Perceived Demands Associated with Emotion Regulation Strategies Among Young and Cognitively Diverse Older Adults

Abstract

Emotion regulation (ER) is viewed as a cognitively demanding process, with strategies varying in demands. Individuals may prefer strategies perceived as lower in cognitive demands, and selecting low-demand strategies may be particularly adaptive for those with limited cognitive resources. We examine how ER strategies differ in perceived cognitive demands and how perceived demands predict strategy selection and well-being among regulators of varying age and cognitive status. Young adults (aged 21–34, n=66), cognitively normal older adults (CN; aged 70–83, n=90), and older adults with mild cognitive impairment (MCI; aged 70–84, n=60) reported perceived demands and use of ten ER strategies. As expected, early-acting strategies (e.g., situation selection) were generally viewed as less demanding than later-acting strategies (e.g., masking). Younger adults reported higher cognitive demands and effort requirements compared with CN older adults. For younger adults and CN older adults (but not those with MCI), strategies perceived as less demanding were used more. Older (but not younger) adults who perceived ER to be more demanding experienced poorer well-being. Age-related differences in perceived cognitive demands suggest ER perceptions may change with gained life experience. However, MCI may create ER difficulties by interfering with one’s ability to select easier to implement strategies.

Introduction

Emotion regulation (ER) is theorized to require cognitive resources (Ochsner & Gross, 2005). Although past work has made inferences about how ER strategies vary in cognitive demands, it is unclear whether individuals perceive strategies as differentially demanding and whether these perceptions guide strategy selection. Given normative age-related declines in fluid cognitive ability (Salthouse, 2009), older adults may differ from younger individuals in their perceptions of ER demands and prefer strategies perceived as less demanding. Notably, there are individual differences within older adults in terms of fluid cognitive ability, with some showing little or no decline, while others experience significant decline (Reuter-Lorenz & Park, 2024). Little is known about ER processes in more cognitively impaired older adults, but cognitive deficits are expected to create difficulty selecting appropriate strategies (Pruessner et al., 2020). The present study examines how young adults, cognitively normal (CN) older adults, and older adults with mild cognitive impairment (MCI) vary in their perceptions of ER demands, as well as how perceived ER demands relate to frequency of ER strategy use across these groups.

Cognitive Demands of ER Strategies

ER strategies enacted earlier in the emotion generative process have been proposed to be lower in effort requirements compared with those enacted after emotions have fully taken hold (Sheppes & Gross, 2011). According to the process model of ER (Gross, 2015), strategies can be grouped based on where they exist on a continuum from early-acting to later-acting strategies. The earliest acting type of strategy is situation selection, which involves engaging in or avoiding situations to meet emotional goals. Within a situation, one can engage in situation modification or altering aspects of the situation. Attentional deployment entails directing attention toward or away from aspects of a situation based on one’s emotional goals. Cognitive change involves thinking about a situation in a way that changes its meaning. Finally, response modulation entails direct attempts to alter aspects of one’s emotional response (e.g., expression).

Although little is known about perceived ER demands, one can make inferences from studies examining how cognitive ability and contextual demands relate to strategy preferences. In laboratory tasks, fluid cognitive ability has been associated with higher use of ER strategies involving engaging cognitive resources toward the source of emotions being regulated (Growney & English, 2023) and a stronger preference for reappraisal over distraction (Scheibe et al., 2015). Because regulators are theorized to select strategies consistent with their available resources (Urry & Gross, 2010), these finding may reflect higher cognitive demands associated with engagement or cognitive change strategies. Prior work has focused on a limited set of strategies and, rather than directly assessing perceived demands, has relied on inferences about perceptions through observing aspects of ER (e.g., strategy selection). Additionally, ER perceptions may vary by individual characteristics such as age, cognitive status, and emotional well-being.

Preference for Less Demanding ER Strategies

Perceptions of cognitive demands can have consequences for willingness to engage in tasks (Hess, 2014). The Selection and Optimization with Compensation–Emotion Regulation model (SOC-ER; Urry & Gross, 2010) describes how individuals select ER strategies that are consistent with their available resources. These resources may be influenced by characteristics of an individual, such as fluid cognitive ability. During strategy selection, a regulator may consider several factors including goals, context, and effort requirements associated with different strategies (for review, see Matthews et al., 2021).

