ArticlePDF Available

Meta-Analysis of the Age-Related Positivity Effect: Age Differences in Preferences for Positive Over Negative Information

Authors:

Abstract

In contrast to long-held axioms of old age as a time of "doom and gloom," mounting evidence indicates an age-related positivity effect in attention and memory. However, several studies report inconsistent findings that raise critical questions about the effect's reliability, robustness, and potential moderators. To address these questions, we conducted a systematic meta-analysis of 100 empirical studies of the positivity effect (N = 7,129). Results indicate that the positivity effect is reliable and moderated by theoretically implicated methodological and sample characteristics. The positivity effect is larger in studies that do not constrain (vs. constrain) cognitive processing-reflecting older adults' natural information processing preferences-and in studies incorporating wider (vs. narrower) age comparisons. Analyses indicated that older adults show a significant information processing bias toward positive versus negative information, whereas younger adults show the opposite pattern. We discuss implications of these findings for theoretical perspectives on emotion-cognition interactions across the adult life span and suggest future research directions. (PsycINFO Database Record (c) 2014 APA, all rights reserved).
Meta-Analysis of the Age-Related Positivity Effect: Age Differences in
Preferences for Positive Over Negative Information
Andrew E. Reed and Larry Chan
Stanford University
Joseph A. Mikels
DePaul University
In contrast to long-held axioms of old age as a time of “doom and gloom,” mounting evidence
indicates an age-related positivity effect in attention and memory. However, several studies report
inconsistent findings that raise critical questions about the effect’s reliability, robustness, and
potential moderators. To address these questions, we conducted a systematic meta-analysis
of 100 empirical studies of the positivity effect (N!7,129). Results indicate that the positivity
effect is reliable and moderated by theoretically implicated methodological and sample character-
istics. The positivity effect is larger in studies that do not constrain (vs. constrain) cognitive
processing—reflecting older adults’ natural information processing preferences—and in studies
incorporating wider (vs. narrower) age comparisons. Analyses indicated that older adults show a
significant information processing bias toward positive versus negative information, whereas
younger adults show the opposite pattern. We discuss implications of these findings for theoretical
perspectives on emotion– cognition interactions across the adult life span and suggest future research
directions.
Keywords: aging, positivity effect, meta-analysis, emotion, information processing
Through much of modern history psychologists espoused
views on aging that were less than kind. “Doom and gloom”
perspectives characterizing later life as a time of profound
physical, cognitive, and emotional losses prevailed well into the
20th century (for a discussion see Carstensen, Pasupathi, Mayr,
& Nesselroade, 2000). Yet recent empirical and theoretical
work challenges long-held axioms by illustrating the “bright
side” of aging, from improved psychological well-being and
emotional self-regulation to an age-related positivity effect (for
a review see Charles & Carstensen, 2010). Researchers have
devoted particular attention to the positivity effect, which has
been the subject of numerous empirical studies in the decade
since it was initially observed and conceptualized by Carstensen
and colleagues (Charles, Mather, & Carstensen, 2003;Mather
& Carstensen, 2003). The term “positivity effect” refers to an
observed age-related increase in the preference for positive over
negative information in attention and memory (Carstensen &
Mikels, 2005;Mather & Carstensen, 2005). Among the many
replications of the positivity effect, however, are a handful of
empirical studies that challenged the consistency, size, and
reliability of the effect (for a review, see Reed & Carstensen,
2012). Inconsistencies in the patterns reported raise important
questions for future research in this area: Is the positivity effect
reliable and, if so, how large is the effect and what factors
moderate it? The present meta-analysis was designed to address
these questions.
Age-related differences in the processing of emotionally va-
lenced information (i.e., a preferential shift toward the positive)
were first considered within the context of socioemotional
selectivity theory (SST; Carstensen, 2006), a life span theory of
motivation. According to SST, motivational priorities shift
across the life span as a function of future time horizons. When
individuals perceive their futures as relatively open-ended and
nebulous, as in early adulthood, they tend to prioritize future-
oriented goals such as acquiring information, meeting new
people, and generally expanding their horizons. However, when
individuals increasingly appreciate the fragility of life and
future time horizons narrow, as is typical in later life, they
prioritize present-focused goals related to emotional meaning
and satisfaction. These systematic age differences in goal pri-
orities consequently should alter information processing by
shifting attention and memory toward goal-congruent and away
from goal-incongruent material. Because older adults are espe-
cially motivated by goals related to emotional satisfaction, SST
predicts an information processing shift toward positive infor-
mation in later life. This life-span perspective stands in contrast
to extensive research on the well-documented negativity bias,
which captures how negative information and emotions are
substantially more impactful than positive (Baumeister, Brat-
slavsky, Finkenauer, & Vohs, 2001;Rozin & Royzman, 2001).
SST therefore conceptualizes the positivity effect in terms of the
relative age difference between younger and older adults in the
processing of positive versus negative information: Older adults
attend to and remember positive versus negative information to
Andrew E. Reed and Larry Chan, Department of Psychology, Stanford
University; Joseph A. Mikels, Department of Psychology, DePaul Univer-
sity.
We thank Ulrich Mayr and anonymous reviewers for their constructive
comments on earlier versions of this article.
Correspondence concerning this article should be addressed to Andrew
E. Reed, Department of Psychology, 450 Serra Mall, Stanford University,
Stanford, CA 94305. E-mail: andyreed@stanford.edu
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Psychology and Aging © 2014 American Psychological Association
2014, Vol. 29, No. 1, 1–15 0882-7974/14/$12.00 DOI: 10.1037/a0035194
1
a greater extent than younger adults
1
(Carstensen & Mikels,
2005;Mather & Carstensen, 2005). Based on this conceptual-
ization, the critical contrast for the positivity effect is between
positive and negative information. This perspective on the
positivity effect converges with the theoretical basis of the
negativity bias as “bad is stronger than good” (Baumeister et
al., 2001) and affords comparisons between the two literatures
(for a discussion see Carstensen & Mikels, 2005;Mikels, Reed,
Hardy, & Löckenhoff, in press).
Initial research on age differences in the processing of emotional
information consistently supported the positivity effect. In one of the
earliest illustrations of the positivity effect, Mather and Carstensen
(2003) used a dot-probe visual attention paradigm to demonstrate that
compared with younger adults, older adults preferentially look toward
positive (i.e., happy) and away from negative (i.e., angry or sad) faces.
Subsequent studies replicated the positivity effect in visual attention
using eye-tracking methods (Isaacowitz, Wadlinger, Goren, & Wil-
son, 2006a,2006b). Early studies also found support for the positivity
effect in memory. Compared with younger adults, older adults better
remembered positive versus negative information across paradigms
ranging from autobiographical and long-term memory (Charles et al.,
2003;Kennedy, Mather, & Carstensen, 2004)toworkingmemory
(Mikels, Larkin, Reuter-Lorenz, & Carstensen, 2005). The positivity
effect is also evident in decision making. Older versus younger adults
show relatively greater attention and memory for positive versus
negative attributes when choosing among health-related and everyday
options (Löckenhoff & Carstensen, 2007,2008;Mather, Knight, &
McCaffrey, 2005).
In the years following these early studies, the positivity effect
has been documented across a wide range of experimental para-
digms, tasks, and stimuli. Researchers have found evidence for the
positivity effect in visual stimuli, such as emotional faces and
photographs (Leigland, Schulz, & Janowsky, 2004;Spaniol, Voss,
& Grady, 2008) and lexical stimuli such as word lists and health
messages (Piguet, Connally, Krendl, Huot, & Corkin, 2008;Sha-
maskin, Mikels, & Reed, 2010), suggesting that the effect gener-
alizes across different types of materials. At the same time, the
variety of paradigms used to demonstrate the positivity effect—
from rapid visual attention (Mickley Steinmetz, Muscatell, &
Kensinger, 2010) to autobiographical memory (Kennedy et al.,
2004) to decision making (Löckenhoff & Carstensen, 2007)—
suggests that the effect is not constrained to specific domains of
information processing.
2
Although empirical support for the positivity effect is wide-
ranging, more than a few studies have found minimal or no age
differences in the processing of positive versus negative information
(for a review, see Reed & Carstensen, 2012). Indeed, studies have
failed to observe the positivity effect in memory for words (e.g.,
Grühn, Smith, & Baltes, 2005;Majerus & D’Argembeau, 2011),
images (e.g., Gallo, Foster, & Johnson, 2009)andadvertisements
(e.g., Williams & Drolet, 2005). At first glance these findings appear
to challenge the reliability of the positivity effect. However, when the
experimental designs of these studies are compared with studies
finding the effect, a pattern emerges suggesting a key moderator:
Studies that do not observe a positivity effect typically impose exper-
imental constraints on information processing by altering individuals’
goals, attention and/or resources. For instance, age differences in
emotional information processing fail to appear when participants are
instructed to remember all information either explicitly (Grühn et al.,
2005;Majerus & D’Argembeau, 2011)orimplicitly(Gallo et al.,
2009)orwhentheyarerequiredtomakeexplicitjudgmentsabout
experimental stimuli at encoding (e.g., Kensinger, Brierley, Medford,
Growdon, & Corkin, 2002). By contrast, studies that afford and
encourage naturalistic and unconstrained processing of stimuli (e.g.,
instructing participants to view images as they would a TV) observe
the positivity effect (Charles et al., 2003;Isaacowitz et al., 2006b;
Kwon, Scheibe, Samanez-Larkin, Tsai, & Carstensen, 2009).
Empirical observations of the positivity effect when information
processing is unconstrained versus constrained are entirely consis-
tent with the motivational account offered by SST (for discussions,
see Mather & Carstensen, 2005;Reed & Carstensen, 2012). Ac-
cording to SST, age differences in the processing of emotional
information are driven in a top-down manner by chronically acti-
vated goals. Pursuing such goals requires sufficient cognitive
control and an absence of superseding situation-specific goals.
Thus, SST would not predict a positivity effect when individuals
are imbued with goals that conflict with their chronically activated
goals and/or when their cognitive resources are limited, and em-
pirical work supports these predictions. Across a series of exper-
iments Mather and colleagues demonstrated the positivity effect in
memory and attention among older individuals with high levels of
cognitive control and also those in a full versus divided attention
condition (Knight et al., 2007;Mather & Knight, 2005). Löcken-
hoff and Carstensen (2007) illustrated the moderating role of goals
when they eliminated the positivity effect in decision making by
activating information-seeking goals in younger and older adults
via experimental instructions. These studies illustrate how manip-
ulations of experimental instructions and task characteristics en-
gender versus suppress the positivity effect.
In recent years alternative theoretical accounts of the age-related
positivity effect have been offered based on age-related cognitive
or neural deficits (for a discussion, see Reed & Carstensen, 2012).
Labouvie-Vief and colleagues (Labouvie-Vief et al., 2010;Wurm,
2011) posited in Dynamic Integration Theory that age-related
cognitive declines lead older adults to automatically and preferen-
tially process positive information because it is easier to process
(i.e., less complex) than negative information. The aging-brain
model proposed by Cacioppo, Berntson, Bechara, Tranel, and
Hawkley (2011) attribute the positivity effect to age-related de-
generation of the amygdala inhibiting neural and affective re-
sponses to negative (though not positive) information. However,
these perspectives are inconsistent with key empirical findings
1
Unfortunately, the literature now includes a number of related terms
that are too often used interchangeably with the positivity effect and
obfuscate the issues (see, e.g., Murphy & Isaacowitz, 2008). For instance,
older adults are often referred to as demonstrating a positivity bias,
whereby positive information is attended to and remembered more than
negative information. Yet, despite the fact that in some studies findings are
driven by a preference for positive among older adults (i.e., a positivity
bias; Mather et al., 2004;Mikels et al., 2005), the positivity effect can just
as easily reflect a reduced preference for negative. Relatedly, the positivity
effect has been confused with a positivity preference whereby positive
material is more impactful than neutral material.
2
Empirical evidence of the positivity effect also extends beyond behav-
ioral findings to age-related patterns of neural activity (for a review, see
Samanez-Larkin & Carstensen, 2011). For instance, compared with
younger adults, older adults recruit subcortical regions such as the
amygdala to a greater extent when processing positive versus negative
information (Mather et al., 2004).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2REED, CHAN, AND MIKELS
reviewed above. For instance, although both explanations predict
that the positivity effect would be strongest among individuals
with the greatest age-related cognitive or neural impairments,
evidence supports the opposite prediction: The positivity effect is
most evident when comparing younger and older individuals with
greater versus poorer cognitive control (Mather & Knight, 2005).
Additionally, neither Dynamic Integration Theory nor the aging-
brain model can account for evidence that the positivity effect can
be modulated by experimental manipulations of goals or process-
ing resources (e.g., Löckenhoff & Carstensen, 2007;Knight et al.,
2007). However, these findings are consistent with the motiva-
tional perspective of SST.
Present Study
To address open questions surrounding the positivity effect and
to test theoretically derived moderators, we conducted a systematic
meta-analysis of studies of attention and memory. All studies
included in the present meta-analysis met strict inclusion criteria
based on the operational definition of the positivity effect as an
interaction between age and valence: Older relative to younger
adults show greater attention and memory for positive versus
negative information.
Arecentmeta-analysisbyMurphy and Isaacowitz (2008) likewise
examined age differences in the processing of emotionally valenced
information, but with goals and methods that are distinct from the
present study in important ways.
3
The goals of the prior meta-analysis
were to examine age differences in the emotion salience effect,
positivity preference, and negativity preference. This engendered
broad inclusion criteria that extended beyond direct tests of the pos-
itivity effect. As a result, most of the included studies (roughly 85%)
had experimental designs that precluded testing for the interaction
between age and positive-versus-negative valence (i.e., the age-related
positivity effect). In addition, the analyses by Murphy and Isaacowitz
focused on contrasts between neutral and positive or negative stimuli,
rather than the positive–negative contrast of the positivity effect. As
acknowledged by the researchers (Murphy & Isaacowitz, 2008,p.
265), their findings speak to questions regarding age differences in
positivity and negativity preferences but do not directly address the
positivity effect.
