Comparisons of Self-Report With Objective Measurements Suggest Faster Responding but Little Change in Response Quality Over Time in Ecological Momentary Assessment Studies.

Response times (RTs) to ecological momentary assessment (EMA) items often decrease after repeated EMA administration, but whether this is accompanied by lower response quality requires investigation. We examined the relationship between EMA item RTs and EMA response quality. In one data set, declining response quality was operationalized as decreasing correspondence over time between subjective and objective measures of blood glucose taken at the same time. In a second EMA study data set, declining response quality was operationalized as decreasing correspondence between subjective ratings of memory test performance and objective memory test scores. We assumed that measurement error in the objective measures did not increase across time, meaning that decreasing correspondence across days within a person could be attributed to lower response quality. RTs to EMA items decreased across study days, while no decrements in the mean response quality were observed. Decreasing EMA item RTs across study days did not appear problematic overall.

Can you tell people's cognitive ability level from their response patterns in questionnaires?

Questionnaires are ever present in survey research. In this study, we examined whether an indirect indicator of general cognitive ability could be developed based on response patterns in questionnaires. We drew on two established phenomena characterizing connections between cognitive ability and people's performance on basic cognitive tasks, and examined whether they apply to questionnaires responses. (1) The worst performance rule (WPR) states that people's worst performance on multiple sequential tasks is more indicative of their cognitive ability than their average or best performance. (2) The task complexity hypothesis (TCH) suggests that relationships between cognitive ability and performance increase with task complexity. We conceptualized items of a questionnaire as a series of cognitively demanding tasks. A graded response model was used to estimate respondents' performance for each item based on the difference between the observed and model-predicted response ("response error" scores). Analyzing data from 102 items (21 questionnaires) collected from a large-scale nationally representative sample of people aged 50+ years, we found robust associations of cognitive ability with a person's largest but not with their smallest response error scores (supporting the WPR), and stronger associations of cognitive ability with response errors for more complex than for less complex questions (supporting the TCH). Results replicated across two independent samples and six assessment waves. A latent variable of response errors estimated for the most complex items correlated .50 with a latent cognitive ability factor, suggesting that response patterns can be utilized to extract a rough indicator of general cognitive ability in survey research.

Cognitive Functioning and the Quality of Survey Responses: An Individual Participant Data Meta-Analysis of 10 Epidemiological Studies of Aging.

OBJECTIVES: Self-reported survey data are essential for monitoring the health and well-being of the population as it ages. For studies of aging to provide precise and unbiased results, it is necessary that the self-reported information meets high psychometric standards. In this study, we examined whether the quality of survey responses in panel studies of aging depends on respondents' cognitive abilities. METHODS: Over 17 million survey responses from 157,844 participants aged 50 years and older in 10 epidemiological studies of aging were analyzed. We derived 6 common statistical indicators of response quality from each participant's data and estimated the correlations with participants' cognitive test scores at each study wave. Effect sizes (correlations) were synthesized across studies, cognitive tests, and waves using individual participant data meta-analysis methods. RESULTS: Respondents with lower cognitive scores showed significantly more missing item responses (overall effect size ρ^ = -0.144), random measurement error (ρ^ = -0.192), Guttman errors (ρ^ = -0.233), multivariate outliers (ρ^ = -0.254), and acquiescent responses (ρ^ = -0.078); the overall effect for extreme responses (ρ^ = -0.045) was not significant. Effect sizes were consistent across studies, modes of survey administsration, and different cognitive functioning domains, although some cognitive domain specificity was also observed. DISCUSSION: Lower-quality responses among respondents with lower cognitive abilities add random and systematic errors to survey measures, reducing the reliability, validity, and reproducibility of survey study results in aging research.

Attrition from longitudinal ageing studies and performance across domains of cognitive functioning: an individual participant data meta-analysis.