Several laboratory studies have found that individuals prefer to select strategies that are easier to implement. In one study, when viewing high-arousal negative images, participants were more likely to choose reappraisal over no regulation if told to use a lower-demand versus higher-demand version of reappraisal (rethinking as fake versus rethinking as positive) (Milyavsky, 2019). Another study found that providing concrete distractions or reappraisals (rather than relying on self-generation) increased reappraisal use, potentially due to decreased demands (Sheppes et al., 2014). An experience sampling study of daily life found that both young and older adults reported using situation selection and modification most often, followed by attentional deployment, cognitive change, and response modulation strategies (Livingstone & Isaacowitz, 2021). Taken together, these studies suggest that individuals prefer strategies lower in cognitive demand or enacted earlier in the emotion generative process over those higher in cognitive demand or enacted later in the emotion generative process. Past work, however, has not examined how perceptions of strategies relate to their use in a systematic fashion.

Roles of Age and Cognitive Status

Older adults incur increased costs associated with effortful cognitive engagement. According to Selective Engagement Theory (Hess, 2014), age-related increases in costs lead to reduced motivation to engage in demanding activities. Accordingly, older adults may experience higher costs associated with ER and their strategy use may be particularly driven by perceived demands. Further, older adults may experience disproportionate demands associated with later-acting strategies than earlier acting strategies. Because older adults are theorized to selectively use strategies that help them avoid experiencing high-arousal negative emotions (i.e., earlier-acting strategies) (Charles, 2010), they may be less likely to have as much recent experience implementing and gathering information about effectiveness or demands of later-acting strategies. Furthermore, because there are age-related differences in recovery from physiological reactivity (Charles, 2010), once high-arousal negative emotion is elicited, it may take longer for older adults to down-regulate compared with younger adults.

Despite normative declines in fluid cognitive ability (Salthouse, 2009), older adults tend to maintain or even improve their emotional well-being (Carstensen et al., 2011). As proposed by the SOC-ER model (Urry & Gross, 2010), older adults may maintain emotional well-being by selecting and optimizing specific ER strategies that are consistent with their resources. Similarly, the Strength and Vulnerability Integration framework (SAVI; Charles, 2010) suggests older adults rely on resources that allow them to avoid situations eliciting high-arousal negative emotions and direct attention toward positive information to avoid engaging with stimuli that may evoke negative emotions to be regulated. Through selectivity, older adults may shape their emotional lives in such a way that they avoid the need to implement later-acting strategies.

In laboratory studies, older adults choose distraction over reappraisal to a greater extent than younger adults (Scheibe et al., 2015), perhaps reflecting older adults’ preference for a lower-demand strategy or desire to quickly improve negative emotional states. However, there is little evidence that young and older adults differ in their ER strategy use (for review, see Allen & Windsor, 2019). One potential explanation for the lack of clear age differences in ER strategy preferences is that individuals may select strategies based on their own unique experience of demands with various strategies.

Because age-related differences in perceived cognitive demands of ER are theorized to be driven by differences in cognitive resources, effects might be magnified in older adults with particularly low cognitive capacity, such as those with MCI. Older adults with MCI experience cognitive decline beyond that which is considered part of normal aging, but have largely intact functioning (Petersen, 2004). Although ER processes among older adults with MCI are not well-understood, their relatively poor emotional well-being in comparison to CN older adults (Bárrios et al., 2013) may reflect ER difficulty or dysfunction. Older adults with MCI may perceive strategies as generally more demanding and gravitate more toward strategies they view as easier to implement. Alternatively, they may have difficulty selecting strategies that are consistent with their available resources. That is, given that selecting and optimizing the use of adaptive, low-demand strategies may itself be cognitively demanding (Pruessner et al., 2020), older adults with MCI may experience more difficulty navigating this process.

The Present Study

We assessed perceptions and use of ten ER strategies spanning the process model among young adults, CN older adults, and older adults with MCI. Participants completed a questionnaire about how often they use each strategy and different aspects of the workload they experience when using each strategy, including their perceived cognitive demands (i.e., mental workload), effort requirements (i.e., energy or exertion required), and physical demands (i.e., physical workload). Older adults’ reduced physiological flexibility can create difficulty dealing with high-arousal emotions (Charles, 2010), which may contribute to heightened perception of physical demands when regulating. Older adults with MCI also tend to experience neuroendocrine dysfunction (Ho et al., 2020), which could further exacerbate physical demands of ER.