By contrast, the present meta-analysis was specifically designed
as a direct test of the positivity effect and consequently only
included studies that incorporated comparisons across age (i.e.,
older vs. younger) and valence (i.e., positive vs. negative). Our
hypotheses were grounded principally in the theoretical perspec-
tive offered by SST (for a review, see Reed & Carstensen, 2012).
Overall, we predicted that the positivity effect would be reliable.
However, we remained agnostic about the global size of the effect
because we predicted that it would be moderated by two key
theoretically implicated factors. First and foremost, we hypothe-
sized that the size of the positivity effect would be moderated by
experimental constraints on goals, attention, and/or cognitive re-
sources. Specifically, we predicted that the effect size would be
strongest among studies that did not constrain information pro-
cessing through instructional manipulations and/or task-related
restrictions and weakest among studies with such constraints.
Second, because the positivity effect is theoretically linked to a
gradual change in motivational priorities across adulthood, we also
predicted that the effect size would be moderated by the mean age
difference between the younger and older samples. Studies with
greater age discrepancies between the two subsamples were ex-
pected to yield larger effect sizes.
Method
Literature Search
We conducted an extensive search for empirical studies relevant
to the positivity effect in peer-reviewed journal articles published
as of June 2011. Article databases PsycINFO and Web of Knowl-
edge were searched using the following global keyword combina-
tions: Positivity effect,aging and valence, aging and emotion. In
addition to the global keywords, nested search terms were used.
4
Additional articles were located via ancestor and descendent searches
of highly cited empirical articles (e.g., Charles et al., 2003;Kensinger
et al., 2002;Mather & Carstensen, 2003), early theoretical articles
(Carstensen & Mikels, 2005;Mather & Carstensen, 2005), and recent
review articles (Charles & Carstensen, 2010;Reed & Carstensen,
2012;Scheibe & Carstensen, 2010). Finally, supplementary articles
(both published and in press) were obtained directly from their authors
in response to data requests.
Inclusion Criteria
Our inclusion criteria were grounded in an operational definition
of the positivity effect centered on the interaction between age and
valence: Older relative to younger adults show greater attention
and memory for positive versus negative information. Conse-
quently, all studies were required to meet the following criteria for
inclusion in the meta-analysis:
1. Included studies needed to have incorporated at least one
age comparison (e.g., younger vs. older adults). Studies
with only one age group were excluded. Studies with
continuous age analyses were included if categorical age
data could be obtained from the authors (e.g., Li, Fung, &
Isaacowitz, 2011). Age groups were coded based on the
original definitions and cutoffs provided by included
studies. Consequently, “younger” and “older” samples
were entered based on the authors’ original categoriza-
tions when available. In addition, because the positivity
effect concerns normal aging, older adult samples iden-
tified as having severe cognitive impairments (e.g., Alz-
3
It is especially instructive to consider these differences given the
frequency with which the Murphy and Isaacowitz meta-analysis has been
misinterpreted in the literature as a test of the positivity effect (e.g., Broster
et al., 2012;Ebner et al., 2012;Goeleven, De Raedt, & Dierckx, 2010;
Johnson & Whiting, 2013;Leclerc & Kensinger, 2011;Wurm, 2011).
4
Nested search terms were as follows: (attention OR cognition OR “dot
probe” OR “eye tracking” OR fixation OR memory OR reaction OR recall
OR recognition OR recollection OR remember OR “visual scanning”)
AND (anger OR bias OR contentment OR disgust OR elation OR emotion
OR “facial expression” OR fear OR happy OR joy OR negative OR
negativity OR neutral OR positive OR positivity OR sad OR surprise)
AND (aging OR ageing OR elderly OR “older adult” OR retired OR elder).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
POSITIVITY EFFECT META-ANALYSIS
heimer’s disease patients) were excluded from the anal-
ysis.
5
2. Studies needed to have assessed information processing
for positively and negatively valenced material to be
included. Studies with only positive or negative stimuli
were excluded.
3. All included studies needed to have direct measures of
attention and/or memory for emotionally valenced infor-
mation. Studies that examined age differences in the
evaluation of positive versus negative stimuli without
direct measures of attention or memory were excluded
(e.g., Grühn & Scheibe, 2008).
Following these criteria, only studies that included full factorial
designs with respect to age and valence were included in the
meta-analysis. Some eligible articles could not be included in
the meta-analysis because critical data were not reported and the
original data sets were no longer available (e.g., Rahhal, May, &
Hasher, 2002;Thompson, Aidinejad, & Ponte, 2001). This led to
the inclusion of 100 independent studies with 115 comparisons
(several studies had multiple conditions or subgroups) and a total
Nof 7,129. Methodological and sample characteristics for all
included studies are detailed in Table 1.
Data Entry
All available study data were entered into a database and sub-
sequently double-checked for accuracy. Authors were contacted if
effect size data (i.e., sample sizes, means, and standard deviations)
could not be directly recorded, calculated, or measured (i.e., from
a graph) based on the published article.
Coding
All included studies were coded for experimental constraints on
goals, resources, and/or attention using the following scheme:
Studies that were coded as “constrained” instructed participants
(explicitly or implicitly) to either focus on or away from valenced
material, required participants to rate or judge stimuli, and/or
explicitly mentioned an upcoming memory test prior to the encod-
ing or study phase. Studies that were coded as “unconstrained”
afforded and/or encouraged passive or naturalistic viewing of
valenced materials (e.g., viewing a series of images as you would
a TV), incorporated incidental memory tests, and/or allowed par-
ticipants to attend to versus ignore whichever information they
pleased (either explicitly or implicitly). Two independent raters
(AR and LC) coded all studies (92.7% agreement rate) and dis-
agreements were resolved via discussion.
Data Analytic Approach
Because the moderator analyses require establishing indepen-
dence between effect sizes across studies, for each study we
selected a single outcome using the following rubric:
For each study with multiple operationalizations of a single
outcome (e.g., d=, CR, and hits for recognition memory) we se-
lected the most precise and comprehensive measure of accuracy
(for memory) or amount (for attention). For memory studies we
used the following priority list: d=, CR (or corrected recall), hits (or
proportion/percent recall). This was done to avoid combining
outcomes that are confounded (e.g., hits and CR) or conceptually
incongruent (e.g., hits and false alarms). For attention studies we
selected dependent measures that best encapsulate the sheer
amount of attention, such as fixation ratios and fixation duration,
over measures of attentional shifts (e.g., fixation counts and sac-
cades).
For studies with multiple measurement points, we selected the
earliest outcome measures (e.g., the shortest delay between study
and test phase). Thus, for memory studies with both recall and
recognition measures, we selected the recall measures if they were
administered first.
Studies with multiple between-subjects experimental conditions
or subgroups were entered and coded separately by condition or
subgroup in the meta-analysis (see Table 1).
All effect size data were subjected to a meta-analysis in which
the primary outcome variable was the interaction between age
(older vs. younger) and valence (positive vs. negative). Consistent
with the operational definition of the positivity effect as the rela-
tive age difference in attention and memory for positive versus
negative information, positivity effect sizes were calculated based
on recommendations for independent groups repeated-measures
designs (Becker, 1988;Morris & DeShon, 2002). First, we com-
puted effect sizes for positivity bias scores (d
bias
) within each age
group and study (or study subgroup) as the paired standardized
mean difference between positive and negative scores (see Boren-
stein, 2009):
dbias !
!
MPos "MNeg
SDDif
"
#2(1 "r)
SDDif !#(SDNeg
2#SDPos
2"2r!SDNeg !SDPos)
$bias !
!
1
n
#dbias
2
2n
"
2
$
1"r
%
In the above formulas M
Pos
and M
Neg
refer to mean positive and
negative scores within each age group, SD
Pos
and SD
Neg
refer to
corresponding standard deviations, SD
Dif
refers to the standard
deviation of the positive-minus-negative difference, v
bias
refers to
the variance of the bias score effect size, nrefers to the sample size
within each age group, and rrefers to the correlation between
positive and negative scores. We used a conservative positive-to-
negative correlation of r!.4 derived from the Charles et al.
(2003) data set.
6
After calculating positivity bias scores within each age group,
we then computed the positivity effect size (d
PE
) for each study (or
subgroup) as the difference in positivity bias effect sizes between
older (d
biasO
) and younger (d
biasY
) samples (following Becker,
1988):
dPE !dbiasO "dbiasY
5
Healthy older adult samples from the same studies were included in the
analysis (e.g., Fleming et al., 2003;Kensinger et al., 2002).
6
Analyses were replicated using correlations of r!.2 and r!.6, and
the results were not significantly different from those reported below.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4REED, CHAN, AND MIKELS
Variance for the positivity effect size (v
PE
) was calculated as the
sum of the variance for younger and older adults’ bias scores
(v
biasY
and v
biasO
, respectively; see Morris & DeShon, 2002):
$PE !$
biasO #$
biasY
Finally, we corrected the positivity effect scores for small sam-
ple bias (following Morris & DeShon, 2002). Corrected effect
sizes were used for the overall meta-analysis as well as the mod-
erator and meta-regression analyses. Effect sizes based on Cohen’s
dcan be categorized as small (d!.2), medium (d!.5), and large
(d!.8; Cohen, 1988).
We chose this effect size metric for two reasons: First, it best
reflected our operational definition of the positivity effect as an
interaction between age and valence. Using this approach yielded
effect sizes that were signed and scaled in relation to the relative
strength of the positivity effect. Larger and more positive effect
sizes indicated a strong positivity effect, whereas smaller or neg-
ative effect sizes indicated a weaker positivity effect or a negativ-
ity effect,
7
respectively. The second reason for using this effect
size metric is that it afforded and facilitated the use of moderator
and meta-regression analyses.
Following guidelines by Konstantopoulos and Hedges (2009)
we conducted a meta-regression analysis via a weighted least-
squares linear regression. The positivity effect estimate (d
PE
) was
entered as the dependent variable, the within-study mean age
difference (squared) was the independent variable, and the recip-
rocal of the sampling variance (1/v
PE
above) was used as the
weight. We entered a fixed regression intercept of zero to reflect
the fact that the positivity effect theoretically cannot exist when
comparing samples of equal age (i.e., no mean age difference). A
corrected standard error (S
j
) was computed by dividing the stan-
dard error (SE
j
) of the regression estimate by the square root of the
residual mean square (MS
RMS
) from the regression analysis of
variance:
Sj!SEj#MSRMS
The corrected standard error was used to calculate a 95%
confidence interval (CI) for the regression estimate (b
j
) as follows:
95% CI !b
j
"1.96 S
j
.
Because we anticipated multiple moderators of the positivity
effect size, we used a random effects model for the main analysis.
Effect sizes were calculated from reported means and standard
deviations when available and from inferential statistics when
means and/or standard deviations were unavailable.
8
CIs were
computed based on the standard deviations of the effect sizes and
assessed in our moderator analysis. Bias score effect sizes were
calculated in Comprehensive Meta-Analysis software (Biostat,
Inc.). Positivity effect sizes were computed from bias scores in
SPSS, which was used to conduct the meta-regression. All effect
sizes were meta-analyzed in Comprehensive Meta-Analysis soft-
ware.
Results
As predicted, we observed a reliable positivity effect overall,
indicated by a significant and positive effect size across all studies,
d
PE
!.257, 95% CI [.165, .349], Z!5.48, p#.001. Examination
of the positivity bias scores within each age group revealed that
older adults showed a small but significant positive bias overall,
d
bias
!.128, 95% CI [.042, .214], Z!2.92, p#.01, whereas
younger adults showed a significant negative bias (d
bias
!$.123,
95% CI [$.203, $.043], Z!$3.01, p#.01). Effect sizes for
individual studies are depicted in Table 2.
We tested for publication bias using the tandem method (Fer-
guson & Brannick, 2012): First, we calculated a fail-safe Nof
2,913 (%!.05), Z!10.06, p#.001. This indicates that the
number of additional (i.e., unpublished, missing or new) studies
needed to reduce the overall effect size to nonsignificance (%&
.05) is more than 25 times the number of included studies. We then
computed Egger’s regression (Egger et al., 1997), which yielded a
nonsignificant bias intercept, '!$.14, t(113) !.19, p!.85.
Finally, we conducted the Duval and Tweedie (2000) trim-and-fill
procedure, which yielded an imputed point estimate of d!.43,
95% CI [.39, .48]. This indicates that missing or unpublished
studies would actually increase, rather than decrease, the overall
effect size. Thus, combined results from the tandem method sug-
gest that the likelihood of publication bias is minimal.
To test whether processing constraint moderated the size of the
positivity effect, we computed 95% CIs for effect sizes among
constrained versus unconstrained studies (see Table 3). Consistent
with our hypothesis, when examining the positivity effect non-
overlapping CIs indicated a significantly larger effect size among
unconstrained studies, which yielded a medium average effect
size, d!.482, 95% CI [.323, .640], Z!5.95, p#.001), versus
constrained studies, which yielded a small average effect size, d!
.134, 95% CI [.033, .235], Z!2.61, p#.01. Subsequent exam-
ination of bias scores indicated that older adults showed a signif-
icant positive bias when their processing was unconstrained, d!
.280, 95% CI [.123, .436], Z!3.50, p#.001, but no bias under
constrained conditions, d!.048, 95% CI [$.048, .144], Z!.97,
p!.33. By contrast, younger adults’ negative bias did not differ
significantly based on processing constraint, as indicated by a
nonsignificant between-groups heterogeneity test, Q
b
(1) !1.61,
p!.21. Younger adults displayed a significant negative bias both
when their processing was not constrained, d!$.197, 95% CI
[$.340, $.054], p#.01, and under constrained processing con-
ditions, d!$.085, 95% CI [$.182, .011], p#.05 one-tailed.
To test whether the mean age difference between older and
younger samples moderates the size of the positivity effect across
studies, we conducted a meta-regression (see above for details).
9
As indicated in Figure 1, our hypothesis was supported in the
overall data set: There was a significant association between
the magnitude of the age discrepancy (squared) and the size of the
positivity effect across studies, b!.0001, SE !.000018, 95% CI
7
By negativity effect we mean a greater focus on negative versus
positive among older versus younger adults.