OBJECTIVES: This paper examined the magnitude of differences in performance across domains of cognitive functioning between participants who attrited from studies and those who did not, using data from longitudinal ageing studies where multiple cognitive tests were administered. DESIGN: Individual participant data meta-analysis. PARTICIPANTS: Data are from 10 epidemiological longitudinal studies on ageing (total n=209 518) from several Western countries (UK, USA, Mexico, etc). Each study had multiple waves of data (range of 2-17 waves), with multiple cognitive tests administered at each wave (range of 4-17 tests). Only waves with cognitive tests and information on participant dropout at the immediate next wave for adults aged 50 years or older were used in the meta-analysis. MEASURES: For each pair of consecutive study waves, we compared the difference in cognitive scores (Cohen's d) between participants who dropped out at the next study wave and those who remained. Note that our operationalisation of dropout was inclusive of all causes (eg, mortality). The proportion of participant dropout at each wave was also computed. RESULTS: The average proportion of dropouts between consecutive study waves was 0.26 (0.18 to 0.34). People who attrited were found to have significantly lower levels of cognitive functioning in all domains (at the wave 2-3 years before attrition) compared with those who did not attrit, with small-to-medium effect sizes (overall d=0.37 (0.30 to 0.43)). CONCLUSIONS: Older adults who attrited from longitudinal ageing studies had lower cognitive functioning (assessed at the timepoint before attrition) across all domains as compared with individuals who remained. Cognitive functioning differences may contribute to selection bias in longitudinal ageing studies, impeding accurate conclusions in developmental research. In addition, examining the functional capabilities of attriters may be valuable for determining whether attriters experience functional limitations requiring healthcare attention.

Daily sampling frequency and sampling duration affect reliability of person-level estimates of physical activity outcomes: Optimizing Ecological Momentary Assessment studies of physical activities.

Studies on the interrelationship between physical activity (PA) behaviors and EMA-assessed constructs should use measures with high reliability of both the EMA-assessed constructs and the time-matched accelerometry-assessed PA behavior. The aim of this paper is to evaluate how the reliability of accelerometry-assessed PA outcomes is affected by different EMA sampling schemes. Emulating relevant sampling schemes in EMA studies, multiple random samples of real-world accelerometer data (measured via activPAL worn for ∼7 days) were drawn that varied in the number of daily samples (3, 5, and 7 daily samples) and in the duration of each sample (5 min, 60 min, and 120 min), totaling 9 sampling schemes. The reliability of the resulting PA outcomes was estimated by correlating weekly aggregates of the sampled data with the true parameter values (weekly aggregates of all data). A total of 4231 days were analyzed from 619 participants (mean age 45 ± 16 years; 73.3 % female; 72.7 % Non-Hispanic White). At the lowest sampling duration (5 min), none of the sampling schemes reached the threshold for acceptable reliability (reliabilities ranged from 0.20 to 0.67). Using 120 min sample duration resulted in reliable person-level PA estimates (ranges 0.77 to 0.97) except for the time spent in sedentary behavior (rel = 0.68), and for minimum (rel = 0.66), and maximum (rel = 0.63) PA intensity. These three PA outcomes consistently yielded lower levels of reliability across all sampling schemes. Holding constant the total time covered in a day, sampling schemes that used more frequent samples with shorter duration resulted in greater reliability compared to sampling schemes that used less frequent samples with longer duration. Overall, the results are consistent with sampling theory - reliability increased as more time was covered. However, our results also suggested interactive effects of reliability based on sampling frequency, duration, and PA outcome. Researchers should be aware that the reliability of accelerometer-derived PA outcomes can vary depending on the sampling design of EMA studies.

Just-in-time adaptive ecological momentary assessment (JITA-EMA).

Interest in just-in-time adaptive interventions (JITAI) has rapidly increased in recent years. One core challenge for JITAI is the efficient and precise measurement of tailoring variables that are used to inform the timing of momentary intervention delivery. Ecological momentary assessment (EMA) is often used for this purpose, even though EMA in its traditional form was not designed specifically to facilitate momentary interventions. In this article, we introduce just-in-time adaptive EMA (JITA-EMA) as a strategy to reduce participant response burden and decrease measurement error when EMA is used as a tailoring variable in JITAI. JITA-EMA builds on computerized adaptive testing methods developed for purposes of classification (computerized classification testing, CCT), and applies them to the classification of momentary states within individuals. The goal of JITA-EMA is to administer a small and informative selection of EMA questions needed to accurately classify an individual's current state at each measurement occasion. After illustrating the basic components of JITA-EMA (adaptively choosing the initial and subsequent items to administer, adaptively stopping item administration, accommodating dynamically tailored classification cutoffs), we present two simulation studies that explored the performance of JITA-EMA, using the example of momentary fatigue states. Compared with conventional EMA item selection methods that administered a fixed set of questions at each moment, JITA-EMA yielded more accurate momentary classification with fewer questions administered. Our results suggest that JITA-EMA has the potential to enhance some approaches to mobile health interventions by facilitating efficient and precise identification of momentary states that may inform intervention tailoring.