We were primarily interested in how perceived cognitive demands and effort requirements (1) varied across ER strategies, (2) varied across groups, and (3) predicted frequency of use across groups. Our hypotheses were: (H1) Individuals will view early-acting strategies as less cognitively demanding and requiring of less effort than later-acting ER strategies, (H2a) Older adults will report higher perceived cognitive demands and effort requirements compared with young adults, (H2b) Age differences in perceived cognitive demands and effort requirements will be higher with later-acting than early-acting strategies, (H3a) Individuals in general will report using strategies less frequently that they rate as higher in cognitive demands and effort requirements, and (H3b) The negative association between perceived demands and frequency of use will be present among older adults but not younger adults, or stronger among older adults than younger adults. Additionally, we explored whether age effects vary for older adults with MCI compared to CN older adults, and whether effects extended to perceived physical demands.

Finally, we tested whether individuals with lower well-being perceive ER as more demanding, especially among older adults. Experiencing the use of ER strategies as higher in cognitive demands and effort requirements may signal difficulty managing one’s emotions, particularly for those with more limited cognitive resources to overcome these demands.

Methods

Participants

Participants were 216 individuals recruited from the St. Louis, MO community through phone calls and letters to individuals in our target age groups. We also posted flyers in the community and on databases of research participants. Participants completed the telephone Mini-Mental State Exam and we used the recommended cutoff of scoring at least 22 to be eligible (Newkirk et al., 2004). The sample was 55.1% women, 43.5% men, and 1.4% other gender. In terms of race, 67.1% identified as White, Caucasian, or European American, 26.4% as Black or African American, 2.3% as Asian, Asian American, or Pacific Islander, 2.3% as Hispanic or Latino, 0.5% as American Indian or Alaska Native, 0.5% as Middle Eastern or Arab American, and 1.9% as other race or ethnicity (multiple category selection possible). The sample included 66 young adults (aged 21–34, M=27.15, SD=3.90), 90 cognitively normal older adults (CN older: aged 70–83, M=75.14, SD=3.79), and 60 older adults with mild cognitive impairment (MCI older: aged 70–84, M=77.05, SD=4.50). Sample size was determined based on a power analysis conducted for a larger portion of this study. Additionally, we conducted a power analysis in G*Power 3.1 (Faul et al., 2009) and determined that a sample size of 129 was required to detect within-between interactions with a small effect size. We present additional details about the sample and power analysis in supplemental materials.

Measures

Emotion Regulation Load Index

Participants completed the Emotion Regulation Load Index, a questionnaire developed for use in the present study based on the NASA Task Load Index (TLX; Hart & Staveland, 1988), a measure validated to assess perceived demands and adapted for use in a variety of contexts. Participants rated strategies in terms of their cognitive demands (“How mentally demanding is it for you to…”), physical demands (“How physically demanding is it for you to…”), and effort requirements, (“How hard do you have to work to…”) on a scale of 1 = Not at all to 7 = Very. Participants also reported the frequency with which they typically attempt to implement each strategy (“How often do you try to…”) on a scale of 1 = Not at all to 7 = Very often.

These four questions were asked for each of ten strategies spanning the process model of ER. Situation-focused strategies included behavioral avoidance (“…avoid unpleasant situations.”), situation selection (“…seek out pleasant situations.”), and situation modification (“…take action to change unpleasant situations.”). Attentional deployment strategies included savoring (“…savor pleasant experiences.”) and cognitive distraction (“…shift your attention (or distract yourself) during unpleasant situations.”). Cognitive change strategies included positive reappraisal (“…think about unpleasant situations in a more positive way.” and detached reappraisal (“…think about unpleasant situations objectively (or in a detached way).”). Response modulation strategies included expressive suppression (“…hide your emotional expressions.”), masking (“…show an emotion other than what you are feeling.”), and acceptance (“…accept unpleasant experiences without trying to change them.”).

Well-Being

Participants completed trait-level measures of well-being. Depressive symptoms were assessed using the 20-item Center for Epidemiologic Studies Depression – Revised Scale (CESD-R; Eaton et al., 2004). Perceived stress was assessed using the 10-item Perceived Stress Scale (PSS; Cohen & Williamson, 1988). Due to collinearity concerns, we examined these indices in separate models because they were highly correlated (r=.69, p<.001). We also Z-scored and then averaged the two scores to create a well-being composite, reported in supplemental materials.

