Content uploaded by Judith G Chipperfield
Author content
All content in this area was uploaded by Judith G Chipperfield on Sep 22, 2018
Content may be subject to copyright.
INT’L. J. AGING AND HUMAN DEVELOPMENT, Vol. 68(4) 261-288, 2009
REGRET IN LATER LIFE: EXPLORING
RELATIONSHIPS BETWEEN REGRET FREQUENCY,
SECONDARY INTERPRETIVE CONTROL BELIEFS,
AND HEALTH IN OLDER INDIVIDUALS*
NANCY E. NEWALL
JUDITH G. CHIPPERFIELD
University of Manitoba, Winnipeg
LIA M. DANIELS
University of Alberta, Edmonton
STEVEN HLADKYJ
RAYMOND P. PERRY
University of Manitoba, Winnipeg
ABSTRACT
The present study examined what older people regret, and the relationships
between regret, health and life satisfaction. The study also explored the role
of secondary interpretive control beliefs in relation to regret. Participants
(N= 228; 79-98 years old) were asked to report on the content and frequency
of their regret, secondary interpretive control beliefs (e.g., beliefs in finding
the “silver lining” in a dark cloud), health, and life satisfaction. A content
analysis revealed that participants most commonly reported feeling regret due
*This research was supported by a Canadian Institutes of Health Research (CIHR) Canada
Graduate Scholarships Doctoral Award to the first author, and a CIHR operating grant to the second
author. This project was completed by the first author as partial fulfillment for the requirements of
a Masters degree. Portions of this work were presented in a poster at the Annual Scientific and
Educational Meeting of the Canadian Association on Gerontology, Halifax, Nova Scotia, Canada,
October 2005.
261
Ó2009, Baywood Publishing Co., Inc.
doi: 10.2190/AG.68.4.a
http://baywood.com
to things they had not done, the death of a loved one, and their own or
others’ health problems. Regression analyses indicated that experiencing
regret more frequently was associated with poorer health and life satisfac-
tion. Moreover, evidence for an emotion-modifying role of secondary inter-
pretive control beliefs was shown through its negative association with
regret. Results suggest that older adults may be experiencing age-related
regrets that differ in content from those experienced at younger ages and
that certain control beliefs may serve to lessen regret.
Said the stranger; “For life in general there is but one decree. Youth is a
blunder; manhood a struggle; old age a regret.”
Coningsby (Disraeli, 1911, p. 98)
Old age is often portrayed as a time of regret and despair. Why this is so has no
doubt to do with the many losses (e.g., physical, social) associated with later life.
But how true is the belief that old age is “a regret”? How often do older individuals
experience regret and what do they regret? Do some people experience this
emotion more often than others, and if so, why? What is particularly compelling
about studying the experience of regret in older individuals is that learning about
people’s regrets can ultimately reveal a little about their life story and about
what they value. Indeed, knowing the regrets of an older cohort can serve as a
lesson for younger generations. In addition, studying regrets in older individuals
may be important in understanding how people deal with personal loss in general,
and how emotional experiences can affect physical and psychological well-being.
The purpose of the present study was to extend the small body of research on
regret in older individuals by examining how frequently individuals experience
regret, what they regret, and how regret relates to health and life satisfaction.
The study also focused on a process that theoretically could be involved in modi-
fying emotions such as regret: the process of secondary interpretive control. The
secondary control process, characterized by Rothbaum, Weisz, and Snyder (1982)
as involving efforts to sustain perceptions of control by accommodating to the
environment, has been found in past research to be important for older individuals
in terms of their physical health and well-being (e.g., Chipperfield, Perry, &
Menec, 1999). This study allowed an examination of the association between
secondary control and regret in older individuals.
REGRET
Few individuals will go through life without experiencing regret. This single
emotion is ubiquitous. Its poignancy appears in songs and novels; and it is a
consequence of everyday decision-making (“I wish I had bought the sweater
when it was on sale”) and important life decisions and events (“I regret not
following my doctor’s advice”). In a study of the most commonly mentioned
262 / NEWALL ET AL.
emotions in conversations, regret, together with guilt and sorrow, were the next
most frequently named emotions after love (Shimanoff, 1984).
According to Webster’s Third New International Dictionary of the English
Language, the word regret is of Scandinavian origin, stemming from grata, to
weep. A complex emotion, it is typically associated with personal losses or
mistakes. In her theoretical and conceptual consideration of regret, Landman
(1987) formulated the following broad definition of regret:
Regret is a more or less painful cognitive/affective state of feeling sorry
for losses, transgressions, shortcomings, or mistakes. The regretted matters
may have been sins of commission as well as sins of omission; they may
range from the entirely voluntary to the accidental; they may have been
actually executed deeds or entirely mental ones; they may have been com-
mitted by oneself or by another person or group; they may be moral or
legal transgressions or morally and legally neutral; and the regretted matters
may have occurred in the past, the present, or the future (p. 153).