Using Attributes of Survey Items to Predict Response Times May Benefit Survey Research.

Researchers have become increasingly interested in response times to survey items as a measure of cognitive effort. We used machine learning to develop a prediction model of response times based on 41 attributes of survey items (e.g., question length, response format, linguistic features) collected in a large, general population sample. The developed algorithm can be used to derive reference values for expected response times for most commonly used survey items.

Virtual Human-Delivered Interviews for Patients With Chronic Pain: Feasibility, Acceptability, and a Pilot Randomized Trial of Standard Medical, Psychosocial, and Educational Interviews.

OBJECTIVE: Seminal advances in virtual human (VH) technology have introduced highly interactive, computer-animated VH interviewers. Their utility for aiding in chronic pain care is unknown. We developed three interactive telehealth VH interviews-a standard pain-focused, a psychosocial risk factor, and a pain psychology and neuroscience educational interview. We then conducted a preliminary investigation of their feasibility, acceptability, and efficacy. We also experimentally compared a human and a computer-generated VH voice. METHODS: Patients ( N = 94, age = 22-78 years) with chronic musculoskeletal pain were randomly assigned to the standard ( n = 31), psychosocial ( n = 34), or educational ( n = 29) VH interview and one of the two VH voices. Acceptability ratings included patient satisfaction and expectations/evaluations of the VH interview. Outcomes assessed at baseline and about 1-month postinterview were pain intensity, interference, emotional distress, pain catastrophizing, and readiness for pain self-management. Linear mixed-effects models were used to test between- and within-condition effects. RESULTS: Acceptability ratings showed that satisfaction with the VH and telehealth format was generally high, with no condition differences. Study attrition was low ( n = 5). Intent-to-treat-analyses showed that, compared with the standard interview, the psychosocial interview yielded a significantly greater reduction in pain interference ( p = .049, d = 0.43) and a marginally greater reduction in pain intensity ( p = .054, d = 0.36), whereas the educational interview led to a marginally greater yet nonsignificant increase in readiness for change ( p = .095, d = 0.24), as well as several significant improvements within-condition. Results did not differ by VH voice. CONCLUSIONS: Interactive VH interviewers hold promise for improving chronic pain care, including probing for psychosocial risk factors and providing pain-related education.

Using Item Response Times in Online Questionnaires to Detect Mild Cognitive Impairment.

OBJECTIVES: With the increase in web-based data collection, response times (RTs) for survey items have become a readily available byproduct in most online studies. We examined whether RTs in online questionnaires can prospectively discriminate between cognitively normal respondents and those with cognitive impairment, no dementia (CIND). METHOD: Participants were 943 members of a nationally representative internet panel, aged 50 and older. We analyzed RTs that were passively recorded as paradata for 37 surveys (1,053 items) administered online over 6.5 years. A multilevel location-scale model derived 3 RT parameters for each survey: (1) a respondent's average RT and 2 components of intraindividual RT variability addressing (2) systematic RT adjustments and (3) unsystematic RT fluctuations. CIND status was determined at the end of the 6.5-year period. RESULTS: All 3 RT parameters were significantly associated with CIND, with a combined predictive accuracy of area under the receiver-operating characteristic curve = 0.74. Slower average RTs, smaller systematic RT adjustments, and greater unsystematic RT fluctuations prospectively predicted a greater likelihood of CIND over periods of up to 6.5, 4.5, and 1.5 years, respectively. DISCUSSION: RTs for survey items are a potential early indicator of CIND, which may enhance analyses of predictors, correlates, and consequences of cognitive impairment in online survey research.

Shedding light on participant selection bias in Ecological Momentary Assessment (EMA) studies: Findings from an internet panel study.