Analytic Plan

Our analyses involved a series of repeated measures ANOVAs, examining group membership (younger adults, CN older adults, MCI older adults), strategy, and their interaction predicting perceptions of strategies. We report findings with the Greenhouse-Geisser correction because the assumption of sphericity was violated in these models. When there was a significant interaction, we conducted post-hoc Fisher’s LSD pairwise comparisons with Bonferroni corrections (for details, see supplemental materials).

To examine links between perceptions and frequency of use, we first computed within-person and between-person correlations, then conducted no-intercept multilevel models which estimate a unique intercept and slope for each group (i.e., young adults, CN older adults, MCI older adults) within the same model. We chose this approach given our interest in understanding perceptions within group while minimizing the likelihood of Type 1 error. It allows us to examine effects within age group and conduct post-hoc tests of general linear hypotheses examining overall slopes (interpreted as main effects) and comparing slopes between groups (interpreted as interactions between group and the predictor variable).

At the within-person level, we examine the extent to which an individual’s perception of a strategy’s demands (compared with their perceptions of other strategies’ demands) is associated with their frequency of using that strategy. At the between-person level, we examine the extent to which an individual’s average level of perceived demands across all strategies (compared with other individuals’ average level of perceived demands) is associated with frequency of using strategies in general. We estimate these separately for young adults, CN older adults, and MCI older adults:

$$\begin{gathered} {\text{Frequency of use}}_{ij}=\text{β}1\left({\text{Young}}_i\right)+\text{β}2\left({\text{CN}}_i\right)+\text{β}3\left({\text{MCI}}_i\right) \\ +\text{β}4\left({\text{Young}}_i{\times \text{Demands}}_{ij}\right)+\text{β}5\left({\text{CN}}_i\times {\text{Demands}}_{ij}\right)+\text{β}6\left({\text{MCI}}_i\times {\text{Demands}}_{ij}\right) \\ +\text{β}7\left({\text{Young}}_i{\times \text{Demands}}_i\right)+\text{β}8\left({\text{CN}}_i{\times \text{Demands}}_i\right)+\text{β}9\left({\text{MCI}}_i\times {\text{Demands}}_{ij}\right) \\ +u_{0i}+u_{1i}\left({\text{Demands}}_{ij}\right)+{ϵ}_{ij} \end{gathered}$$

Using post-hoc simultaneous tests of general linear hypotheses with Holm adjustments, we also examined the overall within-person and between-person slopes for each model (i.e., collapsing over all three groups) and compared slopes across groups to see whether associations between perceptions and frequency significantly differed by group.

Finally, we ran no-intercept multilevel models examining trait-level well-being measures as predictors of ER perceptions. At the between-person level, we examine the extent to which an individual’s well-being is associated with their perceived ER demands. We estimate these separate for young adults CN older adults, and MCI older adults:

$$\begin{gathered} {\text{Demands}}_{ij}=\beta 1\left({\text{Young}}_i\right)+\text{β}2\left({\text{CN}}_i\right)+\text{β}3\left({\text{MCI}}_i\right) \\ +\text{β}4\left({\text{Young}}_i{\times \text{Well-Being}}_i\right)+\text{β}5\left({\text{CN}}_i{\times \text{Well-Being}}_i\right)+\text{β}6\left({\text{MCI}}_i{\times \text{Well-Being}}_i\right) \\ +u_{0i}+{ϵ}_{ij} \end{gathered}$$

Results

Preliminary Findings

Means, standard deviations, between-person correlations, and within-person correlations among study variables are presented in supplemental materials. Ratings of effort and cognitive demands were generally not highly correlated across strategies (correlations ranged from r=.38 to .57), so we examined these items separately.¹

Do Perceived Demands Differ Across Strategies and Groups?

Figure 1a displays mean ratings by group for overall perceptions of ER cognitive demands, effort requirements, and physical demands collapsed across strategies. Figure 1b displays mean perceptions of ER cognitive demands, effort requirements, and physical demands for each ER strategy assessed, collapsed across group.

Figure 1.

(A) Perceived cognitive demands, physical demands, and effort associated with emotion regulation (ER), collapsed over all 10 assessed strategies. (B) Perceived cognitive demands, effort requirements, and physical demands associated with different ER strategies, collapsed across participant groups.