8
Means and standard deviations for positive and negative scores were
available for all but the following studies: Kennedy et al. (2004);Löck-
enhoff & Carstensen (2007,2008), and McKay-Nesbitt et al. (2011). The
first three studies reported means and standard deviations for composite
measures of positive-versus-negative attention and memory, and the latter
study reported test statistics for the positive-versus-negative ttest compar-
ison. For these studies the standardized paired difference effect sizes were
calculated based on formulas provided by Borenstein (2009).
9
Effect size data from Kennedy et al. (2004) were excluded from the
meta-regression because the mean age difference could not be calculated
due to missing age data.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
POSITIVITY EFFECT META-ANALYSIS
Table 1
Methodological and Sample Characteristics of Studies Included in the Meta-Analysis
Study Condition Paradigm Stimuli Constraint
Sample size Mean age Mean age
differenceOlder Younger Older Younger
Allard & Isaacowitz, 2008 Full attention EYE IMG N 20 20 70.5 21.6 48.9
Bannerman et al., 2011 BNR FAC N 30 30 70.6 21.8 48.8
Brassen et al., 2011 DIS FAC Y 21 22 65.8 25.2 40.6
Charles et al., 2003, Study 1 LTM IMG N 48 48 71.0 24.6 46.4
Charles et al., 2003, Study 2 LTM IMG N 32 32 74.1 23.5 50.5
Chung, 2010 LTM IMG N 80 43 74.1 21.2 52.8
Comblain et al., 2004 LTM IMG Y 20 20 67.5 22.5 45.0
D’Argembeau & Van der
Linden, 2004 LTM FAC Y 32 32 67.0 24.0 43.0
Ebner & Johnson, 2009,
Study 1 LTM FAC Y 24 32 74.8 19.3 55.5
Ebner & Johnson, 2010,
Study 2 DIS FAC Y 20 32 74.1 19.3 54.8
Emery & Hess, 2008, Study
1 Rating LTM IMG Y 27.5 29 74.1 18.5 55.6
Emery & Hess, 2008, Study
1 Watching LTM IMG N 27.5 29 74.1 18.5 55.6
Emery & Hess, 2008, Study
2 Rating LTM IMG Y 21 23 72.7 20.7 52.0
Emery & Hess, 2008, Study
2 Watching LTM IMG N 21 23 72.7 20.7 52.0
Emery & Hess, 2011 LTM IMG Y 53 48 70.9 19.5 51.4
Feng et al., 2011 LTM IMG N 40 45 74.7 20.0 54.7
Fernandes et al., 2008 LTM, ABM IMG, WOR, ABM Y 48 49 72.3 19.0 53.3
Fleming et al., 2003 LTM WRD Y 19 27 70.1 23.8 46.3
Fung et al., 2008 EYE FAC (SYN) N 57 46 69.0 19.7 49.3
Fung et al., 2010, Study 1 LTM IMG N 103 114 76.5 21.2 55.3
Gallo et al., 2009 LTM IMG Y 24 24 77.7 21.3 56.4
Goeleven et al., 2010 DIS FAC Y 27 27 74.3 32.3 42.0
Grady et al., 2007 LTM FAC Y 30 40 70.0 21.8 48.2
Grühn et al., 2005 LTM WRD Y 72 72 69.3 24.3 45.0
Grühn et al., 2007 LTM IMG Y 48 48 69.8 25.3 44.4
Hahn et al., 2006, Study 1 VIS FAC (SCH) Y 20 20 67.7 21.2 46.5
Hahn et al., 2006, Study 2 VIS FAC (SCH) Y 14 14 65.2 22.7 42.5
Hahn et al., 2006, Study 3 VIS FAC (SCH) Y 15 15 64.5 22.4 42.1
Isaacowitz et al., 2006a EYE FAC (SYN) N 27 37 68.2 18.4 49.8
Isaacowitz et al., 2006b EYE FAC (SYN) N 28 32 71.4 19.8 51.6
Isaacowitz et al., 2008 EYE FAC (SYN) N 52 72 71.4 19.6 51.7
Kapucu et al., 2008 LTM WRD Y 23 22 71.9 19.6 52.3
Kennedy et al., 2004 ABM ABM N 28 28 — —
Kensinger & Schacter, 2008 LTM IMG Y 17 17 73.3 21.6 51.7
Kensinger et al., 2002 LTM IMG Y 20 20 73.3 20.5 52.8
Kensinger et al., 2007,
Study 1 LTM IMG, WRD Y 30 30 74.5 21.5 53.0
Kensinger et al., 2007,
Study 2 LTM WRD Y 30 30 69.7 20.6 49.1
Kensinger, 2008, Study 1 LTM WRD Y 30 30 73.5 26.1 47.4
Kensinger, 2008, Study 2 LTM WRD Y 30 30 72.3 24.6 47.7
Knight et al., 2002 Neutral mood LTM WRD N 33 64 76.5 21.4 55.1
Knight et al., 2002 Sad mood LTM WRD Y 45 55 76.9 20.4 56.5
Knight et al., 2007 Divided attention EYE IMG Y 13.5 16.5 75.0 19.9 55.2
Knight et al., 2007 Full attention EYE IMG N 13.5 16.5 75.0 19.9 55.2
Ko et al., 2011 American sample LTM IMG Y 26 29 72.3 21.7 50.6
Ko et al., 2011 Korean sample LTM IMG Y 26 29 70.4 21.9 48.5
Kwon et al., 2009 LTM IMG N 52 52 70.8 25.1 45.7
Langeslag & Van Strien,
2008 STM IMG Y 20 20 68.5 19.8 48.7
Langeslag & Van Strien,
2009 LTM IMG Y 19 19 71.3 21.2 50.1
Leclerc & Kensinger, 2008 VIS IMG Y 24 24 76.1 19.5 56.6
Leclerc & Kensinger, 2010,
Study 1 EYE IMG Y 18 18 72.2 21.5 50.7
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6REED, CHAN, AND MIKELS
Table 1 (continued)
Study Condition Paradigm Stimuli Constraint
Sample size Mean age Mean age
differenceOlder Younger Older Younger
Leclerc & Kensinger, 2010,
Study 2 EYE IMG Y 24 24 73.7 23.9 49.8
Leclerc & Kensinger, 2011 LTM IMG Y 19 20 71.7 23.4 48.3
Leigland et al., 2004 LTM FAC, WRD Y 36 25 72.3 23.9 48.4
Li et al., 2011 EYE VID N 24 24 68.2 20.8 47.5
Löckenhoff & Carstensen,
2007 Control DEC ATR N 20 20 79.8 27.3 52.5
Löckenhoff & Carstensen,
2007 Emotion-focus DEC ATR Y 20 20 79.8 27.3 52.5
Löckenhoff & Carstensen,
2007 Information-focus DEC ATR Y 20 20 79.8 27.3 52.5
Löckenhoff & Carstensen,
2008 Different Age DEC ATR Y 23.66 23.66 78.1 19.9 58.2
Löckenhoff & Carstensen,
2008 Same Age DEC ATR Y 23.66 23.66 78.1 19.9 58.2
Löckenhoff & Carstensen,
2008 Self (control) DEC ATR N 23.66 23.66 78.1 19.9 58.2
Majerus & D’Argembeau,
2011, Study 2 STM WRD Y 15 15 70.7 26.0 44.7
Majerus & D’Argembeau,
2011, Study 3 STM WRD Y 15 15 69.5 25.0 44.5
Mather & Carstensen, 2003,
Study 1 DOT FAC N 52 52 74.0 25.8 48.2
Mather & Carstensen, 2003,
Study 2 DOT FAC N 44 44 71.5 25.4 46.1
Mather & Knight, 2005,
Study 1 LTM IMG N 48 48 72.7 19.7 53.0
Mather & Knight, 2005,
Study 2 LTM IMG N 31 25 73.6 21.7 51.9
Mather & Knight, 2005,
Study 3 Divided attention LTM IMG Y 16 16 73.8 22.8 51.0
Mather & Knight, 2005,
Study 3 Full attention LTM IMG N 16 16 73.8 22.8 51.0
Mather & Knight, 2006 VIS FAC (SCH) Y 35 33 72.5 20.3 52.2
Mather et al., 2005, Study 3 LTM STA N 40 40 72.5 23.7 48.9
Mather et al., 2005, Study
4b DEC ATR N 48 44 72.6 21.3 51.3
McKay-Nesbitt et al., 2011 LTM ADS N 124 151 70.0 20.0 50.0
Mickley & Kensinger, 2009,
Study 1 LTM IMG Y 26 26 78.2 19.2 59.0
Mickley & Kensinger, 2009,
Study 2 LTM IMG Y 24 25 75.6 19.6 56.0
Mickley Steinmetz et al.,
2010 RVD WRD Y 22 25 74.7 20.6 53.9
Mienaltowski et al., 2011 DIS FAC Y 15 16 69.5 19.9 49.6
Mikels et al., 2005 WM IMG Y 20 20 72.5 22.4 50.2
Nashiro & Mather, 2011a,
Study 1 LTM IMG Y 18 18 72.7 20.7 52.0
Nashiro & Mather, 2011a,
Study 2 LTM IMG Y 24 24 74.9 19.2 55.7
Nashiro & Mather, 2011b LTM IMG Y 24 24 77.1 20.2 56.9
Nashiro et al., 2011, Study 1 RL FAC Y 17 18 72.7 20.7 52.0
Nashiro et al., 2011, Study 2 RL FAC Y 20 20 81.1 19.6 61.5
Nikitin & Freund, 2011 EYE FAC N 79 89 70.5 25.5 45.0
Noh & Isaacowitz, 2011 CUE FAC Y 44 42 74.1 20.4 53.7
Orgeta, 2011, Study 1 DOT FAC N 40 40 69.8 20.1 49.8
Orgeta, 2011, Study 2 DOT FAC (SCH) N 40 40 69.7 22.4 47.4
Piguet et al., 2008 LTM WRD Y 36 36 72.2 21.4 50.8
Pruis et al., 2009 LTM IMG, STO Y 26 25 72.7 29.4 43.3
Ready et al., 2007, Study 1 ABM ABM N 28 21 71.6 24.6 47.0
Ready et al., 2007 Study 2 ABM ABM N 17 53 68.7 22.9 45.8
Rendell et al., 2011 PM IMG, TSK Y 30 30 75.0 21.9 53.1
Ritchey et al., 2011 LTM IMG Y 16 20 66.7 23.2 43.5
(table continues)
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
POSITIVITY EFFECT META-ANALYSIS
[.000083, .000121], Z!10.32, p#.001. Age discrepancy was a
stronger predictor of the size of the positivity effect in studies
without processing constraints, b!.00019, SE !.00003, 95% CI
[.00016, .00022], Z!12.23, p#.001, versus studies with pro-
cessing constraints, b!.00004, SE !.00002, 95% CI [.00002,
.00007], Z!3.46, p#.001.
Discussion
The present study was designed to address open questions
regarding the reliability, size, and moderators of the positivity
effect. Using a meta-analytic approach with strict inclusion criteria
and an expansive set of studies, we found evidence for the posi-
tivity effect overall. Converging analyses indicated a minimal
likelihood of publication bias. Critically, however, the positivity
effect appears to be moderated by experimental constraints on
information processing and the magnitude of the age difference
between younger and older subsamples, supporting the motiva-
tional perspective on the positivity effect offered by SST.
According to SST, the positivity effect results from chronically
activated goals shifting across the adult life span: Whereas
younger people tend to pursue goals concerning knowledge acqui-
sition and expanding horizons, older people prioritize goals related
to emotional meaning and satisfaction. Because these age-related
goals and their corresponding information processing patterns re-
flect top-down motivational processes and not fixed declines in
cognitive or neural capacity, they are malleable and require cog-
nitive resources to pursue (e.g., Mather & Knight, 2006;Löcken-
hoff & Carstensen, 2007). Based on this theoretical framework, we
compared experiments that constrained the goals and/or resources
of research participants (either intentionally or inadvertently) ver-
sus experiments that afforded the unfettered expression and pursuit
of chronically activated goals. Results of this comparison indicated
that the positivity effect is strongest (medium-sized) among studies
that do not constrain information processing (i.e., via goals, atten-
tion and/or resources) and weakest (small) among studies that do
impose experimental constraints. The size of the positivity effect
appears to scale with the magnitude of the age difference within
studies, supporting the view that the effect may reflect a gradual
life span shift in information processing. Examination of bias
scores within each age group revealed that older adults attend to
and better remember positive more than negative information,
whereas the opposite pattern was observed among younger adults. As
with the positivity effect, positivity and negativity biases were mod-
erated by processing constraints: Older adults appear to preferentially
Table 1 (continued)
Study Condition Paradigm Stimuli Constraint
Sample size Mean age Mean age
differenceOlder Younger Older Younger
Rösler et al., 2005 EYE IMG N 12 12 64.4 26.5 37.9
Samanez-Larkin et al., 2009 DIS WRD Y 12 12 73.3 22.2 51.1
Savaskan et al., 2007 LTM FAC Y 15 15 76.7 26.9 49.9
Shamaskin et al., 2010 LTM HRM Y 25 24 74.5 20.4 54.1
Spaniol et al., 2008, Study 1 LTM FAC, IMG, WOR Y 24 24 67.5 22.5 45.0
Spaniol et al., 2008, Study 2 LTM FAC, IMG, WOR N 23 24 71.8 22.3 49.4
St. Jacques et al., 2010 LTM IMG Y 15 15 70.2 24.8 45.4
Thapar & Rouder, 2009 LTM WRD Y 30 30 67.7 19.9 47.9
Thomas & Hasher, 2006 DIS WRD Y 48 48 67.6 21.4 46.2
Tomaszczyk et al., 2008 Active processing LTM IMG Y 36 36 72.3 19.7 52.7
Tomaszczyk et al., 2008 Passive processing LTM IMG N 36 36 72.3 19.7 52.7
Van Gerven & Murphy,
2010 DIS WRD Y 48 48 68.1 21.9 46.2
Waring & Kensinger, 2009 LTM IMG Y 24 24 72.8 19.6 53.2
Werheid et al., 2010,
Study 1 LTM FAC Y 20 20 66.2 24.4 41.8
Werheid et al., 2010,
Study 2 LTM FAC Y 20 20 66.4 24.5 41.9
Werheid et al., 2010,
Study 4 LTM FAC Y 12 12 65.2 26.6 38.6
Williams & Drolet, 2005,
Study 2 Control LTM ADS Y 40.33 41.66 70.0 20.0 50.0
Williams & Drolet, 2005,
Study 2 Expansive time LTM ADS Y 40.33 41.66 70.0 20.0 50.0
Williams & Drolet, 2005,
Study 2 Limited time LTM ADS Y 40.33 41.66 70.0 20.0 50.0
Yang & Hasher, 2011 SEM WRD N 51 55 67.94 19.13 48.81
Yang & Ornstein, 2011 Control LTM IMG N 12 12 68.5 21.92 46.58
Yang & Ornstein, 2011 Emotion-focused LTM IMG Y 20 19 71.3 19.95 51.35
Yang & Ornstein, 2011 Information-focused LTM IMG Y 12 12 69.33 20.25 49.08
Note. Study Type: ATT !attention; MEM !memory. Paradigm: EYE !eye-tracking; BNR !binocular rivalry; DIS !distraction; LTM !long-term
memory; ABM !autobiographical memory; VIS !visual search task; STM !short-term memory; DEC !decision grid; DOT !dot-probe task; WM !
working memory; RL !reversal learning; CUE !spatial cueing; PM !prospective memory; RVD !rapid visual detection; SEM !semantic memory.