Although the potential for participant selection bias is readily acknowledged in the momentary data collection literature, very little is known about uptake rates in these studies or about differences in the people that participate versus those who do not. This study analyzed data from an existing Internet panel of older people (age 50 and greater) who were offered participation into a momentary study (n = 3,169), which made it possible to compute uptake and to compare many characteristics of participation status. Momentary studies present participants with brief surveys multiple times a day over several days; these surveys ask about immediate or recent experiences. A 29.1% uptake rate was observed when all respondents were considered, whereas a 39.2% uptake rate was found when individuals who did not have eligible smartphones (necessary for ambulatory data collection) were eliminated from the analyses. Taking into account the participation rate for being in this Internet panel, we estimate uptake rates for the general population to be about 5%. A consistent pattern of differences emerged between those who accepted the invitation to participate versus those who did not (in univariate analyses): participants were more likely to be female, younger, have higher income, have higher levels of education, rate their health as better, be employed, not be retired, not be disabled, have better self-rated computer skills, and to have participated in more prior Internet surveys (all p < .0026). Many variables were not associated with uptake including race, big five personality scores, and subjective well-being. For several of the predictors, the magnitude of the effects on uptake was substantial. These results indicate the possibility that, depending upon the associations being investigated, person selection bias could be present in momentary data collection studies.

Evaluation of Pressing Issues in Ecological Momentary Assessment.

The use of repeated, momentary, real-world assessment methods known as the Experience Sampling Method and Ecological Momentary Assessment (EMA) has been broadly embraced over the last few decades. These methods have extended our assessment reach beyond lengthy retrospective self-reports as they can capture everyday experiences in their immediate context, including affect, behavior, symptoms, and cognitions. In this review we evaluate nine conceptual, methodological, and psychometric issues about EMA with the goal of stimulating conversation and guiding future research on these matters: the extent to which participants are actually reporting momentary experiences, respondents' interpretation of momentary questions, the use of comparison standards in responding, efforts to increase the EMA reporting period beyond the moment to longer periods within a day, training of EMA study participants, concerns about selection bias of respondents, the impact of missing EMA assessments, the reliability of momentary data, and for which purposes EMA might be considered a gold standard for assessment. Resolution of these issues should have far-reaching implications for advancing the field.

Quality of Survey Responses at Older Ages Predicts Cognitive Decline and Mortality Risk.

Background and Objectives: It is widely recognized that survey satisficing, inattentive, or careless responding in questionnaires reduce the quality of self-report data. In this study, we propose that such low-quality responding (LQR) can carry substantive meaning at older ages. Completing questionnaires is a cognitively demanding task and LQR among older adults may reflect early signals of cognitive deficits and pathological aging. We hypothesized that older people displaying greater LQR would show faster cognitive decline and greater mortality risk. Research Design and Methods: We analyzed data from 9, 288 adults 65 years or older in the Health and Retirement Study. Indicators of LQR were derived from participants' response patterns in 102 psychosocial questionnaire items administered in 2006-2008. Latent growth models examined whether LQR predicted initial status and change in cognitive functioning, assessed with the modified Telephone Interview for Cognitive Status, over the subsequent 10 years. Discrete-time survival models examined whether LQR was associated with mortality risk over the 10 years. We also examined evidence for indirect (mediated) effects in which LQR predicts mortality via cognitive trajectories. Results: After adjusting for age, gender, race, marital status, education, health conditions, smoking status, physical activity, and depressive symptoms, greater LQR was cross-sectionally associated with poorer cognitive functioning, and prospectively associated with faster cognitive decline over the follow-up period. Furthermore, greater LQR was associated with increased mortality risk during follow-up, and this effect was partially accounted for by the associations between LQR and cognitive functioning. Discussion and Implications: Self-report questionnaires are not formally designed as cognitive tasks, but this study shows that LQR indicators derived from self-report measures provide objective, performance-based information about individuals' cognitive functioning and survival. Self-report surveys are ubiquitous in social science, and indicators of LQR may be of broad relevance as predictors of cognitive and health trajectories in older people.

Momentary social interactions and affect in later life varied across the early stages of the COVID-19 pandemic.

The COVID-19 pandemic has impacted many different facets of life. The infectious nature of the disease has led to significant changes in social interactions in everyday life. The present study examined how older adults' patterns of everyday momentary social interactions (i.e., with no one, partner, family, and friends) and their affect varied across the early stages of the pandemic and whether the magnitude of affective benefits associated with social interactions changed across time. A total of 188 adults aged 50 or above (Mage = 62.05) completed momentary assessments in early March, late March, May, and July 2020. Overall, older adults spent more time in solitude and less time interacting with their friends after the declaration of the pandemic. Further, negative affect (NA) spiked after the pandemic declaration and then returned to pre-pandemic level. Finally, momentary interactions with close social ties were consistently associated with higher positive affect (PA) and lower NA whereas momentary solitude was associated with lower PA, but not related to NA. The magnitude of associations between specific social interactions (or solitude) and affect varied across time, and the onset of the pandemic appeared associated with this variation. During the presumably most stressful period, solitude was not associated with lower PA and family interaction was not associated with higher PA as they were at other times. Further, interactions with friends seemed to have diminished affective benefits following the onset of the pandemic.