Perceptions of Cognitive Demands

We hypothesized individuals would view early-acting strategies as less cognitively demanding than later-acting strategies (H1). There was a main effect of strategies for perceived cognitive demands, F(6.93, 1420.20)=29.99, p<.001, η²=.13, with a large effect size. Acceptance was rated highest in cognitive demands (M=4.43, SD=1.80) and savoring was rated as lowest (M=2.84, SD=2.04). In line with our hypothesis, the strategies perceived as lower in cognitive demands tended to be early-acting (e.g., savoring, situation selection, behavioral avoidance) while the ones rated as higher in cognitive demands tended to be later-acting (e.g., masking, expressive suppression, acceptance).

We hypothesized that older adults would perceive higher cognitive demands compared with young adults (H2a). Perceived cognitive demands varied significantly by group, F(2, 205)=4.25, p=.015, η²=.003, with a small effect size. In contrast with our hypothesis, post-hoc comparisons revealed that young adults viewed ER strategies as more cognitively demanding than CN older adults (Mean Difference=0.49, SE=0.17, p=.005, 95% CI[0.15, 0.83]). Perceived cognitive demands reported by MCI older adults did not differ significantly from young adults or CN older adults (ps>.09).

We hypothesized that age group differences in perceived cognitive demands would be higher with later-acting than earlier-acting strategies (H2b). However, there was not a significant Group X Strategy interaction, F(13.86, 1420.20)=1.56, p=.086, η²=.01, indicating that the group difference in perceived cognitive demands did not vary by strategy.

Perceptions of Effort Requirements

We hypothesized that individuals would view early-acting strategies as requiring less effort than later-acting strategies (H1). There was a main effect of strategy type on ER effort, F(6.95, 1382.83)=23.92, p<.001, η²=.11, with a medium-large effect size. As expected, perceptions of effort requirements tended to be lowest in early-acting strategies and highest in later-acting strategies, with acceptance rated the highest in effort requirements (M=4.41, SD=1.82) and savoring rated as the lowest (M=2.90, SD=1.96).

We hypothesized that older adults would perceive higher effort requirements compared with young adults (H2a). There was a main effect of group on ER effort, F(2, 199)=3.30, p=.039, η²=.002, with a small effect size. In contrast with our hypothesis, young adults reported that deploying ER strategies was more effortful compared to CN older adults, Mean Difference=0.46, SE=0.18, p=.033. The perceived effort requirements of older adults with MCI did not differ significantly from younger adults or CN older adults (ps>.64).

We hypothesized that age group differences in perceived effort requirements would be higher with later-acting than with earlier-acting strategies (H2b). The Strategy X Group interaction, F(13.90, 1382.83)=1.66, p=.060, η²=.02, was not significant, indicating that the group difference in effort requirements did not vary by strategy.

Perceptions of Physical Demands

We explored whether strategies and groups differed in perceived physical demands. Physical demands varied significantly by strategy, F(6.86 1358.00)=11.45, p<.001, η²=.05, with a medium effect size. Masking was rated as highest in physical demands (M=3.50, SD=1.94), and savoring was rated as lowest (M=2.39, SD=1.91).

Perceived physical demands also varied significantly by group, F(2, 198)=8.76, p<.001, η²=.01, with a small effect size. Older adults with MCI perceived higher physical demands compared with both young adults, Mean Difference=0.89, SE=0.22, p<.001, and CN older adults, Mean Difference=0.64, SE=0.21, p=.007.

There was also a Strategy X Group interaction, indicating that group differences in physical demands varied significantly by strategy, F(13.72, 1358.00)=3.83, p<.001, η²=.04, with a small-medium effect size. Older adults with MCI perceived relatively higher physical demands specifically for cognitive distraction, positive reappraisal, and detached reappraisal. For situation modification, older adults with MCI perceived higher physical demands compared with CN older adults, but not young adults. For acceptance, older adults with MCI and CN older adults perceived higher physical demands compared with young adults.

Links Between ER Perceptions and Frequency of Use

Analyses examining group differences in frequency of ER strategy use are presented in supplemental materials. Within a given strategy, perceptions of demands were generally not associated with frequency of use, as shown in supplemental materials. However, such correlations do not consider how a participant views and uses a strategy in relation to other strategies. Results from no-intercept models and post-hoc simultaneous tests of general linear hypotheses are presented in Table 1.

Table 1.

