.43, SD = 0.15) than participants in both the stereotype threat condition (M = 0.36, SD = 0.11), t(58) = 2.09, p = .041, d = 0.539, and the intervention + stereotype threat condition (M = 0.33, SD = 0.13), t(58) = −2.72, p = .009, d = 0.702. There was no significant difference in recall of positive pictures between participants in the stereotype threat and intervention + stereotype threat conditions, t(58) = 0.86, p = .395.

In summary, these results replicate the pattern reported for the older adult participants in Experiment 1. As shown in Figure 2, the effects of age-based stereotype threat about cognitive decline were valence specific. Stereotype threat impaired older adults’ memory for positive pictures but did not affect their memory for the negative pictures. However, because Experiment 2 did not include younger adults, these results cannot directly address the impact of stereotype threat on the positivity effect, which is defined as an interaction between age and valence.

Results of Experiment 2 also showed that our intervention was not effective. In fact, the adverse effects of stereotype threat on memory for negative pictures were numerically strongest in this condition. This has some similarities to younger adult studies showing that making salient positive stereotypes salient can sometimes lead to “choking” under the pressure of high expectations (e.g., Cheryan & Bodenhausen, 2000). It is also possible that our positive stereotype intervention was ineffective because it highlighted a positive stereotype (i.e., older adults are wise) that pertained to the same self-identify as the negative stereotype (i.e., older adults are senile). This is in contrast to the prior studies by Rydell and colleagues, in which the highlighted positive stereotype (i.e., college students are good at math) pertained to a different self-identity than the negative stereotype (i.e., women are bad at math). As described in the Multiple Self-Aspects Framework (McConnell, 2011), a person’s self-concept is a collection of multiple, context-dependent selves, which can be organized in a network of distributed nodes. For instance, people can think of themselves in terms of their social roles (e.g., husband or mother), gender identity, racial/ethnic identity, occupational roles, religious affiliations, political affiliations, and so on. The extent to which any of these selves is activated depends on the context. However, activating any one aspect of the self-concept (e.g., college student) should temporarily suppress accessibility of other aspects of the self-concept (e.g., woman). Consistent with this, Rydell and colleagues (2009) found that their positive stereotype intervention (i.e., women are good at math) increased the accessibility of women’s college student self-identity while simultaneously inhibiting the accessibility of their gender identity. Thus, positive stereotype interventions may only be effective when they highlight a different aspect of the self-identity than the one implicated in the negative stereotype. Future research is needed to test this hypothesis.

General Discussion

Older adults are often stereotyped as being hopeless, despondent, and depressed. However, old age is often associated with enhanced emotional well-being. As people get older, they tend to become emotionally more positive and less negative (see Charles & Carstensen, 2010). They also tend to display positivity effects by preferentially favoring positive over negative information in attention and memory (see Mather, 2016). Although positivity effects are reliable and robust (Reed et al., 2014), the current results suggest that they are less likely to appear in the presence of ageist stereotypes. Across two experiments, we found that when the testing situation emphasized age-related cognitive declines, older adults recalled fewer positive pictures. In contrast, stereotype threat did not significantly affect older adults’ recall of the negative pictures. In Experiment 1, we also found that stereotype threat did not affect younger adults’ recall of either positive or negative picture. Because of this, the positivity effect in Experiment 1 (which is defined as an interaction between age and valence) was only present in our stereotype alleviation condition and was absent in our stereotype threat condition.

The fact that stereotype threat preferentially reduced older adults’ recall of positive, relative to negative, pictures is consistent with both the theory that stereotype threat reduces the availability of cognitive control resources (Schmader et al., 2008) and the theory that stereotype threat induces a prevention regulatory focus (Seibt & Förster, 2004). Because these two accounts converged in their predictions, in the current studies we did not attempt to disentangle them or identify their relative contributions to these effects. Although prior results from our lab have tended to support the regulatory focus account (see Barber, 2017), future research should examine this issue within the context of emotional memory, for example, by including measures of cognitive control and regulatory focus or by varying the tasks’ reward structure.