Regret can be characterized as a self-focused negative emotion sharing simi-
larities with guilt, shame, and remorse (Landman, 1987; Mandel, 2003). Regret
can also be considered a cognitive phenomenon related to counterfactual think-
ing, a process whereby individuals focus on how events may have been different
or counter to “fact.” Kahneman and Tversky (1982) used the term “counterfactual
emotion” to describe emotions such as regret, frustration, indignation, grief, and
envy that appear to depend on “a comparison of reality with what might or should
have been” (p. 206).
Much of the research on regret has focused on two types of events that are
thought to influence counterfactual thinking and ultimately lead to regret—actions
and inactions (also called commissions and omissions, respectively). A consistent
trend, termed the “action effect,” is that people report in scenario-based studies
that they would feel more regret over actions than inactions (e.g., Landman, 1988);
however, there is debate over this issue (Gilovich & Medvec, 1995; Roese, 1997;
Zeelenberg, van den Bos, van Dijk, & Pieters, 2002). For example, in contrast to
scenario-based studies, when people are asked to list their life regrets they tend
to report more regrettable inactions than actions (Gilovich & Medvec, 1995).
Implied in the studies on regrettable actions vs. inactions is the notion that a
defining feature of a regrettable event is its association with decision-making
(to act or not) and a sense of personal responsibility (e.g., Zeelenberg, van Dijk,
van der Pligt, Manstead, Empelen, & Reinderman, 1998). Although there has
been some disagreement over this issue (see Ordonez & Connolly, 2000;
Zeelenberg, van Dijk, & Manstead, 2000), research appears to show that regret
typically involves a sense of personal responsibility over a decision or outcome
(e.g., Gilovich & Medvec, 1995). Interestingly, however, when discussing the
theoretical distinction between regret and guilt, Landman (1987) concluded that
regret is the more overarching of the two emotions because people can feel regret
REGRET IN LATER LIFE / 263
for events that they had no control over, whereas guilt is typically experienced
after a transgression for which an individual feels responsible. Cases in which
individuals experience regret without having a sense of responsibility may be
“exceptions rather than the rule” (Zeelenberg et al., 2000, p. 152); however, it
appears that a sense of responsibility need not be a necessary condition for regret.
What Individuals Regret
What events do individuals regret? To answer this question, many researchers
have classified people’s regrets on the basis of the life domain to which they relate.
In a meta-analysis of nine studies performed between 1989 and 2003 which
categorized regrets based on life domains, Roese and Summerville (2005) found
that people (at least Americans) most frequently report regrets related to (in
order of most to least frequently): education, career, romance, parenting, self,
leisure, finance, family, health, friends, spirituality, and community. To explain
this ranking, the authors offered an opportunity principle such that people will
be more apt to have regrets in those life domains in which they have greater
continued opportunities (e.g., education). That is, they argued that “opportunity
breeds regret” (p. 1274).
However, Roese and Summerville (2005) noted several limitations of past
research. First, they noted that all of the nine studies in the meta-analysis involved
convenience samples, which limits generalizability of results. Second, the authors
noted that five of the nine studies used a less than ideal methodology of generating
regrets in that participants were given a pre-existing list of life domains rather
than generating them on their own. Finally, we would add that a third major
limitation of past research is that often researchers have restricted their definition
of regrets to inactions (Roese & Summerville, 2005), with few studies simply
asking participants about their feelings of regret (Jokisaari, 2003). For example,
some studies have focused only on regrettable inactions (e.g., Lecci, Okun, &
Karoly, 1994) or have asked participants what they would do differently if it were
possible to relive their lives (e.g., DeGenova, 1992; Kinnier & Metha, 1989),
perhaps predisposing respondents to think of inactions (Gilovich & Medvec,
1995). Given the earlier explanation that regret can be either due to action or
inaction, this represents another lapse in past research.
Functionality of Regret
Despite regret being an unpleasant feeling, like all emotions it can be inform-
ative and even functional. Thus, there can be positive consequences of regret.
The functionality of regret can perhaps best be illustrated through its connection
with counterfactual thinking, in particular “upwards” counterfactual thinking
(Markman, Gavanski, Sherman, & McMullen, 1993) which involves thinking
about how events could have turned out better (e.g., “If only I had followed
her advice I wouldn’t have gotten ill . . .”) versus “downwards” counterfactual
264 / NEWALL ET AL.
thinking whereby individuals think of worse scenarios than what actually occurred
(e.g., “At least I did not break both of my arms when I fell . . .”).
Roese (1997) argued that through a contrast effect, thinking about how events
could have been better (i.e., upwards counterfactual thinking) could result in
negative affect like regret. He also argued that through a causal inference effect,
upwards counterfactual thinking can provide information for why events occurred
(e.g., I got ill because I did not follow certain advice); and the particular causes
attributed to events can, in turn, have emotional and motivational consequences
(Weiner, 1985). In sum, (upwards) counterfactual thinking can amplify negative
emotions like regret through contrast and causal inference effects; however,
counterfactual thinking can also be functional in that it can help individuals
identify what could cause and what could potentially prevent particular
negative events in the future (Mandel, 2003; Roese, 1997). In this way, at least
in cases in which regret is felt in the context of goal attainment, the negative
thinking surrounding regret may lead to positive responding (Stewart &
Vandewater, 1999).