Stimuli: ABM !autobiographical memories; ADS !advertisements; ATR !attributes; FAC !faces; FAC (SCH) !schematic faces; FAC (SYN) !
synthetic faces; HRM !health-related messages; IMG !images; WRD !words; VID !videos; STA !statements; STO !stories; TSK !tasks; g
PE
!
Positivity effect size; — !data unavailable.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8REED, CHAN, AND MIKELS
Table 2
Effect Size Data for Individual Studies Included in the Meta-Analysis
Positivity bias
(d
bias
)Positivity effect
Study Condition Constraint NOlder Younger d
PE
[95% CI] Zp
Allard & Isaacowitz, 2008 Full attention N 40 0.55 0.49 0.06 [$0.59, 0.71] 0.18 0.86
Bannerman et al., 2011 N 60 0.65 $0.03 0.67 [0.15, 1.18] 2.53 0.01
Brassen et al., 2011 Y 43 0.22 0.11 0.11 [$0.47, 0.69] 0.37 0.71
Charles et al., 2003, Study 1 N 96 0.95 $0.01 0.95 [0.51, 1.38] 4.25 #.001
Charles et al., 2003, Study 2 N 64 $0.14 $0.55 0.41 [$0.1, 0.92] 1.55 0.12
Chung, 2010 N 123 0.34 $1.00 1.33 [0.89, 1.76] 5.95 #.001
Comblain et al., 2004 Y 40 $0.43 $0.44 0 [$0.65, 0.65] 0.00 1.00
D’Argembeau & Van der Linden, 2004 Y 64 0.06 0.32 $0.26 [$0.74, 0.22] $1.06 0.29
Ebner & Johnson, 2009, Study 1 Y 56 0.43 0.31 0.12 [$0.43, 0.67] 0.42 0.67
Ebner & Johnson, 2010, Study 2 Y 52 0.16 0.01 0.15 [$0.4, 0.7] 0.53 0.60
Emery & Hess, 2008, Study 1 Rating Y 56.5 $0.08 $0.34 0.25 [$0.26, 0.76] 0.94 0.34
Emery & Hess, 2008, Study 1 Watching N 56.5 0.00 $0.19 0.19 [$0.32, 0.7] 0.72 0.47
Emery & Hess, 2008, Study 2 Rating Y 44 0.31 $0.18 0.48 [$0.1, 1.06] 1.60 0.11
Emery & Hess, 2008, Study 2 Watching N 44 $0.23 $0.08 $0.15 [$0.73, 0.43] $0.50 0.62
Emery & Hess, 2011 Y 101 $0.38 $0.56 0.18 [$0.21, 0.57] 0.90 0.37
Feng et al., 2011 N 85 $0.44 $0.85 0.41 [$0.07, 0.89] 1.67 0.09
Fernandes et al., 2008 Y 97 0.08 0.28 $0.2 [$0.59, 0.19] $1.00 0.32
Fleming et al., 2003 Y 46 $0.19 0.03 $0.22 [$0.8, 0.36] $0.73 0.46
Fung et al., 2008 N 103 $0.19 $0.20 0.02 [$0.37, 0.41] 0.10 0.92
Fung et al., 2010, Study 1 N 217 1.60 $0.29 1.87 [1.53, 2.2] 10.80 #.001
Gallo et al., 2009 Y 48 $0.25 $0.41 0.15 [$0.43, 0.73] 0.50 0.62
Goeleven et al., 2010 Y 54 0.77 0.29 0.47 [$0.08, 1.02] 1.66 0.10
Grady et al., 2007 Y 70 $0.36 $0.78 0.42 [$0.09, 0.93] 1.59 0.11
Grühn et al., 2005 Y 144 $0.04 0.55 $0.59 [$0.92, $0.25] $3.41 #.001
Grühn et al., 2007 Y 96 0.02 $0.42 0.44 [0.04, 0.83] 2.20 0.03
Hahn et al., 2006, Study 1 Y 40 $0.31 $0.26 $0.05 [$0.66, 0.56] $0.16 0.87
Hahn et al., 2006, Study 2 Y 28 $0.09 $0.10 0 [$0.73, 0.73] 0.00 1.00
Hahn et al., 2006, Study 3 Y 30 $0.30 $0.21 $0.09 [$0.79, 0.61] $0.25 0.80
Isaacowitz et al., 2006a N 64 0.54 $0.13 0.66 [0.14, 1.17] 2.49 0.01
Isaacowitz et al., 2006b N 60 0.85 0.22 0.62 [0.06, 1.17] 2.19 0.03
Isaacowitz et al., 2008 N 124 0.45 0.12 0.32 [$0.01, 0.65] 1.85 0.06
Kapucu et al., 2008 Y 45 $0.31 $0.74 0.42 [$0.19, 1.03] 1.33 0.18
Kennedy et al., 2004 N 56 0.51 $0.25 0.75 [0.19, 1.3] 2.65 0.01
Kensinger & Schacter, 2008 Y 34 $0.57 $0.57 0 [$0.7, 0.7] 0.00 1.00
Kensinger et al., 2002 Y 40 $0.05 0.07 $0.12 [$0.73, 0.49] $0.38 0.70
Kensinger et al., 2007, Study 1 Y 60 $0.09 $0.48 0.38 [$0.13, 0.89] 1.44 0.15
Kensinger et al., 2007, Study 2 Y 60 0.03 $0.50 0.53 [0.01, 1.04] 2.00 0.05
Kensinger, 2008, Study 1 Y 60 0.42 $0.50 0.9 [0.38, 1.41] 3.40 #.001
Kensinger, 2008, Study 2 Y 60 0.41 $0.23 0.63 [0.11, 1.14] 2.38 0.02
Knight et al., 2002 Neutral mood N 97 0.31 0.02 0.28 [$0.15, 0.71] 1.25 0.21
Knight et al., 2002 Sad mood Y 100 0.13 0.30 $0.17 [$0.56, 0.22] $0.85 0.40
Knight et al., 2007 Divided attention Y 30 $0.17 0.27 $0.43 [$1.13, 0.27] $1.19 0.23
Knight et al., 2007 Full attention N 30 0.22 $0.21 0.41 [$0.29, 1.11] 1.14 0.26
Ko et al., 2011 American sample Y 55 $0.04 $0.95 0.89 [0.3, 1.47] 2.97 0.00
Ko et al., 2011 Korean sample Y 55 $0.14 $0.59 0.45 [$0.1, 1] 1.59 0.11
Kwon et al., 2009 N 104 0.12 $0.28 0.4 [0, 0.79] 2.00 0.05
Langeslag & Van Strien, 2008 Y 40 $0.13 $0.16 0.03 [$0.58, 0.64] 0.09 0.92
Langeslag & Van Strien, 2009 Y 38 0.15 $0.25 0.39 [$0.22, 1] 1.23 0.22
Leclerc & Kensinger, 2008 Y 48 $0.01 0.17 $0.18 [$0.73, 0.37] $0.64 0.52
Leclerc & Kensinger, 2010, Study 1 Y 36 $0.36 $0.03 $0.32 [$0.97, 0.33] $0.96 0.33
Leclerc & Kensinger, 2010, Study 2 Y 48 0.04 0.04 0 [$0.55, 0.55] 0.00 1.00
Leclerc & Kensinger, 2011 Y 39 0.11 $0.24 0.35 [$0.26, 0.96] 1.11 0.27
Leigland et al., 2004 Y 61 1.16 0.75 0.41 [$0.24, 1.06] 1.24 0.22
Li et al., 2011 N 62 $0.78 $1.23 0.44 [$0.17, 1.05] 1.39 0.16
Löckenhoff & Carstensen, 2007 Control N 40 0.82 0.13 0.67 [0, 1.34] 1.93 0.05
Löckenhoff & Carstensen, 2007 Emotion-focus Y 40 0.67 0.30 0.36 [$0.29, 1.01] 1.09 0.28
Löckenhoff & Carstensen, 2007 Information-focus Y 40 0.34 0.04 0.29 [$0.36, 0.94] 0.87 0.38
Löckenhoff & Carstensen, 2008 Different Age Y 47.32 0.53 0.14 0.38 [$0.2, 0.96] 1.27 0.21
Löckenhoff & Carstensen, 2008 Same Age Y 47.32 0.89 $0.08 0.95 [0.33, 1.56] 3.00 0.00
Löckenhoff & Carstensen, 2008 Self (control) N 47.32 0.89 0.03 0.84 [0.22, 1.45] 2.66 0.01
Majerus & D’Argembeau, 2011, Study 2 Y 30 0.19 0.20 $0.01 [$0.71, 0.69] $0.03 0.98
(table continues)
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
9
POSITIVITY EFFECT META-ANALYSIS
process positive (vs. negative) information when their processing is
unconstrained but process positive and negative information equally
when constraints are present. By contrast, younger adults preferen-
tially process negative (vs. positive) information under constrained
and unconstrained processing conditions.