There are also limitations that will need to be addressed in future research. For instance, in the current studies we did not include a control condition. Rather, our memory test instructions were designed to either amplify or alleviate older adults’ performance concerns about age-related memory decline. This was done to clearly separate the amount of threat experienced across the conditions. However, it leaves open the question of whether our stereotype threat instructions decreased the positivity of older adults’ recall, whether the stereotype alleviation instructions increased the positivity of older adults’ recall, or whether both effects simultaneously occurred. Although some prior research has suggested that older adults’ memory performance is statistically similar in control and alleviation conditions (see Hess et al., 2003), future research is needed to test this within the context of emotional memory performance. Another limitation is that we limited our critical stimuli to positive and negative pictures. However, in a previous study by Krendl, Ambady, and Kensinger (2015), subliminally priming older adults with ageist stereotypes selectively increased false alarms to neutral words, but not to positive or negative emotional words. Although the processes underlying subliminal priming differ from those underlying stereotype threat (see Barber & Mather, 2014), it is possible that stereotype threat–related memory impairments for older adults may also be greatest for neutral, compared with emotional, information.

In the current studies we also used a blatant manipulation of stereotype threat. It is unclear whether a more subtle manipulation would have changed the magnitude of our results. One meta-analysis examined age based stereotype threat effects across multiple performance domains and found smaller effects sizes when blatant, rather than subtle, manipulations were used (Lamont, Swift, & Abrams, 2015). However, a more recent meta-analysis found that within the domain of episodic memory there were larger effect sizes when blatant manipulations were used as compared to when subtle manipulations were used (Armstrong, Gallant, Li, Patel, & Wong, 2017). Future research is needed that more directly examines how these methodological variations affect the magnitude of age based stereotype threat effects.

We also did not include assessments of mood, anxiety, or arousal. However, we do note that changes in affective responding cannot readily account for the observed effects. Prior studies have shown that older adults tend to engage in mood-incongruent processing; they attend more to positive pictures when in a negative mood state (Isaacowitz, Toner, Goren, & Wilson, 2008; Isaacowitz, Toner, & Neupert, 2009; for a review, see Isaacowitz, 2012). Thus, if stereotype threat induced a negative mood, older adults would probably have responded by increasing the amount of attention paid toward the positive stimuli. Presumably, this would have increased (rather than decreased) the positivity of their recall.

More generally, the conclusion that ageist stereotypes can selectively reduce the memorability of positive information for older adults has important implications for health interventions. Capitalizing on the robust nature of the positivity effect, health interventions for older adults often use positively framed, rather than negatively framed, messages (see Reed & Carstensen, 2015). However, some older adults report feeling concerned that their health care provider was negatively evaluating them because of their age (Abdou, Fingerhut, Jackson, & Wheaton, 2016; Phibbs & Hooker, 2017). Because of this, the current results suggest that positively framed messages may not be beneficial in promoting the desired health changes. To best capitalize on the positivity effect, testing/intervention environments should aim to mitigate older adults’ concerns about ageist evaluations and stereotypes.

Funding

This work was supported by a grant from the National Institute on Aging to S.J.B. (grant number R01-AG046464).

Conflict of Interest

None reported.

Acknowledgments

Our thanks to Natalie Taylor-Ketcham, Ismael Tellez, and Kevin Tiziani for assistance in data collection and entry.