Results from No-Intercept Multilevel Models with ER Perceptions Predicting Frequency of Use

	Perceived Cognitive Demands		Perceived Physical Demands		Perceived Effort Requirements
	Est.	SE	Est.	SE	Est.	SE
Intercept: Young Adults	4.88	0.10	4.91	0.10	4.88	0.10
ER perception (w)	–0.20***	0.05	–0.03	0.07	–0.26***	0.06
ER perception (b)	0.01	0.09	0.08	0.09	–0.05	0.09
Intercept: CN Older Adults	4.74	0.08	4.72	0.08	4.70	0.08
ER perception (w)	–0.26***	0.05	–0.18**	0.06	–0.26***	0.05
ER perception (b)	0.18*	0.07	0.16*	0.06	0.03	0.08
Intercept: MCI Older Adults	4.70	0.10	4.67	0.11	4.72	0.10
ER perception (w)	–0.07	0.06	–0.13	0.07	–0.23***	0.07
ER perception (b)	0.19*	0.10	0.11	0.08	0.19*	0.09
Post-Hoc Simultaneous Tests of General Linear Hypotheses
Overall Slope: ER perception (w)	–0.18***	0.03	–0.12**	0.04	–0.25***	0.03
Overall Slope: ER perception (b)	0.12*	0.05	0.12*	0.04	0.06	0.05
Slope Comparison: ER perception (w)	0.16*	0.07	0.08	0.04	0.02	0.05
Slope Comparison: ER perception (b)	–0.18	0.11	–0.04	0.05	0.13	0.07

Note. Table presents models from three separate models examining perceived cognitive demands, physical demands, and effort requirements predicting how often participants reported using ER strategies. Four post-hoc simultaneous tests of general linear hypotheses were conducted per model, and effects are adjusted for multiple tests using the Holm method. In the Perceived Cognitive Demands model, Slope Comparison (w) compares the within-person slope for older adults with MCI to the within-person slope for the other two groups and Slope Comparison (b) compares the between-person slope for young adults with the between-person slope for the other two groups. In the Perceived Physical Demands model, Slope Comparison (w) and (b) both compare slopes for young adults with CN older adults. In the Perceived Effort Requirements model, Slope Comparison (w) and (b) both compare slopes for MCI older adults with slopes for the other two groups. Within-person effects = (w) and between-person effects = (b). p < .05,

^**p < .01,

^***p < .001

Perceptions and Frequency of Use by Group

We hypothesized that individuals in general will report using strategies less frequently that they rate as higher in cognitive demands and effort requirements (H3a), and that this negative association between perceived demands and frequency of use will be present among older adults but not younger adults, or stronger among older adults than younger adults (H3b).

A test of the overall slope of within-person cognitive demands revealed that participants used strategies more often that they rated as lower in cognitive demands, b= –0.18(0.03), p<.001, consistent with H3a. However, in contrast with H3b, the effect was present for young adults, b= –0.20(0.05), p<.001, and CN older adults, b= –0.26(0.05), p<.001, but not MCI older adults, b= –0.07(0.06), p=.220, and a test comparing slopes revealed that the effect of cognitive demands on frequency of use was significantly stronger among young adults and CN older adults than among older adults with MCI, b= 0.16(0.07), p=.049. We also explored between-person effects: participants who rated strategies as higher in cognitive demands on average reported using ER strategies more often, b = –0.37 (0.15), p = .040, and this between-person effect did not vary significantly by group, b= –0.18(0.11), p=.101.

Similarly, a test of the overall slope of within-person effort requirements revealed that participants used strategies more often that they rated as lower in effort requirements, b= –0.25(0.03), p<.001, consistent with H3a. However, inconsistent with H3b, this effect was significant among all groups: young adults, b= –0.26(0.06), p<.001, CN older adults, b= –0.26(0.05), p<.001, and MCI older adults, b= –0.23(0.09), p<.001, and the test comparing the slopes was not significant (p=.707). We explored between-person effects: overall, participants who regulated their emotions more often did not vary in their average rating of strategy effort requirements, b= 0.17(0.15), p=.241. However, MCI older adults who rated effort requirements as higher on average tended to use ER strategies more often, b= 0.19(0.09), p=.032; this association was not significant among young adults or CN older adults (ps>.58).