The present study represents to our knowledge the first system-
atic meta-analysis of the age-related positivity effect and supports
key theoretically based predictions from a precise operational
definition. These predictions were previously supported by some,
but not all, published studies. For instance, although Löckenhoff
and Carstensen (2007) demonstrated that experimental constraints
on goals eliminate positivity in decision making, Emery and Hess
(2008) found minimal evidence for the influence of goal manipu-
lations on emotional memory. Studies that explicitly constrain
Table 2 (continued)
Positivity bias
(d
bias
)Positivity effect
Study Condition Constraint NOlder Younger d
PE
[95% CI] Zp
Majerus & D’Argembeau, 2011, Study 3 Y 30 0.34 $0.03 0.36 [$0.34, 1.06] 1.00 0.32
Mather & Carstensen, 2003, Study 1 N 104 0.17 0.12 0.05 [$0.34, 0.44] 0.25 0.80
Mather & Carstensen, 2003, Study 2 N 88 0.14 0.07 0.07 [$0.36, 0.5] 0.31 0.75
Mather & Knight, 2005, Study 1 N 96 0.15 $0.31 0.46 [0.06, 0.85] 2.30 0.02
Mather & Knight, 2005, Study 2 N 56 $0.32 $1.34 1.01 [0.35, 1.66] 3.05 0.00
Mather & Knight, 2005, Study 3 Divided attention Y 32 $0.79 $0.43 $0.35 [$1.08, 0.38] $0.94 0.35
Mather & Knight, 2005, Study 3 Full attention N 32 0.15 $0.25 0.39 [$0.28, 1.06] 1.13 0.26
Mather & Knight, 2006 Y 68 $0.16 $0.25 0.09 [$0.39, 0.57] 0.37 0.71
Mather et al., 2005, Study 3 N 80 0.31 $0.07 0.38 [$0.05, 0.81] 1.70 0.09
Mather et al., 2005, Study 4b N 92 1.13 0.76 0.37 [$0.11, 0.85] 1.51 0.13
McKay-Nesbitt et al., 2011 N 259 0.00 $0.25 0.25 [$0.02, 0.52] 1.77 0.08
Mickley & Kensinger, 2009, Study 1 Y 52 $0.24 0.18 $0.42 [$0.97, 0.13] $1.48 0.14
Mickley & Kensinger, 2009, Study 2 Y 49 $0.50 0.03 $0.52 [$1.07, 0.03] $1.84 0.07
Mickley Steinmetz et al., 2010 Y 47 $0.10 $0.28 0.18 [$0.4, 0.76] 0.60 0.55
Mienaltowski et al., 2011 Y 31 $0.01 0.01 $0.03 [$0.7, 0.64] $0.09 0.93
Mikels et al., 2005 Y 40 0.55 $0.43 0.96 [0.3, 1.61] 2.89 0.00
Nashiro & Mather, 2011a, Study 1 Y 36 0.48 0.11 0.37 [$0.28, 1.02] 1.12 0.26
Nashiro & Mather, 2011a, Study 2 Y 48 0.04 0.30 $0.26 [$0.81, 0.29] $0.92 0.36
Nashiro & Mather, 2011b Y 48 0.09 $0.47 0.55 [$0.03, 1.13] 1.83 0.07
Nashiro et al., 2011, Study 1 Y 35 0.79 0.04 0.73 [0.02, 1.43] 2.02 0.04
Nashiro et al., 2011, Study 2 Y 40 0.67 $0.07 0.73 [0.07, 1.38] 2.20 0.03
Nikitin & Freund, 2011 N 168 1.02 1.18 $0.16 [$0.55, 0.23] $0.80 0.42
Noh & Isaacowitz, 2011 Y 86 $0.05 0.03 $0.08 [$0.51, 0.35] $0.36 0.72
Orgeta, 2011, Study 1 N 80 0.00 $0.02 0.02 [$0.41, 0.45] 0.09 0.93
Orgeta, 2011, Study 2 N 80 0.08 0.00 0.08 [$0.35, 0.51] 0.36 0.72
Piguet et al., 2008 Y 72 $0.05 0.04 $0.09 [$0.52, 0.34] $0.40 0.69
Pruis et al., 2009 Y 51 $0.32 $0.92 0.59 [0, 1.17] 1.97 0.05
Ready et al., 2007, Study 1 N 49 0.28 $0.34 0.61 [0.05, 1.16] 2.16 0.03
Ready et al., 2007, Study 2 N 34 1.08 $1.00 2.04 [1.23, 2.84] 4.95 #.001
Rendell et al., 2011 Y 60 0.58 0.32 0.25 [$0.3, 0.8] 0.88 0.38
Ritchey et al., 2011 Y 36 $0.46 $0.69 0.23 [$0.47, 0.93] 0.64 0.52
Rösler et al., 2005 N 24 0.06 $0.30 0.35 [$0.43, 1.13] 0.88 0.38
Samanez-Larkin et al., 2009 Y 24 $0.18 0.14 $0.31 [$1.09, 0.47] $0.78 0.44
Savaskan et al., 2007 Y 45 0.29 0.56 $0.26 [$0.91, 0.39] $0.78 0.43
Shamaskin et al., 2010 Y 49 2.41 1.14 1.25 [0.33, 2.16] 2.67 0.01
Spaniol et al., 2008, Study 1 Y 48 $0.65 $0.52 $0.13 [$0.78, 0.52] $0.39 0.70
Spaniol et al., 2008, Study 2 N 47 $0.53 $0.68 0.15 [$0.46, 0.76] 0.47 0.64
St Jacques et al., 2009 Y 30 $0.17 $0.20 0.02 [$0.68, 0.72] 0.06 0.96
Thapar & Rouder, 2009 Y 60 0.22 0.03 0.19 [$0.32, 0.7] 0.72 0.47
Thomas & Hasher, 2006 Y 96 0.25 $0.25 0.49 [0.09, 0.88] 2.45 0.01
Tomaszczyk et al., 2008 Active processing Y 72 0.23 $0.16 0.39 [$0.09, 0.87] 1.59 0.11
Tomaszczyk et al., 2008 Passive processing N 72 0.14 $0.32 0.45 [$0.03, 0.93] 1.84 0.07
Van Gerven & Murphy, 2010 Y 96 0.06 $0.18 0.23 [$0.16, 0.62] 1.15 0.25
Waring & Kensinger, 2009 Y 48 $0.10 $0.06 $0.04 [$0.59, 0.51] $0.14 0.89
Werheid et al., 2010, Study 1 Y 40 $0.35 $0.46 0.1 [$0.51, 0.71] 0.32 0.75
Werheid et al., 2010, Study 2 Y 40 $0.12 $0.21 0.09 [$0.52, 0.7] 0.28 0.78
Werheid et al., 2010, Study 4 Y 24 0.45 0.00 0.44 [$0.34, 1.22] 1.10 0.27
Williams & Drolet, 2005, Study 2 Control Y 82 $0.92 0.96 $1.87 [$2.38, $1.35] $7.07 #.001
Williams & Drolet, 2005, Study 2 Expansive time Y 82 1.33 1.39 $0.06 [$0.64, 0.52] $0.20 0.84
Williams & Drolet, 2005, Study 2 Limited time Y 82 $1.14 $0.83 $0.31 [$0.82, 0.2] $1.17 0.24
Yang & Hasher, 2011 N 106 0.13 $0.60 0.72 [0.32, 1.11] 3.60 #.001
Yang & Ornstein, 2011 Control N 24 $0.31 $0.63 0.31 [$0.49, 1.11] 0.75 0.45
Yang & Ornstein, 2011 Emotion-focused Y 39 $0.38 $0.13 $0.24 [$0.85, 0.37] $0.76 0.45
Yang & Ornstein, 2011 Information-focused Y 24 $0.09 $0.57 0.47 [$0.33, 1.27] 1.14 0.25
Note. Sample size (N) represents total sample (older (younger) within each study.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
10 REED, CHAN, AND MIKELS
processing to test the resulting effects on positivity are few and far
between in the research literature, yet studies that inadvertently
constrained processing are relatively common. Using a meta-
analytic approach circumvents this issue because it ensures that the
latter category of studies is evaluated in the appropriate context.
This is a key contribution of the present study because ignoring
differences in task constraint across experiments leads one to
significantly underestimate the size of the positivity effect. Not
only was the positivity effect significantly mitigated in studies
with information processing constraints, but older adults’ prefer-
ential processing of positive information (over negative) was com-
pletely eliminated under these conditions. Thus, the inclusion of
experimental constraints on processing not only skew age differ-
ences but also provide a misleading portrait of information pro-
cessing tendencies within each age group.
Our finding that the size of the positivity effect scales with
the magnitude of age differences is also theoretically consistent
despite scant empirical support from individual studies. Re-
search by Fung and colleagues using adult life span samples
found a linear association between age and positivity in visual
attention (Li et al., 2011) as well as age differences in positivity
between older and younger— but not middle-aged—adults
(Fung, Isaacowitz, Lu, & Li, 2010), suggesting a gradual age
trend. Critically, though the present analysis focuses primarily
on comparisons between young and older adults, the positivity
effect theoretically reflects relative age comparisons. As a
consequence one could observe a positivity effect among a
sample comprised entirely of older adults if it contained a
sufficiently wide age range (e.g., 60 to 100). Of course, the vast
majority of studies on the positivity effect neither include a
broad life span sample nor a comparison of younger, middle-
aged and older adults, thus complicating efforts to identify the
developmental trajectory of positivity. Applying a meta-
analytic perspective partially remedies this problem by allow-
ing between-study comparisons of variable age differences to
compensate for limited within-study age comparisons. How-
ever, the cross-sectional nature of the studies included in the
meta-analysis limits the ability to draw conclusions regarding
developmental changes. Longitudinal or sequential studies are
therefore needed to directly test whether the positivity effect is
indeed a developmental trend.
Although our meta-analysis was intended to examine the
reliability, size and moderators of the positivity effect, our
approach precluded addressing questions about underlying
mechanisms, such as whether the effect is driven by age dif-
ferences in increased positive and/or decreased negative pro-
cessing. In contrast to the prior meta-analysis by Murphy and
Isaacowitz (2008), we included studies that lacked measures of
neutral information processing (e.g., Kennedy et al., 2004;
Shamaskin et al., 2011) that would serve as necessary controls
for separately comparing positive versus negative information
processing across age groups. Our results therefore leave open
questions concerning the mechanisms underlying the positivity
effect. For instance, is the positivity effect driven by discrete
changes in the processing of positive and negative information
and, if so, what factors moderate whether age differences are
manifest for one valence versus the other? These questions
represent fertile ground for future meta-analyses, and the dra-
matic expansion of the empirical literature in the years since
Murphy and Isaacowitz’s (2008) study renders a systematic
replication of the latter particularly useful.
Although the present meta-analysis is not able to directly
address specific mechanisms underlying the positivity effect, it
does provide novel insights into the overall pattern. Consistent
with previous reviews (see, e.g., Baumeister et al., 2001), the
present analysis provides additional evidence of a negativity
bias in youth. As discussed earlier, the positivity effect captures
an age-related shift in favor of positive versus negative infor-
mation. Until now, though, it was unclear whether the shift is a
reduced negativity bias (or no negativity bias) or a full-blown
positivity bias. Our overall findings support the latter: Collaps-
ing across all studies, older adults showed a positivity bias on
average. However, in studies that constrained information pro-
cessing they showed neither a positivity bias nor a negativity
bias, whereas younger adults showed a negativity bias. Thus,
the exact configuration of the age-related positivity effect—
whether a shift from negativity in youth to positivity in later life
Table 3
Positivity Effect and Positivity Bias Effect Sizes by Processing Constraint and Age Group
Positivity bias (d
bias
)
Processing constraint kOlder Younger Positivity effect (d
PE
)
Unconstrained 38 .280 [.123, .436] $.197 [$.340, $.054] .482 [.323, .640]
Constrained 77 .048 [$.048, .144] $.085 [$.182, .011] .134 [.033, .235]
Total 115 .128 [.042, .214] $.123 [$.203, $.043] .257 [.165, .349]
Note. 95% confidence intervals are in brackets.
Figure 1. Meta-regression of mean age difference on positivity effect
size. Each circle represents a single study sample. Circle sizes are propor-
tional to the sample size. Regression fit line '!.005, t(113) !5.69, p#
.001.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
POSITIVITY EFFECT META-ANALYSIS
or an elimination of negativity in later life—may hinge on
outside influences on information processing. This result places
a significant life span qualification on the pervasiveness of the
negativity bias; although in youth “bad is stronger than good”
(Baumeister et al., 2001), in later life good is just as strong as
bad—if not stronger.
Results of our meta-analysis have prescriptive value for
future research on the positivity effect. Because of the dispro-
portionate impact of processing constraints on positivity, it is
imperative to recognize and assess the consequences of often-
unintended methodological alterations in this area of study. As
illustrated by the present meta-analysis, merely notifying par-
ticipants that their memory will be tested before presenting
stimuli significantly reduces age differences in preferential
processing, in large part by eliminating older adults’ natural
prioritization of positive over negative information. Studies that
intend to examine normative age differences in the processing
of emotional information would therefore benefit from encour-
aging and enabling participants to process information in
whichever way they prefer and removing all potential con-
straints or barriers.
The present findings bring needed clarity to an increasingly
crowded and complex literature and reinforce the utility of
meta-analysis for summarizing extant work. However, future
empirical studies are needed to address critical questions re-
garding alternative influences on the positivity effect. For in-
stance, while individual studies have shown that the positivity
effect depends on sufficient cognitive control resources (Knight
et al., 2007;Mather & Knight, 2005), the present meta-analysis
could not examine this moderator. This was due to a lack of
consistent selection and reporting of background cognitive abil-
ity measures across studies. Future studies incorporating com-
mon measures of cognitive control resources
10
would afford
testing this key tenet of the motivational perspective on posi-
tivity effect. Another limitation of the present meta-analysis is
that it could not examine cross-cultural differences in positivity
because of the relatively few number of studies with non-
Western samples. Understanding the influence of culture on
positivity from a meta-research perspective is especially impor-
tant given the inconsistent results stemming from individual
studies (e.g., Fung et al., 2008,2010;Kwon et al., 2009).
Though cultural comparisons and perspectives are becoming
increasingly common in this area, more work is needed to
understand how the positivity effect is manifest (or not) across
cultures and the underlying mechanisms therein. Another lim-
itation of the current meta-analysis is that it could not address
the context sensitivity of the positivity effect beyond experi-
mental constraints. Although the motivational perspective pro-
poses that age differences in emotional information processing
will be mitigated in high-stakes versus low-stakes situations
(i.e., deciding on a life-saving medical procedure vs. choosing
a hypothetical car; for a discussion, see Reed & Carstensen,
2012), the vast majority of extant studies use tasks with mini-
mal significance to participants. Future research with relatively
consequential information processing contexts could trace the
conditions under which older and younger adults differ in their
emotional memory and attention.
In sum, the present meta-analysis brings some measure of
clarity to a domain that has produced a remarkable quantity of
research—and no shortage of debate—in the mere decade since
its inception. Although evaluating studies in a serial manner
might indeed lend the impression of disorder in the literature
and raise doubts over the validity of the positivity effect, using
a meta-analytic perspective paints a different picture: Discrep-
ant results observed across tests of the positivity effect become
both meaningful and relatively predictable, reflecting theoreti-
cally implicated moderators. When these factors are taken into
consideration, empirical findings coalesce to reveal a positivity
effect that is both reliable and robust.
10
For example, the Attentional Network Task by Fan and colleagues
(Fan, McCandliss, Sommer, Raz, & Posner, 2002).
References
!
References included in the meta-analysis are marked with an asterisk
below.
!
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, Neuropsy-
chology, and Cognition, 15, 725–743. doi:10.1080/13825580802348562
!
Bannerman, R. L., Regener, P., & Sahraie, A. (2011). Binocular rivalry:
A window into emotional processing in aging. Psychology and Aging,
26, 372–380. doi:10.1037/a0022029
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001).
Bad is stronger than good. Review of General Psychology, 5, 323–370.
doi:10.1037/1089-2680.5.4.323
Becker, B. J. (1988). Synthesizing standardized mean-change measures.
British Journal of Mathematical and Statistical Psychology, 41, 257–
278. doi:10.1111/j.2044-8317.1988.tb00901.x
Borenstein, M. (2009). Effect sizes for continuous data. In H. Cooper, L. V.
Hedges, & J. C. Valentine (Eds.), The handbook of research synthesis
and meta-analysis (2nd ed.; pp. 221–235). New York, NY: Russell Sage
Foundation.
!
Brassen, S., Gamer, M., & Büchel, C. (2011). Anterior cingulate activa-
tion is related to a positivity bias and emotional stability in successful
aging. Biological Psychiatry, 70, 131–137. doi:10.1016/j.biopsych.2010
.10.013
Broster, L. S., Blonder, L. X., & Jiang, Y. (2012). Does emotional memory
enhancement assist the memory-impaired? Frontiers in Aging Neurosci-
ence, 4, 2. doi:10.3389/fnagi.2012.00002
Cacioppo, J. T., Berntson, G. G., Bechara, A., Tranel, D., & Hawkley, L. C.
(2011). Could an aging brain contribute to subjective well-being? The
value added by a social neuroscience perspective. In A. Todorov, S. T.
Fiske, & D. A. Prentice (Eds.), Social neuroscience: Toward under-
standing the underpinnings of the social mind (pp. 249 –262). New
York, NY: Oxford University Press.
Carstensen, L. L. (2006). The influence of a sense of time on human
development. Science, 312, 1913–1915. doi:10.1126/science.1127488
Carstensen, L. L., & Mikels, J. A. (2005). At the intersection of emotion
and cognition: Aging and the positivity effect. Current Directions in
Psychological Science, 14, 117–121. doi:10.1111/j.0963-7214.2005
.00348.x
Carstensen, L. L., Pasupathi, M., Mayr, U., & Nesselroade, J. R. (2000).