Vague quantifiers demonstrate little susceptibility to frame of reference effects.

Comparison standards that people use when responding to survey questions, also called Frames of Reference (FoRs), can influence the validity of self-report responses. The effects of FoRs might be the stronger for items using vague quantifier (VQ) scales, which are particularly prominent in quality of life research, compared with numeric responses. This study aims to investigate the impact of FoRs on self-report measures by examining how imposing a specific FoR in survey questions affects (a) the response levels of VQ and numeric scales and (b) the relationship between VQs and a quantitative responses to the same question. A sample of 1,869 respondents rated their education, commute and sleep duration, medication use, and level of physical activity using both VQ and numeric formats. Participants were asked to compare themselves with the average US adult, with their friends who are about their age, or did not receive specific instructions regarding a reference for comparison. We found that FoR conditions did not influence the numeric responses. Among the VQ responses, only education attainment was affected by FoR. The association between the numeric responses and vague quantifiers was comparable across different FoR conditions. Our results showed that manipulating the use of interpersonal FoRs had limited effect on the responses, which suggests that at least some comparisons do not have a strong biasing effect on self-report measures. However, future research should confirm this finding for using other FoRs (e.g., historical or hypothetical comparisons) and other outcome measures.

Little evidence for consistent initial elevation bias in self-reported momentary affect: A coordinated analysis of ecological momentary assessment studies.

Response bias characterized by decreases in self-reported subjective states when measured repeatedly over short time-frames is a potential concern in social science. Recent work suggests that this initial elevation bias (IEB) is pronounced among young adult students' self-reports of affect when using ambulatory methods, but it is unclear if such bias extends broadly across samples, designs, and constructs. We examined the conditions wherein reliable and robust IEB may manifest by conducting a coordinated analysis of seven ecological momentary assessment (EMA) studies with diverse lifespan samples to test the generalizability of IEB across study designs and affective constructs (momentary negative and positive affect). Overall, evidence for substantial IEB across studies was weak. No reliable evidence emerged for IEB in negative affect, with evidence for a small magnitude IEB for positive affect when comparing initial reports with reports made 1 week later, although the latter was not evident in other comparisons and was attenuated to nonsignificance when controlling for temporal factors. The magnitude and direction of IEB varied, but in mostly nonsystematic ways, as a function of study design and affective valence. Meta-analytic summary revealed consistently low combined effect sizes (Cohen's ds ranging from -.05 to .14). We found little evidence that IEB in momentary affect is sufficiently reliable, robust, or generalizable across designs and constructs to pose broad and/or serious concerns for EMA studies. Nonetheless, we recommend systematically examining the potential for IEB across study designs and constructs to help identify the conditions/contexts where IEB may or may not manifest. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Achieving reliable pain change scores for individuals in the postoperative phase: carefully choose sampling density, test length, and administration mode.

ABSTRACT: Despite tremendous efforts to increase the reliability of pain measures and other self-report instruments, improving or even evaluating the reliability of change scores has been largely neglected. In this study, we investigate the ability of 2 instruments from the Patient-Reported Outcomes Measurement Information System, pain interference (6 items) and pain behavior (7 items), to reliably detect individual changes in pain during the postsurgical period of a hernia repair in 98 patients who answered daily diaries over almost 3 weeks after surgery. To identify the most efficient strategy for obtaining sufficiently reliable estimates of change (reliability >0.9), the number of measurement occasions over the study period (sampling density), the number of items (test length), and the mode of administration (ie, static short form vs Computer adaptive testing) were manipulated in post-hoc simulations. Reliabilities for different strategies were estimated by comparing the observed change with the best approximation of "real" (ie, latent) change. We found (1) that near perfect reliability can be achieved if measures from all days over the whole study period, obtained with all pain interference or pain behavior items, were used to estimate the observed change, (2) that various combinations of the number of items and the number of measurement occasions could achieve acceptable reliability, and (3) that computer adaptive testings were superior to short forms in achieving sufficient reliability. We conclude that the specific strategy for assessing individual postoperative change in pain experience must be selected carefully.

Contrasting Effects of Finding Meaning and Searching for Meaning, and Political Orientation and Religiosity, on Feelings and Behaviors During the COVID-19 Pandemic.