We also explored these questions with perceived physical demands. A test of the overall slope of the within-person effect of physical demands revealed participants used strategies more often that they rated as lower in physical demands, b= –0.35(0.11), p=.004. This effect varied by group such that it was only significant among CN older adults, b= –0.19(0.06), p=.003 (not MCI older adults or younger adults, ps>.06). At the between-person level, participants who used strategies more often rated strategies as higher in physical demands, b= 0.35(0.14), p=.010. This effect was also significant among CN older adults, b= 0.16(0.06), p=.012, but not MCI older adults or young adults (ps>.20).

Links Between Well-Being and ER Strategy Perceptions

Finally, we explored associations between well-being and strategy perceptions. We present results from no-intercept models predicting ER perceptions from well-being indices and group membership in Tables S10–S12. With models examining depressive symptoms as a predictor, there was an overall slope predicting cognitive demands, b=0.02(0.01), p=.002, and effort requirements, b=0.02(0.01), p=.041, but not physical demands, b=0.02(0.01), p=.060. However, examining stress as a predictor, those who reported higher stress perceived higher cognitive demands, b=0.46(0.11), p<.001, effort requirements, b=0.37(0.11), p=.003, and physical demands, b=0.54(0.13), p<.001, associated with ER strategies. Regarding cognitive demands, effects of both well-being indices varied by group. Associations with depressive symptoms were significant among older adults with MCI, b=0.05(0.02), p=.010, but not among CN older adults, b=0.02(0.01), p=.106, or young adults, b=0.01(0.01), p=.300, whereas associations with perceived stress were significant among CN older adults, b=0.47(0.19), p=.013, and older adults with MCI, b=0.63(0.22), p=.005, but not young adults, b=0.29(0.17), p=.088. In sum, effects were stronger among older adults than younger adults, and stress was more consistently linked with ER perceptions than was depressive symptoms.

Discussion

In the present study, young adults, CN older adults, and older adults with MCI reported on their perceptions and use of ER strategies. We examined perceived demands associated with different strategies, group differences in perceived demands, and relationships between perceived demands and frequency of use. We also explored associations with well-being.

Early-Acting ER Strategies Viewed as Less Demanding than Later-Acting ER Strategies

Based on theoretical reasoning (Sheppes & Gross, 2011), we expected that early-acting ER strategies would be viewed as less cognitively demanding than later-acting ER strategies, and generally found this to be the case. Strategies rated as lowest in cognitive demands and effort requirements—seeking out pleasant situations and savoring pleasant experiences—were those that do not specifically entail dealing with negative emotions. Individuals may view enhancing or up-regulating positive emotions as less demanding than down-regulating negative emotions, as the former typically does not involve emotions inconsistent with one’s emotional goals.

Attentional deployment strategies (aside from savoring) and cognitive change strategies generally fell into the middle range of perceived demands, with situation selection and response modulation strategies respectively falling at the low and high end of perceived demands. Surprisingly, cognitive distraction and detached reappraisal, the two strategies most frequently pitted against one another in laboratory tasks examining ER choice (Sheppes et al., 2014), did not differ significantly in their ratings of cognitive demands, physical demands, or effort requirements. It may be easier to shift attention away from negative stimuli in laboratory settings than in daily life, where negative information is often more self-relevant.

The three response modulation strategies were generally rated as highest in cognitive demands, effort requirements, and physical demands. These later-acting strategies are enacted after emotions fully take hold, with expressive suppression and masking viewed as putatively maladaptive (Aldao et al., 2010). High demands associated with acceptance may reflect the perceived difficulty that comes with a situation over which one does not have control or demands associated with experiencing negative emotions and not regulating.

Group Differences in Perceived ER Demands

Given age-related differences in fluid cognitive ability, we expected that older individuals would perceive higher demands associated with ER than younger individuals (Hess, 2014). However, we found that young adults perceived higher ER mental demands and effort requirements compared with CN older adults. There are several potential reasons for this group difference. First, engaging in ER may generally be a more well-practiced behavior among CN older adults than young adults, and CN older adults’ expertise may result in lower perceived demands. Second, older adults experience fewer stressors compared with younger adults (Stawski et al., 2008), and there may be objective differences in demands therein. Finally, retired older adults may have more control over situations and fewer time-sensitive responsibilities compared with younger adults, who may be more likely to struggle with balancing ER strategy deployment with other necessary tasks. Older adults with MCI did not differ significantly from other groups in overall perceived cognitive demands and effort requirements. This intermediary position may reflect that on average, those with MCI experience decreases in the efficiency of ER processes, but they may retain some advantages gained from life experience.