Emotional experience in everyday life across the adult life span. Journal
of Personality and Social Psychology, 79, 644 – 655. doi:10.1037/0022-
3514.79.4.644
Charles, S. T., & Carstensen, L. L. (2010). Social and emotional aging.
Annual Review of Psychology, 61, 383– 409. doi:10.1146/annurev.psych
.093008.100448
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
12 REED, CHAN, AND MIKELS
!
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
!
Chung, C. (2010). Effects of view of life and selection bias on emotional
memory in old age. GeroPsych, 23, 161–168. doi:10.1024/1662-9647/
a000019
Cohen, J. (1988). Statistical power analysis for the behavioral sciences
(2nd ed.). Hillsdale, NJ: Erlbaum.
!
Comblain, C., D’Argembeau, A., Van der Linden, M., & Aldenhoff, L.
(2004). The effect of ageing on the recollection of emotional and neutral
pictures. Memory, 12, 673– 684. doi:10.1080/09658210344000477
!
D’Argembeau, A., & Van der Linden, M. (2004). Identity but not expres-
sion memory for unfamiliar faces is affected by ageing. Memory, 12,
644 – 654. doi:10.1080/09658210344000198
Duval, S., & Tweedie, R. (2000). Trim and fill: A simple funnel-plot-based
method of testing and adjusting for publication bias in meta-analysis.
Biometrics, 56, 455– 463. doi:10.1111/j.0006-341X.2000.00455.x
!
Ebner, N. C., & Johnson, M. K. (2009). Young and older emotional faces:
Are there age group differences in expression identification and mem-
ory? Emotion, 9, 329 –339. doi:10.1037/a0015179
!
Ebner, N. C., & Johnson, M. K. (2010). Age-group differences in inter-
ference from young and older emotional faces. Cognition and Emotion,
24, 1095–1116. doi:10.1080/02699930903128395
Ebner, N. C., Johnson, M. K., & Fischer, H. (2012). Neural mechanisms of
reading facial emotions in young and older adults. Frontiers in Psychol-
ogy, 3, 1–19.
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in
meta-analysis detected by a simple, graphical test. British Medical
Journal, 315, 629 – 634. doi:10.1136/bmj.315.7109.629
!
Emery, L., & Hess, T. M. (2008). Viewing instructions impact emotional
memory differently in older and young adults. Psychology and Aging,
23, 2–12. doi:10.1037/0882-7974.23.1.2
!
Emery, L., & Hess, T. M. (2011). Cognitive consequences of expressive
regulation in older adults. Psychology and Aging, 26, 388 –396. doi:
10.1037/a0020041
Fan, J., McCandliss, B. D., Sommer, T., Raz, A., & Posner, M. I.
(2002). Testing the efficiency and independence of attentional net-
works. Journal of Cognitive Neuroscience, 14, 340 –347. doi:
10.1162/089892902317361886
!
Feng, M. C., Courtney, C. G., Mather, M., Dawson, M. E., & Davison,
G. C. (2011). Age-related affective modulation of the startle eyeblink
response: Older adults startle most when viewing positive pictures.
Psychology and Aging, 26, 752–760. doi:10.1037/a0023110
Ferguson, C. J., & Brannick, M. T. (2012). Publication bias in psycholog-
ical science: Prevalence, methods for identifying and controlling, and
implications for the use of meta-analyses. Psychological Methods, 17,
120 –128. doi:10.1037/a0024445
!
Fernandes, M., Ross, M., Wiegand, M., & Schryer, E. (2008). Are the
memories of older adults positively biased? Psychology and Aging, 23,
297–306. doi:10.1037/0882-7974.23.2.297
!
Fleming, K., Kim, S. H., Doo, M., Maguire, G., & Potkin, S. G. (2003).
Memory for emotional stimuli in patients with Alzheimer’s disease.
American Journal of Alzheimer’s Disease and Other Dementias, 18,
340 –342. doi:10.1177/153331750301800604
!
Fung, H. H., Isaacowitz, D. M., Lu, A. Y., & Li, T. (2010). Interdependent
self-construal moderates the age-related negativity reduction effect in
memory and visual attention. Psychology and Aging, 25, 321–329.
doi:10.1037/a0019079
!
Fung, H. H., Lu, A. Y., Goren, D., Isaacowitz, D. M., Wadlinger, H. A.,
& Wilson, H. R. (2008). Age-related positivity enhancement is not
universal: Older Chinese look away from positive stimuli. Psychology
and Aging, 23, 440 – 446. doi:10.1037/0882-7974.23.2.440
!
Gallo, D. A., Foster, K. T., & Johnson, E. L. (2009). Elevated false
recollection of emotional pictures in young and older adults. Psychology
and Aging, 24, 981–988. doi:10.1037/a0017545
!
Goeleven, E., De Raedt, R., & Dierckx, E. (2010). The positivity effect in
older adults: The role of affective interference and inhibition. Aging &
Mental Health, 14, 129 –137. doi:10.1080/13607860903228754
!
Grady, C. L., Keightley, M., Hongwanishkul, D., Lee, W., & Hasher, L.
(2007). The effect of age on memory for emotional faces. Neuropsy-
chology, 21, 371–380. doi:10.1037/0894-4105.21.3.371
Grühn, D., & Scheibe, S. (2008). Age-related differences in valence and
arousal ratings of pictures from the International Affective Picture
System (IAPS): Do ratings become more extreme with age? Behavior
Research Methods, 40, 512–521. doi:10.3758/BRM.40.2.512
!
Grühn, D., Scheibe, S., & Baltes, P. B. (2007). Reduced negativity effect
in older adults’ memory for emotional pictures: The heterogeneity-
homogeneity list paradigm. Psychology and Aging, 22, 644 – 649. doi:
10.1037/0882-7974.22.3.644
!
Grühn, D., Smith, J., & Baltes, P. B. (2005). No aging bias favoring
memory for positive material: Evidence from a heterogeneity-
homogeneity list paradigm using emotionally toned words. Psychology
and Aging, 20, 579 –588. doi:10.1037/0882-7974.20.4.579
!
Hahn, S., Carlson, C., Singer, S., & Gronlund, S. D. (2006). Aging and
visual search: Automatic and controlled attentional bias to threat faces.
Acta Psychologica, 123, 312–336. doi:10.1016/j.actpsy.2006.01.008
!
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
!
Isaacowitz, D. M., Wadlinger, H. A., Goren, D., & Wilson, H. R. (2006a).
Is there an age-related positivity effect in visual attention? A comparison
of two methodologies. Emotion, 6, 511–516. doi:10.1037/1528-3542.6
.3.511
!
Isaacowitz, D. M., Wadlinger, H. A., Goren, D., & Wilson, H. R. (2006b).
Selective preference in visual fixation away from negative images in old
age? An eye-tracking study. Psychology and Aging, 21, 40 – 48. doi:
10.1037/0882-7974.21.1.40
Johnson, D. R., & Whiting, W. L. (2013). Detecting subtle expressions:
Older adults demonstrate automatic and controlled positive response
bias in emotional perception. Psychology and Aging, 28, 172–178.
doi:10.1037/a0029914
!
Kapucu, A., Rotello, C. M., Ready, R. E., & Seidl, K. N. (2008). Response
bias in “remembering” emotional stimuli: A new perspective on age
differences. Journal of Experimental Psychology: Learning, Memory,
and Cognition, 34, 703–711. doi:10.1037/0278-7393.34.3.703
!
Kennedy, Q., Mather, M., & Carstensen, L. L. (2004). The role of
motivation in the age-related positivity effect in autobiographical mem-
ory. Psychological Science, 15, 208 –214. doi:10.1111/j.0956-7976.2004
.01503011.x
!
Kensinger, E. A. (2008). Age differences in memory for arousing and
nonarousing emotional words. The Journals of Gerontology: Series B:
Psychological Sciences and Social Sciences, 63B, P13–P18. doi:
10.1093/geronb/63.1.P13
!
Kensinger, E. A., Brierley, B., Medford, N., Growdon, J. H., & Corkin, S.
(2002). Effects of normal aging and Alzheimer’s disease on emotional
memory. Emotion, 2, 118 –134. doi:10.1037/1528-3542.2.2.118
!
Kensinger, E. A., O’Brien, J. L., Swanberg, K., Garoff-Eaton, R. J., &
Schacter, D. L. (2007). The effects of emotional content on reality-
monitoring performance in young and older adults. Psychology and
Aging, 22, 752–764. doi:10.1037/0882-7974.22.4.752
!
Kensinger, E. A., & Schacter, D. L. (2008). Neural processes supporting
young and older adults’ emotional memories. Journal of Cognitive
Neuroscience, 20, 1161–1173. doi:10.1162/jocn.2008.20080
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
13
POSITIVITY EFFECT META-ANALYSIS
!
Knight, B. G., Maines, M. L., & Robinson, G. S. (2002). The effects of sad
mood on memory in older adults: A test of the mood congruence effect.
Psychology and Aging, 17, 653– 661. doi:10.1037/0882-7974.17.4.653
!
Knight, M., Seymour, T. L., Gaunt, J. T., Baker, C., Nesmith, K., &
Mather, M. (2007). Aging and goal-directed emotional attention: Dis-
traction reverses emotional biases. Emotion, 7, 705–714. doi:10.1037/
1528-3542.7.4.705
!
Ko, S.-G., Lee, T.-H., Yoon, H.-Y., Kwon, J.-H., & Mather, M. (2011).
How does context affect assessments of facial emotion? The role of
culture and age. Psychology and Aging, 26, 48 –59. doi:10.1037/
a0020222
Konstantopoulos, S., & Hedges, L. V. (2009). Fixed effects models in
meta-analysis. In H. Cooper, L. V. Hedges, & J. C. Valentine (Eds.), The
handbook of research synthesis and meta-analysis (2nd ed., pp. 279
294). New York, NY: Russell Sage Foundation.
!
Kwon, Y., Scheibe, S., Samanez-Larkin, G. R., Tsai, J. L., & Carstensen,
L. L. (2009). Replicating the positivity effect in picture memory in
Koreans: Evidence for cross-cultural generalizability. Psychology and
Aging, 24, 748 –754. doi:10.1037/a0016054
Labouvie-Vief, G., Grühn, D., & Studer, J. (2010). Dynamic integration of
emotion and cognition: Equilibrium regulation in development and
aging. In R. M. Lerner, M. E. Lamb, & A. M. Freund (Eds.), The
handbook of life-span development (Vol. 2, pp. 79 –115). Hoboken, NJ:
Wiley. doi:10.1002/9780470880166.hlsd002004
!
Langeslag, S. J. E., & Van Strien, J. W. (2008). Age differences in the
emotional modulation of ERP old/new effects. International Journal of
Psychophysiology, 70, 105–114. doi:10.1016/j.ijpsycho.2008.07.022
!
Langeslag, S. J. E., & Van Strien, J. W. (2009). Aging and emotional
memory: The co-occurrence of neurophysiological and behavioral pos-
itivity effects. Emotion, 9, 369 –377. doi:10.1037/a0015356
!
Leclerc, C. M., & Kensinger, E. A. (2008). Effects of age on detection of
emotional information. Psychology and Aging, 23, 209 –215.
!
Leclerc, C. M., & Kensinger, E. A. (2010). Age-related valence-based rever-
sal in recruitment of medial prefrontal cortex on a visual search task. Social
Neuroscience, 5, 560 –576. doi:10.1080/17470910903512296
!
Leclerc, C. M., & Kensinger, E. A. (2011). Neural processing of emo-
tional pictures and words: A comparison of young and older adults.
Developmental Neuropsychology, 36, 519 –538. doi:10.1080/87565641
.2010.549864
!
Leigland, L. A., Schulz, L. E., & Janowsky, J. S. (2004). Age related
changes in emotional memory. Neurobiology of Aging, 25, 1117–1124.
doi:10.1016/j.neurobiolaging.2003.10.015
!
Li, T., Fung, H. H., & Isaacowitz, D. M. (2011). The role of dispositional
reappraisal in the age-related positivity effect. The Journals of Geron-
tology: Series B: Psychological Sciences and Social Sciences, 66B,
56 – 60. doi:10.1093/geronb/gbq074
!
Löckenhoff, C. E., & Carstensen, L. L. (2007). Aging, emotion, and
health-related decision strategies: Motivational manipulations can re-
duce age differences. Psychology and Aging, 22, 134 –146. doi:10.1037/
0882-7974.22.1.134
!
Löckenhoff, C. E., & Carstensen, L. L. (2008). Decision strategies in
health care choices for self and others: Older but not younger adults
make adjustments for the age of the decision target. The Journals of
Gerontology: Series B: Psychological Sciences and Social Sciences,
63B, P106 –P109. doi:10.1093/geronb/63.2.P106
!
Majerus, S., & D’Argembeau, A. (2011). Verbal short-term memory
reflects the organization of long-term memory: Further evidence from
short-term memory for emotional words. Journal of Memory and Lan-
guage, 64, 181–197. doi:10.1016/j.jml.2010.10.003
Mather, M., Canli, T., English, T., Whitfield, S., Wais, P., Ochsner, K.,...
Carstensen, L. L. (2004). Amygdala responses to emotionally valenced
stimuli in older and younger adults. Psychological Science, 15, 259 –
263. doi:10.1111/j.0956-7976.2004.00662.x
!
Mather, M., & Carstensen, L. L. (2003). Aging and attentional biases for
emotional faces. Psychological Science, 14, 409 – 415. doi:10.1111/
1467-9280.01455
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
!
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:10.1037/0882-7974.20.4.554
!
Mather, M., & Knight, M. (2006). Angry faces get noticed quickly: Threat
detection is not impaired among older adults. The Journals of Geron-
tology: Series B: Psychological Sciences and Social Sciences, 61B,
P54 –P57. doi:10.1093/geronb/61.1.P54
!
Mather, M., Knight, M., & McCaffrey, M. (2005). The allure of the
alignable: Younger and older adults’ false memories of choice features.