Perceiving life as meaningful can buffer against negative experiences, whereas searching for meaning in life is often associated with negative outcomes. We examined how these individual differences, along with religiosity and political orientation, are associated with feelings and health-related behaviors during the COVID-19 pandemic (N = 7,220; U.S. nationally representative sample). Conservatism and religiosity predicted less negative effect; conservatives (but not the highly religious) were less likely to engage in preventive actions such as wearing face masks and social distancing. Controlling for political orientation, religiosity, and demographics, the presence of meaning in life predicted less negative affect and greater healthy preventive actions, whereas searching for meaning predicted greater negative affect and more preventive and risky health behaviors. Thus, the perception that life is meaningful not only predicts an individual's emotional well-being but is also associated with beneficial actions that can help protect others from the spread of the coronavirus.

Inferring Cognitive Abilities from Response Times to Web-Administered Survey Items in a Population-Representative Sample.

Monitoring of cognitive abilities in large-scale survey research is receiving increasing attention. Conventional cognitive testing, however, is often impractical on a population level highlighting the need for alternative means of cognitive assessment. We evaluated whether response times (RTs) to online survey items could be useful to infer cognitive abilities. We analyzed >5 million survey item RTs from >6000 individuals administered over 6.5 years in an internet panel together with cognitive tests (numerical reasoning, verbal reasoning, task switching/inhibitory control). We derived measures of mean RT and intraindividual RT variability from a multilevel location-scale model as well as an expanded version that separated intraindividual RT variability into systematic RT adjustments (variation of RTs with item time intensities) and residual intraindividual RT variability (residual error in RTs). RT measures from the location-scale model showed weak associations with cognitive test scores. However, RT measures from the expanded model explained 22−26% of the variance in cognitive scores and had prospective associations with cognitive assessments over lag-periods of at least 6.5 years (mean RTs), 4.5 years (systematic RT adjustments) and 1 year (residual RT variability). Our findings suggest that RTs in online surveys may be useful for gaining information about cognitive abilities in large-scale survey research.

Subtle mistakes in self-report surveys predict future transition to dementia.

INTRODUCTION: We investigate whether indices of subtle reporting mistakes derived from responses in self-report surveys are associated with dementia risk. METHODS: We examined 13,831 participants without dementia from the prospective, population-based Health and Retirement Study (mean age 69 ± 10 years, 59% women). Participants' response patterns in 21 questionnaires were analyzed to identify implausible responses (multivariate outliers), incompatible responses (Guttman errors), acquiescent responses, random errors, and the proportion of skipped questions. Subsequent incident dementia was determined over up to 10 years of follow-up. RESULTS: During follow-up, 2074 participants developed dementia and 3717 died. Each of the survey response indices was associated with future dementia risk controlling for confounders and accounting for death as a competing risk. Stronger associations were evident for participants who were younger and cognitively normal at baseline. DISCUSSION: Mistakes in the completion of self-report surveys in longitudinal studies may be early indicators of dementia among middle-aged and older adults.

Explaining age differences in the memory-experience gap.

Emotions and symptoms are often overestimated in retrospective ratings, a phenomenon referred to as the "memory-experience gap." Some research has shown that this gap is less pronounced among older compared to younger adults for self-reported negative affect, but it is not known whether these age differences are evident consistently across domains of well-being and why these age differences emerge. In this study, we examined age differences in the memory-experience gap for emotional (positive and negative affect), social (loneliness), and physical (pain, fatigue) well-being. We also tested four variables that could plausibly explain age differences in the gap: (a) episodic memory and executive functioning, (b) the age-related positivity effect, (c) variability of daily experiences, and (d) socially desirable responding. Adults (n = 477) from three age groups (21-44, 45-64, 65+ years old) participated in a 21-day diary study. Participants completed daily end-of-day ratings and retrospective ratings of the same constructs over different recall periods (3, 7, 14, and 21 days). Results showed that, relative to young and middle-aged adults, older adults had a smaller memory-experience gap for negative affect and loneliness. Lower day-to-day variability partly explained why the gap was smaller for older adults. There was no evidence that the magnitude of the memory-experience gap for positive affect, pain or fatigue depended on age. We recommend that future research considers how variability in daily experiences can impact age differences in retrospective self-reports of well-being. (PsycInfo Database Record (c) 2021 APA, all rights reserved).