References

  1. Abdou C. M., Fingerhut A. W., Jackson J. S., & Wheaton F (2016). Healthcare stereotype threat in older adults in the health and retirement study. American Journal of Preventive Medicine, 50, 191–198. doi:10.1016/j.amepre.2015.07.034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Allard E. S., & Isaacowitz D. M (2008). Are preferences in emotional processing affected by distraction? Examining the age-related positivity effect in visual fixation within a dual-task paradigm. Aging, Neuropsychology, and Cognition, 15, 725–743. doi:10.1080/13825580802348562 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Armstrong B., Gallant S. N., Li L., Patel K., & Wong B. I (2017). Stereotype threat effects on older adults’ episodic and working memory: A meta-analysis. The Gerontologist, 57, 193–205. doi:10.1093/geront/gnx056 [DOI] [PubMed] [Google Scholar]
  4. Asendorpf J. B., Conner M., De Fruyt F., De Houwer J., Denissen J. J., Fiedler K., … Perugini M (2013). Recommendations for increasing replicability in psychology. European Journal of Personality, 27, 108–119. doi:10.1002/per.1919 [Google Scholar]
  5. Baltes P. B., & Staudinger U. M (2000). Wisdom: A metaheuristic (pragmatic) to orchestrate mind and virtue toward excellence. American Psychologist, 55, 122–136. doi:10.1037/0003-066X.55.1.122 [DOI] [PubMed] [Google Scholar]
  6. Barber S. J. (2017). An examination of age-based stereotype threat about cognitive decline: Implications for stereotype threat research and theory development. Perspectives on Psychological Science, 12, 62–90. doi:10.1177/1745691616656345 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Barber S. J., Castrellon J. J., Opitz P., & Mather M (2017). Younger and older adults’ collaborative recall of shared and unshared emotional pictures. Memory & Cognition, 45, 716–730. doi:10.3758/s13421-017-0694-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Barber S. J., & Mather M (2014). Stereotype threat in older adults: When and why does it occur, and who is most affected? In P. Verhaegen & C. Hertzog (Eds.), The Oxford handbook of emotion, social cognition, and problem solving in adulthood (pp. 302–320). Oxford, UK: Oxford University Press. [Google Scholar]
  9. Barber S. J., Opitz P. C., Martins B., Sakaki M., & Mather M (2016). Thinking about a limited future enhances the positivity of younger and older adults’ recall: Support for socioemotional selectivity theory. Memory & Cognition, 44, 869–882. doi:10.3758/s13421-016-0612-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bohn L., Kwong See S. T., & Fung H. H (2016). Time perspective and positivity effects in Alzheimer’s disease. Psychology and Aging, 31, 574–582. doi:10.1037/pag0000084 [DOI] [PubMed] [Google Scholar]
  11. Carstensen L. L. (1995). Evidence for a life-span theory of socioemotional selectivity. Current Directions in Psychological Science, 4, 151–156. doi:10.1111/1467-8721.ep11512261 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Carstensen L. L. (2006). The influence of a sense of time on human development. Science, 312, 1913–1915. doi:10.1126/science.1127488 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Charles S., & Carstensen L. L (2010). Social and emotional aging. Annual Review of Psychology, 61, 383–409. doi:10.1146/annurev.psych.093008.100448 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Charles S. T., Mather M., & Carstensen L. L (2003). Aging and emotional memory: The forgettable nature of negative images for older adults. Journal of Experimental Psychology: General, 132, 310–324. doi:10.1037/0096-3445.132.2.310 [DOI] [PubMed] [Google Scholar]
  15. Chasteen A. L., Kang S. K., & Remedios J. D (2011). Aging and stereotype threat: Development, process, and interventions. In M. Inzlicht & T. Schmader (Eds.), Stereotype threat: Theory, process, and application (pp. 202–216). New York, NY: Oxford University Press. [Google Scholar]
  16. Cheryan S., & Bodenhausen G. V (2000). When positive stereotypes threaten intellectual performance: The psychological hazards of “model minority” status. Psychological Science, 11, 399–402. doi:10.1111/1467-9280.00277 [DOI] [PubMed] [Google Scholar]