Journal of Experimental Psychology: General, 134, 38 –51. doi:10.1037/
0096-3445.134.1.38
!
McKay-Nesbitt, J., Manchanda, R. V., Smith, M. C., & Huhmann, B. A.
(2011). Effects of age, need for cognition, and affective intensity on
advertising effectiveness. Journal of Business Research, 64, 12–17.
doi:10.1016/j.jbusres.2009.09.013
!
Mickley, K. R., & Kensinger, E. A. (2009). Phenomenological character-
istics of emotional memories in younger and older adults. Memory, 17,
528 –543. doi:10.1080/09658210902939363
!
Mickley Steinmetz, K. R., Muscatell, K. A., & Kensinger, E. A. (2010).
The effect of valence on young and older adults’ attention in a rapid
serial visual presentation task. Psychology and Aging, 25, 239 –245.
doi:10.1037/a0018297
!
Mienaltowski, A., Corballis, P. M., Blanchard-Fields, F., Parks, N. A., &
Hilimire, M. R. (2011). Anger management: Age differences in emo-
tional modulation of visual processing. Psychology and Aging, 26,
224 –231. doi:10.1037/a0021032
!
Mikels, J. A., Larkin, G. R., Reuter-Lorenz, P. A., & Carstensen, L. L.
(2005). Divergent trajectories in the aging mind: Changes in working
memory for affective versus visual information with age. Psychology
and Aging, 20, 542–553. doi:10.1037/0882-7974.20.4.542
Mikels, J. A., Reed, A. E., Hardy, L. N., & Löckenhoff, C. E. (in press).
Positive emotions across the adult life span. In M. Tugade, L. Shiota, &
L. Kirby (Eds.), Handbook of positive emotions. New York, NY: Guil-
ford Press.
Morris, S. B., & DeShon, R. P. (2002). Combining effect size estimates in
meta-analysis with repeated measures and independent-groups designs.
Psychological Methods, 7, 105–125. doi:10.1037/1082-989X.7.1.105
Murphy, N. A., & Isaacowitz, D. M. (2008). Preferences for emotional
information in older and younger adults: A meta-analysis of memory and
attention tasks. Psychology and Aging, 23, 263–286. doi:10.1037/0882-
7974.23.2.263
!
Nashiro, K., & Mather, M. (2011a). Effects of emotional arousal on
memory binding in normal aging and Alzheimer’s disease. The Ameri-
can Journal of Psychology, 124, 301–312. doi:10.5406/amerjpsyc.124
.3.0301
!
Nashiro, K., & Mather, M. (2011b). How arousal affects younger and
older Adults’ memory binding. Experimental Aging Research, 37,
108P54 –128. doi:10.1080/0361073X.2011.536746
!
Nashiro, K., Mather, M., Gorlick, M. A., & Nga, L. (2011). Negative
emotional outcomes impair older adults’ reversal learning. Cognition
and Emotion, 25, 1014 –1028. doi:10.1080/02699931.2010.542999
!
Nikitin, J., & Freund, A. M. (2011). Age and motivation predict gaze
behavior for facial expressions. Psychology and Aging, 26, 695–700.
doi:10.1037/a0023281
!
Noh, S. R., & Isaacowitz, D. M. (2011). Age differences in the emotional
modulation of attention: Effects of own-age versus other-age emotional
face cues on the alerting and orienting networks. Journal of Cognitive
Psychology, 23, 709 –722.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
14 REED, CHAN, AND MIKELS
!
Orgeta, V. (2011). Avoiding threat in late adulthood: Testing two life span
theories of emotion. Experimental Aging Research, 37, 449 – 472. doi:
10.1080/0361073X.2011.590759
!
Piguet, O., Connally, E., Krendl, A. C., Huot, J. R., & Corkin, S. (2008).
False memory in aging: Effects of emotional valence on word recogni-
tion accuracy. Psychology and Aging, 23, 307–314. doi:10.1037/0882-
7974.23.2.307
!
Pruis, T. A., Neiss, M. B., Leigland, L. A., & Janowsky, J. S. (2009).
Estrogen modifies arousal but not memory for emotional events in older
women. Neurobiology of Aging, 30, 1296 –1304. doi:10.1016/j
.neurobiolaging.2007.11.009
!
Rahhal, T. A., May, C. P., & Hasher, L. (2002). Truth and character:
Sources that older adults can remember. Psychological Science, 13,
101–105. doi:10.1111/1467-9280.00419
!
Ready, R. E., Weinberger, M. I., & Jones, K. M. (2007). How happy have
you felt lately? Two diary studies of emotion recall in older and younger
adults. Cognition and Emotion, 21, 728 –757. doi:10.1080/
02699930600948269
!
Reed, A. E., & Carstensen, L. L. (2012). The theory behind the age-related
positivity effect. Frontiers in Psychology, 3, 1–9.
!
Rendell, P. G., Phillips, L. H., Henry, J. D., Brumby-Rendell, T., de la
Piedad Garcia, X., Altgassen, M., & Kliegel, M. (2011). Prospective
memory, emotional valence and ageing. Cognition and Emotion, 25,
916 –925. doi:10.1080/02699931.2010.508610
!
Ritchey, M., Bessette-Symons, B., Hayes, S. M., & Cabeza, R. (2011).
Emotion processing in the aging brain is modulated by semantic elab-
oration. Neuropsychologia, 49, 640 – 650. doi:10.1016/j.neuropsycholo-
gia.2010.09.009
Rösler, A., Ulrich, C., Billino, J., Sterzer, P., Weidauer, S., Bernhardt, T.,
. . . Kleinschmidt, A. (2005). Effects of arousing emotional scenes on the
distribution of visuospatial attention: Changes with aging and early
subcortical vascular dementia. Journal of the Neurological Sciences,
229 –230, 109 –116. doi:10.1016/j.jns.2004.11.007
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity domi-
nance, and contagion. Personality and Social Psychology Review, 5,
296 –320. doi:10.1207/S15327957PSPR0504_2
Samanez-Larkin, G. R., & Carstensen, L. L. (2011). Socioemotional func-
tioning and the aging brain. In J. Decety & J. T. Cacioppo (Eds.), The
Oxford handbook of social neuroscience (pp. 507–521). New York, NY:
Oxford University Press. doi:10.1093/oxfordhb/9780195342161.013
.0034
!
Samanez-Larkin, G. R., Robertson, E. R., Mikels, J. A., Carstensen, L. L.,
& Gotlib, I. H. (2009). Selective attention to emotion in the aging brain.
Psychology and Aging, 24, 519 –529. doi:10.1037/a0016952
!
Savaskan, E., Müller, S. E., Böhringer, A., Philippsen, C., Müller-Spahn,
F., & Schächinger, H. (2007). Age determines memory for face identity
and expression. Psychogeriatrics, 7, 49 –57. doi:10.1111/j.1479-8301
.2007.00179.x
Scheibe, S., & Carstensen, L. L. (2010). Emotional aging: Recent findings
and Future trends. The Journals of Gerontology: Series B: Psychological
Sciences and Social Sciences, 65B, 135–144. doi:10.1093/geronb/
gbp132
!
Shamaskin, A. M., Mikels, J. A., & Reed, A. E. (2010). Getting the
message across: Age differences in the positive and negative framing of
health care messages. Psychology and Aging, 25, 746 –751. doi:10.1037/
a0018431
!
Spaniol, J., Voss, A., & Grady, C. L. (2008). Aging and emotional
memory: Cognitive mechanisms underlying the positivity effect. Psy-
chology and Aging, 23, 859 – 872. doi:10.1037/a0014218
!
St. Jacques, P., Dolcos, F., & Cabeza, R. (2010). Effects of aging on
functional connectivity of the amygdala during negative evaluation: A
network analysis of fMRI data. Neurobiology of Aging, 31, 315–327.
doi:10.1016/j.neurobiolaging.2008.03.012
!
Thapar, A., & Rouder, J. N. (2009). Aging and recognition memory for
emotional words: A bias account. Psychonomic Bulletin & Review, 16,
699 –704. doi:10.3758/PBR.16.4.699
!
Thomas, R. C., & Hasher, L. (2006). The influence of emotional valence
on age differences in early processing and memory. Psychology and
Aging, 21, 821– 825. doi:10.1037/0882-7974.21.4.821
Thompson, L. A., Aidinejad, M. R., & Ponte, J. (2001). Aging and the
effects of facial and prosodic cues on emotional intensity ratings and
memory reconstructions. Journal of Nonverbal Behavior, 25, 101–125.
doi:10.1023/A:1010749711863
!
Tomaszczyk, J. C., Fernandes, M. A., & Macleod, C. M. (2008). Personal
relevance modulates the positivity bias in recall of emotional pictures in
older adults. Psychonomic Bulletin & Review, 15, 191–196. doi:10.3758/
PBR.15.1.191
!
Van Gerven, P. W. M., & Murphy, D. R. (2010). Aging and distraction by
irrelevant speech: Does emotional valence matter? The Journals of
Gerontology: Series B:Psychological Sciences and Social Sciences,
65B, 667– 670. doi:10.1093/geronb/gbq048
!
Waring, J. D., & Kensinger, E. A. (2009). Effects of emotional valence
and arousal upon memory trade-offs with aging. Psychology and Aging,
24, 412– 422. doi:10.1037/a0015526
!
Werheid, K., Gruno, M., Kathmann, N., Fischer, H., Almkvist, O., &
Winblad, B. (2010). Biased recognition of positive faces in aging and
amnestic mild cognitive impairment. Psychology and Aging, 25, 1–15.
doi:10.1037/a0018358
!
Williams, P., & Drolet, A. (2005). Age-related differences in responses to
emotional advertisements. Journal of Consumer Research, 32, 343–354.
doi:10.1086/497545
Wurm, L. H. (2011). Decreasing complexity of affective space in older
adults lower on cognitive control: Affective effects in a nonaffective task
and with nonaffective stimuli. Psychology and Aging, 26, 716 –730.
doi:10.1037/a0022513
!
Yang, L., & Hasher, L. (2011). Age differences in the automatic acces-
sibility of emotional words from semantic memory. Cognition and
Emotion, 25, 3–9. doi:10.1080/02699930903523348
!
Yang, L., & Ornstein, T. J. (2011). The effect of emotion-focused orien-
tation at retrieval on emotional memory in young and older adults.
Memory, 19, 305–313. doi:10.1080/09658211.2011.561803
Received March 3, 2013
Revision received September 4, 2013
Accepted September 9, 2013 !
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
15
POSITIVITY EFFECT META-ANALYSIS
... Consequently, older adults are more motivated to elaborate on emotionally meaningful and especially on positive information, which in turn improves their memory for such information. This emotionality effect (and positivity effect in particular) has been well established for item memory (i.e., memory for central information; see, Reed et al., 2014 for a meta-analysis). Two studies suggest that older adults also remember presumably emotional sources, but these studies did not use normed emotional stimuli and did not distinguish between positive and negative valence (May et al., 2005;Rahhal et al., 2002). ...
... For instance, research on emotional item memory has shown that younger and older adults both preferably process emotional over neutral items (e.g., pictures or words), but differ in the type of emotional material on which they focus (Carstensen, 2006). More specifically, studies have shown that younger adults typically show a negativity bias in item memory (i.e., better memory for negative items), whereas older adults seem to preferably process and memorize positive items (i.e., positivity bias) or put less emphasis on negative items (i.e., reduced negativity bias; Reed et al., 2014). This phenomenon, in both manifestations, is termed positivity effect (Mather & Carstensen, 2005) and, in short, describes older compared to younger adults' relative preference for positive over negative information. ...
... The positivity effect has been well investigated for item memory (see meta-analysis by Reed et al., 2014) but not for source memory. This is surprising given that the age-related source memory deficit might be reduced if older adults' processing preference for emotional (and especially positive) material (see SST) extends to emotional sources. ...
Article
The goal of our research was to investigate whether older adults show a source memory enhancement for emotionally valenced sources. Additionally, building on research on the socioemotional selectivity theory and the age-related positivity effect, we tested whether older adults show a larger enhancement for positive compared to negative (and neutral) sources than younger adults. In Experiment 1 (nold = 25, nyoung = 27), we used one positive, one negative, and one neutral picture to manipulate source valence (many-to-one mapping of items to sources), whereas, in Experiment 2 (nold = 62, nyoung = 62), we used multiple pictures per source valence category (one-to-one mapping of items to sources) to counteract potential habituation effects. In both experiments, sources had medium and matching arousal levels. Items were neutral words superimposed on the source pictures. To support an implicit, natural information processing, participants rated the words in terms of pleasantness. We analyzed memory data with a multinomial processing tree model to disentangle memory processes from guessing bias. Across both experiments, an age-related positivity effect occurred in participants' pleasantness ratings. This effect, however, did not carry over to older adults' source memory. That is, in source memory, we found a general emotionality effect for younger but not for older adults and no age-related positivity effect. We propose that due to older adults' pronounced difficulties in remembering the item-to-source link (i.e., associative deficit), even a greater focus on an inherently emotional source might be insufficient to boost source memory.
... Previous studies on age-related findings regarding emotional functioning describe information processing bias toward positive versus negative information among older adults. The fact that our findings were not in line with this observation could be explained by the low mean age (35.28 years) of the participants (Reed et al. 2014). Taken together, our results indicate that viewing nature videos was successful in producing a comparable experience to the theoretical heterogeneity of subjective experience that has been previously found during mindfulness meditation practices without the key ingredients of MM (meditation, acceptance, awareness). ...