Older adults with MCI reported higher physical demands for ER compared with young adults and CN older adults. Surprisingly these heightened physical demands were not specific to strategies that inherently involve physicality (i.e., situation-focused and response modulation). Rather, older adults with MCI reported higher physical demands with strategies spanning the process model of ER. This finding may reflect overall reductions in physiological flexibility that lead cognitively impaired older adults to experience ER difficulties (Charles, 2010).

ER Demands are Negatively Tied to Frequency of ER Strategy Use and Well-Being

In support of our hypothesis based on the idea that individuals prefer to regulate using strategies that they view as easier to implement (e.g., Milyavsky, 2019; Sheppes et al., 2014), participants reported higher use of strategies they rated as relatively less demanding. We note that there are several other possible interpretations of this finding, First, strategies may become easier to implement with practice. Second, individuals may experience greater success implementing strategies that are less demanding and select them more often due to their expertise or wisdom. We expected older adults would be more likely than younger adults to use strategies they view as lower in cognitive demands. Consistent with expectations, the preference for strategies viewed as less demanding was most apparent among CN older adults. Older adults with MCI did not demonstrate selective use of strategies they rated as lower in cognitive or physical demands. Future work is needed to determine whether their level of executive functioning interferes with the ability to efficiently process available resources for strategy selection or benefit from practice as much as CN individuals (Machulda et al., 2013). Interestingly, the average perceived demands of a given strategy (e.g., expressive suppression) across the sample was generally not associated with frequency of its use. Instead, frequency of use was associated with a person’s own unique perception of the strategy’s demands in relation to their perception of other strategies. Perceived demands may therefore be best viewed as a person-specific factor when considering its contribution to ER choice.

The relationship between perceived demands and strategy use also varied by group differently at the within-person and between-person level. Young and CN older adults used strategies less that they perceived as more demanding than other strategies, but there were individual differences among older adults—both CN and MCI—suggesting that those who generally view ER strategies as higher in demands regulate more often than those who view ER strategies as less demanding. Accordingly, older adults who regulate more frequently may be those who are less effective in leveraging life experience to facilitate effective regulation efforts. Those who view ER strategies as typically high in demands may have developed this perception due to repeated failed attempts to regulate. Consistent with this idea, higher perceived ER demands was also generally associated with lower well-being, especially among older adults. Younger adults reported relatively high cognitive demands associated with ER strategies regardless of their well-being levels. Older individuals who experience high-intensity negative emotions accompanying poor well-being may also perceive ER as more demanding, as difficulty dealing with these emotions is theorized to increase with age (Charles, 2010).

Limitations, Future Directions, and Conclusions

One limitation of the present study is that our focus on trait-level assessments of ER strategy use and perceptions did not allow for consideration of context (e.g., emotion being regulated, physical setting). Cognitive demands and effort requirements associated with implementing a given strategy likely vary based on environmental affordances. Notably, age-related differences in emotional experiences and situational contexts likely affected participants’ reports of perceived demands. Experience sampling approaches may be useful in examining how age and context together play roles in perceived regulation demands, as well as identifying causal links between ER and emotional well-being. We found that associations between stress and perceived ER demands were stronger than associations between depressive symptoms and perceived ER demands. The robustness of associations between well-being and ER demands should be tested across different measures of well-being as well as among clinical different samples (e.g., individuals with anxiety or major depressive disorder). We caution against inferring causality from these cross-sectional findings. Age-related differences in ER demands should also be considered at different stages of the process model. Demands may be high when determining which strategy to select and switching from one strategy to another (Pruessner et al., 2020). To better understand the role of cognition in ER, future work should consider the interplay among potential determinants, including the regulator’s fluctuating cognitive resources, the regulator’s perceived demands associated with different ER components, and the environmental affordances or aspects of the regulation context that inherently lend themselves to implementing specific strategies.

This study extends understanding of ER processes, highlighting the potential role of perceived demands in strategy selection. Younger regulators perceived higher cognitive demands and effort requirements compared with CN older adults, but perceived demands of a strategy were only consistently associated with lower use among CN older adults. Rather than viewing strategy groups as inherently demanding or not, researchers should consider regulators’ unique perceptions of cognitive demands.