  17. Earp B. D., & Trafimow D (2015). Replication, falsification, and the crisis of confidence in social psychology. Frontiers in Psychology, 6, 621. doi:10.3389/fpsyg.2015.00621 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Folstein M. F., Folstein S. E., & McHugh P. R (1975). “Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189–198. doi:10.1016/0022-3956(75)90026-6 [DOI] [PubMed] [Google Scholar]
  19. French S. L., Floyd M., Wilkins S., & Osato S (2012). The fear of Alzheimer’s disease scale: A new measure designed to assess anticipatory dementia in older adults. International Journal of Geriatric Psychiatry, 27, 521–528. doi:10.1002/gps.2747 [DOI] [PubMed] [Google Scholar]
  20. Hess T. M., Auman C., Colcombe S. J., & Rahhal T. A (2003). The impact of stereotype threat on age differences in memory performance. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 58, 3–11. doi:10.1093/geronb/58.1.P3 [DOI] [PubMed] [Google Scholar]
  21. Hess T. M., Emery L., & Queen T (2009). Task demands moderate stereotype threat effects on memory performance. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 64B, 482–486. doi:10.1093/geronb/gbp044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Higgins E. T. (1997). Beyond pleasure and pain. American Psychologist, 52, 1280–1300. doi:10.1037/0003-066X.52.12.1280 [DOI] [PubMed] [Google Scholar]
  23. Hummert M. L. (1990). Multiple stereotypes of elderly and young adults: A comparison of structure and evaluations. Psychology and Aging, 5, 182–193. doi:10.1037/0882-7974.5.2.182 [DOI] [PubMed] [Google Scholar]
  24. Isaacowitz D. M. (2012). Mood regulation in real time: Age differences in the role of looking. Current Directions in Psychological Science, 21, 237–242. doi:10.1177/0963721412448651 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Isaacowitz D. M., Toner K., Goren D., & Wilson H. R (2008). Looking while unhappy: Mood-congruent gaze in young adults, positive gaze in older adults. Psychological Science, 19, 848–853. doi:10.1111/j.1467-9280.2008.02167.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Isaacowitz D. M., Toner K., & Neupert S. D (2009). Use of gaze for real-time mood regulation: Effects of age and attentional functioning. Psychology and Aging, 24, 989–994. doi:10.1037/a0017706 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jeste D. V., Ardelt M., Blazer D., Kraemer H. C., Vaillant G., & Meeks T. W (2010). Expert consensus on characteristics of wisdom: A Delphi method study. The Gerontologist, 50, 668–680. doi:10.1093/geront/gnq022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Knight M., Seymour T. L., Gaunt J. T., Baker C., Nesmith K., & Mather M (2007). Aging and goal-directed emotional attention: Distraction reverses emotional biases. Emotion, 7, 705–714. doi:10.1037/1528-3542.7.4.705 [DOI] [PubMed] [Google Scholar]
  29. Krendl A. C. Ambady N. & Kensinger E. A (2015). The dissociable effects of stereotype threat on older adults’ memory encoding and retrieval. Journal of Applied Research in Memory and Cognition, 4, 103–109. doi:10.1016/j.jarmac.2015.02.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lamont R. A., Swift H. J., & Abrams D (2015). A review and meta-analysis of age-based stereotype threat: Negative stereotypes, not facts, do the damage. Psychology and Aging, 30, 180–193. doi:10.1037/a0038586 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lang P. J., Bradley M. M., & Cuthbert B. N (2008). International Affective Picture System (IAPS): Affective ratings of pictures and instruction manual. Technical Report A-8. University of Florida, Gainesville, FL. [Google Scholar]
  32. Leal S. L. Noche J. A. Murray E. A. & Yassa M. A (2016). Positivity effect specific to older adults with subclinical memory impairment. Learning & Memory, 23, 415–421. doi:10.1101/lm.042010.116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Makel M. C., Plucker J. A., & Hegarty B (2012). Replications in psychology research: How often do they really occur?Perspectives on Psychological Science, 7, 537–542. doi:10.1177/1745691612460688 [DOI] [PubMed] [Google Scholar]
  34. Mantantzis K., Schlaghecken F., & Maylor E. A (2017). Food for happy thought: Glucose protects against age-related positivity effects under cognitive load. Psychology and Aging, 32, 203–209. doi:10.1037/pag0000169 [DOI] [PubMed] [Google Scholar]