Article
Full-text available
Objectives Various active control interventions for Mindfulness-Based Stress Reduction (MBSR) have been developed, though many can fall short in controlling for non-specific or placebo effects. We developed a Nature-Based Stress Reduction (NBSR) program based on previously reported positive results from virtual natural environment exposure on mental health. Methods In the present study, we present the NBSR program with its components that were matched with MBSR to ensure equality in structure, duration, contacts, and intensity, but not in specific active components (i.e., mindfulness meditation). Furthermore, we characterized the nature video component of NBSR (videos consisting of scenes of nature) as an attention-matched activity equivalent to the formal meditation practice components of MBSR. Videos were edited with creator permissions and freely online available content to include ten 3-min clips for scenes of nature from 8 different biomes. All clips were viewed by 3 different staff members and rated based on hedonic valence (pleasant to unpleasant). Each 30-min video set was designed to have a ratio of 4 pleasant, 3 unpleasant, and 3 neutral valence clips consistent with the documented heterogeneity of affective experiences during mindfulness meditation. Amazon Mechanical Turk Workers (n = 127) rated hedonic valence and self-reported arousal for individual video clips. We conducted ANOVA and t-tests to establish how hedonic valence differed by proposed valence category. Results Mean valence ratings significantly differed between the three categories of nature video clips using an ANOVA test (p < .001). Follow-up pairwise t-tests revealed significant differences between valence ratings for pleasant vs. unpleasant (p < .001), neutral vs. unpleasant (p < .001), and pleasant vs. neutral (p < .01). Conclusions The subjective experience of NBSR nature videos was reported as pleasant, with higher variability reported for unpleasant clips. This pattern generally parallels the variability and heterogeneity of subjective experiences during mindfulness meditation. These findings demonstrate that the nature video component of NBSR provides promising attention- and valence-matched placebo activity unrelated to mindfulness meditation. A comparison of NBSR versus MBSR in a randomized controlled trial is needed to validate NBSR; however, the freely available nature videos may be a useful component to match mindfulness meditation practice in studies.
... Our finding that older compared to young adults evaluated somewhat untrustworthy-looking faces as more trustworthy is in line with emerging evidence of decreased sensitivity to deceptive cues in aging (Ruffman et al., 2012;Frazier et al., 2021;see Ebner et al., 2022, for an overview). This finding also suggests that the age-related positivity effect (Reed et al., 2014) applies to face trustworthiness evaluation (see also Zebrowitz et al., 2013Zebrowitz et al., , 2017, in that older adults rated untrustworthy-looking faces, as faces depicting negative facial cues, as less negative/more positive than young adults did. ...
Article
Full-text available
The amygdala has been shown to be responsive to face trustworthiness. While older adults typically give higher face trustworthiness ratings than young adults, a direct link between amygdala response and age-related differences in face trustworthiness evaluation has not yet been confirmed. Additionally, there is a possible modulatory role of the neuropeptide oxytocin in face trustworthiness evaluation, but the results are mixed and effects unexplored in aging. To address these research gaps, young, and older adults were randomly assigned to oxytocin or placebo self-administration via a nasal spray before rating faces on trustworthiness while undergoing functional magnetic resonance imaging. There was no overall age-group difference in face trustworthiness ratings, but older compared to young participants gave higher trustworthiness ratings to ambivalently untrustworthy-looking faces. In both age groups, lower face trustworthiness ratings were associated with higher left amygdala activity. A comparable negative linear association was observed in right amygdala but only among young participants. Also, in the right amygdala, lower and higher, compared to moderate, face trustworthiness ratings were associated with greater right amygdala activity (i.e., positive quadratic (U-shaped) association) for both age groups. Neither the behavioral nor the brain effects were modulated by a single dose of intranasal oxytocin administration, however. These results suggest dampened response to faces with lower trustworthiness among older compared to young adults, supporting the notion of reduced sensitivity to cues of untrustworthiness in aging. The findings also extend evidence of an age-related positivity effect to the evaluation of face trustworthiness.
... Thus, older people pay more attention to positive compared to negative stimuli. This so-called "positivity effect" explains an empirical phenomenon, stating that information processing exhibits a positive bias in aging (Mather and Carstensen, 2005;Reed et al., 2014). ...
Article
Full-text available
Healthy aging is accompanied by multi-faceted changes. Especially within the brain, healthy aging exerts substantial impetus on core parts of cognitive and motivational networks. Rewards comprise basic needs, such as food, sleep, and social contact. Thus, a functionally intact reward system remains indispensable for elderly people to cope with everyday life and adapt to their changing environment. Research shows that reward system function is better preserved in the elderly than most cognitive functions. To investigate the compensatory mechanisms providing reward system stability in aging, we employed a well-established reward paradigm (Monetary Incentive Delay Task) in groups of young and old participants while undergoing EEG measurement. As a new approach, we applied EEG connectivity analyses to assess cortical reward-related network connectivity. At the behavioral level, our results confirm that the function of the reward system is preserved in old age. The mechanisms identified for maintaining reward system function in old age do not fit into previously described models of cognitive aging. Overall, older adults exhibit lower reward-related connectivity modulation, higher reliance on posterior and right-lateralized brain areas than younger adults, and connectivity modulation in the opposite direction than younger adults, with usually greater connectivity during non-reward compared to reward conditions. We believe that the reward system has unique compensatory mechanisms distinct from other cognitive functions, probably due to its etymologically very early origin. In summary, this study provides important new insights into cortical reward network connectivity in healthy aging.
Article
While a great deal of research has explored how different formats of online reviews can influence the review reader, less research has explored how writing in different styles of reviews can influence the review writer. This research contributes to the online review literature by exploring how informational versus story review formats influence the review writer's product or service evaluation. Although prior research has found that stories are effective at persuading readers, the current research finds evidence to the contrary for review writers. Writers of negative informational reviews reported evaluations that were less negative after writing the review compared to those who wrote story reviews. Surprisingly, negative informational reviews can make positive aspects of the product or service experience more salient, hence tempering the effects of a negative product experience and positively influencing evaluations. If retailers are interested in improving consumers' product or service evaluations after a negative experience, this research suggests that they would be better served by asking consumers to write an informational review as compared to a story review.
Chapter
The 2020 Census was the first U.S. census to use an online reporting option as the primary data collection vehicle. The online census questionnaire was available between March and October of 2020. User satisfaction with the online census experience was measured via an online follow-up survey called the 2020 Census User Experience (UX) Survey. A sample of 153,000 web respondents from the 2020 Census were invited to complete the UX survey, and notified with up to three text message invitations. We selected a representative nationwide sample of respondents likely to have had different experiences filling out the Census, such as those who answered the census in March and those who waited until October, or those who answered using a smartphone and those who used a PC. We did not take into consideration the demographics of the respondents when selecting the sample, however, post analysis showed some age-related differences in both satisfaction with the online census experience and in responding to the UX survey through a link sent via text message. Older adults who already answered their census online and who likely owned a cell phone were more likely to interact with the text message (either by replying STOP or answering the survey) than either younger users or middle-age users. Older adults were also more satisfied than younger adults. However, satisfaction did not differ between older and middle-aged adults. These findings add to the general research on older adults and technology.
Chapter
Digital media technologies have been gradually integrated into teaching activities over the past few years, providing teachers with more possibilities for teaching. This study examines the teaching effects of an interactive AI based image-processing platform in assisting as a teaching aid for children painting education. In this study, we compared the learning interest, learning attitude, and continuous learning intention of 96 children aged 5 to 13 in the process of painting education. The subjects were divided into two groups: the experimental group used AI image processing for painting education, and the control group utilized traditional teaching methods for painting learning. Results showed that the use of AI image-processing tools in painting education reduces girls’ learning attitudes and continuous learning intention, while stimulating boys’ learning interest.
Article
The power of episodic memories is that they bring a past moment into the present, providing opportunities for us to recall details of the experiences, reframe or update the memory, and use the retrieved information to guide our decisions. In these regards, negative and positive memories can be especially powerful: Life’s highs and lows are disproportionately represented in memory, and when they are retrieved, they often impact our current mood and thoughts and influence various forms of behavior. Research rooted in neuroscience and cognitive psychology has historically focused on memory for negative emotional content. Yet the study of autobiographical memories has highlighted the importance of positive emotional memories, and more recently, cognitive neuroscience methods have begun to clarify why positive memories may show powerful relations to mental wellbeing. Here, we review the models that have been proposed to explain why emotional memories are long-lasting (durable) and likely to be retrieved (accessible), describing how in overlapping—but distinctly separable—ways, positive and negative memories can be easier to retrieve, and more likely to influence behavior. We end by identifying potential implications of this literature for broader topics related to mental wellbeing, education, and workplace environments.
Preprint
Full-text available
Background: We examined age group differences in hedonic adaptation trajectories of positive and negative affect (PA/NA) at different arousal levels during the severe societal restrictions that governments implemented to contain the first wave of the COVID-19 pandemic (March to June 2020). Method: Data from 10509 participants from 33 countries and 12 weekly assessments were used (67% women, aged 18 to 85+, on average 318 participants per country (SD= 434) and 5.6 assessments (SD= 2.5) per participant). Multilevel models (level 1: assessments, level 2: participants, level 3: countries) were fit to examine trajectories of low to high arousal PA and NA during the phase of tightening societal restrictions, the phase of stable peak restrictions, and the phase of easing restrictions separately. Results: During the entire study period mean levels of PA were lower in emerging and young adults (aged 18-44) than older adults, while mean NA levels were higher. During peak societal restrictions, participants reported increasingly more PA, especially high-arousal emotions (d= 0.36 per month vs. 0.19 unaroused). NA levels decreased over time, especially high-arousal emotions (d= 0.35 vs. 0.14 p/month). These hedonic adaptation trajectories were largely similar across age groups. Nevertheless, up to 30% of the participants increased in NA and up to 6% decreased in PA, against the general trend, demonstrating substantial individual differences in emotional adaptation. Finally, heterogeneity in the effects of time on affect was larger on the individual level than the country level. Conclusion: Emotional recovery trajectories during the first lockdown in the COVID-19 pandemic were virtually similar across age groups in 33 countries, across valence and arousal levels, suggesting age advantages in emotional well-being remain restricted to mean-level differences rather than emotion dynamics.
Article
We examined the degree to which differences in face recognition rates across emotional expression conditions varied concomitantly with differences in mean objective image similarity. Effects of emotional expression on face recognition performance were measured via an old/new recognition paradigm in which stimuli at both learning and testing had happy, neutral, and angry expressions. Results showed an advantage for faces learned with neutral expressions, as well as for angry faces at testing. Performance data was compared to three quantitative image-similarity indices. Findings showed that mean human performance was strongly correlated with mean image similarity, suggesting that the former may be at least partly explained by the latter. Our findings sound a cautionary note regarding the necessity of considering low-level stimulus properties as explanations for findings that otherwise may be prematurely attributed to higher order phenomena such as attention or emotional arousal.
Article
Mood congruence effects have long been studied in younger adults, but not in older adults. Socioemotional selectivity theory (SST) suggests that mood congruence could operate differently in older adults. One hundred and nineteen younger and 78 older adults were randomly assigned to sad or neutral mood inductions, using combined Velten and music induction procedures. Results indicated that during sad mood induction both older and younger adults showed enhanced recall of sad words on delayed word list recall task and in autobiographical memory. However, only older adults displayed mood congruence effects on lexical ambiguity and lower recall of positive words in the word list task. Results provided partial support for developmental effects on mood congruence derived from SST.
Article
Although a hallmark of Alzheimer's disease (AD) is memory impairment, there is speculation that recall may be enhanced when an emotional component is associated with an event. The current study aims to assess whether patients with AD would recall emotionally laden material better than neutral stimuli. DSMIVdiagnosed AD patients with mild to moderate dementia, as well as groups of young and elderly healthy controls, participated in this study. All subjects were administered three word lists for three trials each. The words were positive, negative, or neutral in valence and matched for concreteness, emotionality, and pleasantness. As expected, the controls performed significantly better than the AD patients. Importantly, the pattern of recall for the emotions was different, such that both control groups recalled all emotions equally, whereas the AD patients recalled significantly more negative words than positive or neutral. These findings of improved immediate memory for emotional material in AD lends support to the notion that mnemonic functions are differentially affected in the disease.
Chapter
In any meta-analysis, we start with summary data from each study and use it to compute an effect size for the study. An effect size is a number that reflects the magnitude of the relationship between two variables. For example, if a study reports the mean and standard deviation for the treated and control groups, we might compute the standardized mean difference between groups. Or, if a study reports events and nonevents in two groups we might compute an odds ratio. It is these effect sizes that are then compared and combined in the meta-analysis. Consider figure 12.1, the forest plot of a fictional metaanalysis to assess the impact of an intervention. In this plot, each study is represented by a square, bounded on either side by a confidence interval. The location of each square on the horizontal axis represents the effect size for that study. The confidence interval represents the precision with which the effect size has been estimated, and the size of each square is proportional to the weight that will be assigned to the study when computing the combined effect. This figure also serves as the outline for this chapter, in which I discuss what these items mean and how they are computed. This chapter addresses effect sizes for continuous outcomes such as means and correlations (for effect sizes for binary outcomes, see chapter 13, this volume).
Article
Divergent trajectories characterize the aging mind: Processing capacity declines, while judgment, knowledge, and emotion regulation are relatively spared. We maintain that these different developmental trajectories have implications for emotion–cognition interactions. Following an overview of our theoretical position, we review empirical studies indicating that (a) older adults evidence superior cognitive performance for emotional relative to non-emotional information, (b) age differences are most evident when the emotional content is positively as opposed to negatively valenced, and (c) differences can be accounted for by changes in motivation posited in socioemotional selectivity theory.
Article
This study used a modification of an attentional cueing task with emotional faces as cues to examine whether emotional cues influence the efficiency of alerting and spatial orienting, and whether the effects vary with the age of the faces and/or the age of the subjects. In this task, younger and older adults responded to the location of a target (left or right) preceded by a brief emotional cue (happy, sad, or neutral faces) or by no cue. Older adults showed a larger alerting effect than younger adults and this pattern was not further moderated by the valence of cues or the age of the faces. However, the results for the orienting effects indicated both age similarities and differences as a function of age of the faces. Both age groups exhibited orienting benefits from valid cueing by neutral and positive own-age faces, and showed orienting benefits for negative other-age face cues. Older adults appeared to be differentially influenced by orienting cues compared to younger adults as suggested by the cue-validity effect (i.e., response times slower for invalidly cued targets than for validly cued targets), especially for the own-age face cues. Whereas younger adults demonstrated the cue-validity effects for neutral and happy own-age face cues, older adults showed the cue-validity effects for the own-age face cues regardless of valence. The results highlight the importance of considering the age of the faces when assessing age differences in attention to emotional face cues.