  35. Mather M. (2016). The affective neuroscience of aging. Annual Review of Psychology, 67, 213–238. doi:10.1146/annurev-psych-122414-033540 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mather M., & Carstensen L. L (2005). Aging and motivated cognition: The positivity effect in attention and memory. Trends in Cognitive Science, 9, 496–502. doi:10.1016/j.tics.2005.08.005 [DOI] [PubMed] [Google Scholar]
  37. Mather M., & Knight M (2005). Goal-directed memory: The role of cognitive control in older adults’ emotional memory. Psychology and Aging, 20, 554–570. doi:10.1016/j.tics.2005.08.005 [DOI] [PubMed] [Google Scholar]
  38. Mazerolle M., Régner I., Barber S. J., Paccalin M., Miazola A. C., Huguet P., & Rigalleau F (2017). Negative aging stereotypes impair performance on brief cognitive tests used to screen for predementia. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 72, 932–936. doi:10.1093/geronb/gbw083 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. McConnell A. R. (2011). The multiple self-aspects framework: Self-concept representation and its implications. Personality and Social Psychology Review, 15, 3–27. doi:10.1177/1088868310371101 [DOI] [PubMed] [Google Scholar]
  40. Phibbs S., & Hooker K (2017). An exploration of factors associated with ageist stereotypes in a medical setting. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences. doi:10.1093/geronb/gbx034 [DOI] [PubMed] [Google Scholar]
  41. Popham L. E., & Hess T. M (2016). Age differences in the underlying mechanisms of stereotype threat effects. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 70, 223–232. doi:10.1093/geronb/gbt093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Reed A. E., & Carstensen L. L (2012). The theory behind the age-related positivity effect. Frontiers in Psychology, 3, 339. doi:10.3389/fpsyg.2012.00339 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Reed A. E., & Carstensen L. L (2015). Age-related positivity effect and its implications for social and health gerontology. In N. Pachana (Ed.), Encyclopedia of Geropsychology. Singapore: Springer. [Google Scholar]
  44. Reed A. E., Chan L., & Mikels J. A (2014). Meta-analysis of the age-related positivity effect: Age differences in preferences for positive over negative information. Psychology and Aging, 29, 1–15. doi:10.1037/a0035194 [DOI] [PubMed] [Google Scholar]
  45. Rydell R. J., & Boucher K. L (2010). Capitalizing on multiple social identities to prevent stereotype threat: The moderating role of self-esteem. Personality and Social Psychology Bulletin, 36, 239–250. doi:10.1177/0146167209355062 [DOI] [PubMed] [Google Scholar]
  46. Rydell R. J., McConnell A. R., & Beilock S. L (2009). Multiple social identities and stereotype threat: Imbalance, accessibility, and working memory. Journal of Personality and Social Psychology, 96, 949–966. doi:10.1037/a0014846 [DOI] [PubMed] [Google Scholar]
  47. Schmader T., Johns M., & Forbes C (2008). An integrated process model of stereotype threat effects on performance. Psychological Review, 115, 336–356. doi:10.1037/0033-295X.115.2.336 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Schmidt S. (2009). Shall we really do it again? The powerful concept of replication is neglected in the social sciences. Review of General Psychology, 13, 90–100. doi:10.1037/a0015108 [Google Scholar]
  49. Seibt B., & Förster J (2004). Stereotype threat and performance: How self-stereotypes influence processing by inducing regulatory foci. Journal of Personality and Social Psychology, 87, 38–56. doi:10.1037/0022-3514.87.1.38 [DOI] [PubMed] [Google Scholar]
  50. Spencer S. J., Logel C., & Davies P. G (2016). Stereotype threat. Annual Review of Psychology, 67, 415–437. doi:10.1146/annurev-psych-073115-103235 [DOI] [PubMed] [Google Scholar]
  51. Steele C. M., & Aronson J (1995). Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology, 69, 797–811. doi:10.1037/ 0022-3514.69.5.797 [DOI] [PubMed] [Google Scholar]
  52. Touryan S. R., Johnson M. K., Mitchell K. J., Farb N., Cunningham W. A., & Raye C. L (2007). The influence of self-regulatory focus on encoding of, and memory for, emotional words. Social Neuroscience, 2, 14–27. doi:10.1080/17470910601046829 [DOI] [PubMed] [Google Scholar]

Articles from The Journals of Gerontology Series B: Psychological Sciences and Social Sciences are provided here courtesy of Oxford University Press

RESOURCES