Negative Implications of Regret:
Relationship with Physical Health and Well-Being
Although regret can be seen to be functional or motivating, experiencing this
emotion may also have negative implications. For example, there are several
pathways by which positive and negative emotions may be linked to health and
well-being. Emotions may affect health through physiological mechanisms, for
example, or indirectly through health behaviors and cognitions (Mayne, 2001).
Emotional responses can also have an impact on social support and relationships
which, in turn, can influence health (Tucker & Friedman, 1996).
A small but growing body of research suggests that regret may be negatively
associated with individuals’ physical and psychological well-being, particularly
among older adults (e.g., Jokisaari, 2003; Lecci et al., 1994; Torges, Stewart,
& Miner-Rubino, 2005; Wrosch, Bauer, Miller, & Lupien, 2007). Lecci et al.
(1994) found that the number of reported regrets was negatively related to
life satisfaction; however, this was true only for the oldest individuals in the
sample (ages 33-59 years). From these results, the authors suggested that having
more regrets may be particularly profound or detrimental later in life because
older adults may have fewer opportunities and less time to overcome their
regrets. This idea was supported by Jokisaari’s (2003) study in which older
adults perceived being less likely to fulfill their unattained goal or change
their regret. Torges et al. (2005) found, in their study involving 60-year-olds,
that those who reported a “missed opportunity” or a “lifestyle pattern” life
regret had poorer physical health and life satisfaction that those who did not.
Moreover, in a study on life regrets among older adults (ages 63-94 years),
Wrosch et al. (2007) found that participants’ regret intensity was associated with
REGRET IN LATER LIFE / 265
poorer physical health, as indicated by presence of physical problems and
cortisol secretion levels.
In sum, studies indicate that regret may have negative consequences for health
and well-being, particularly for older individuals. This leads to the question
of: How can individuals prevent or reduce the detrimental effects of regret on
outcomes such as health and life satisfaction?
Managing Regret
As noted by Gross, Carstensen, Tsai, Skorpen, and Hsu (1997), people do not
simply let emotions come and go, they actively regulate their emotions by making
“attempts to influence which emotions they have, when they have them, and
how these emotions are experienced or expressed” (p. 597). Emotion regulation
processes are thought to include conscious and even unconscious mechanisms
that increase or decrease both positive and negative emotions (Gross, 1998).
In the regret literature, researchers have theorized about several broad mech-
anisms that could lessen or dissipate the feeling of regret. Landman (1987)
suggested that individuals could attempt to “undo” regrets either through a
physical or mental act. Similarly, Gilovich and Medvec (1995) cited two major
ways that individuals could reduce the pain of regrettable situations. First, they
described behavioral repair as overcoming or compensating for the regrettable
situation through action. For example, if a person were to regret saying some-
thing rude to someone, to overcome the regret, s/he might apologize. Second,
they described psychological repair as involving several potential cognitive
mechanisms that could reduce regret such as identifying “silver linings” (i.e.,
seeing the positive in the negative situation). For example, a person who regrets
moving to a smaller apartment might reduce regret by thinking about a positive
aspect of the situation such as the fact that they have less cleaning to do.
Torges et al. (2005) discussed the possibility that older people may not feel they
have the time and opportunity to make external changes (i.e., behavioral repair)
that would address their regrets. In this regard, in their study of 60-year-olds, they
focused on some of the internal processes (i.e., psychological repair) that people
may use to come to terms with “missed opportunities” and “lifestyle change”
regrets. They categorized responses as signifying that people had not come to
terms with regret, had put the best face on things,orhad come to terms with their
past regrets. Included in their assessment of “putting the best face on things” was
the idea that people had been able to find a positive outcome related to their regret
(e.g., regretting marrying young but being happy about having children). The
results of their study showed that those who acknowledged having any type of
regret had a lower level of well-being. In addition, however, they found that those
who had either come to terms or put the best face on regrets had better well-being
than those who had not come to terms with their regrets. This suggests that
people’s efforts to manage their regrets can have implications for well-being.
266 / NEWALL ET AL.
Wrosch and colleagues (Wrosch et al., 2005; Wrosch, Bauer, & Scheier, 2007;
Wrosch & Heckhausen, 2002) have studied self-regulation factors that appear
to be important in managing life regrets. In one study, Wrosch et al. (2005) found
that older individuals who were more disengaged from undoing their regret as
well as those individuals who had a greater number of available future goals had
less intense regret. In an innovative writing intervention study, Wrosch et al.
(2007) focused on social comparison processes (e.g., perceiving one’s own
situation as the same or better than another’s situation); minimizing blame; and
having future goals available. Results showed that people in an experimental
condition, who were given writing assignments designed to engage people in
these processes, had less “despair” associated with their regret over time compared
to people in the control condition.
Secondary Control and Regret
The present study, although taking a slightly different approach, complements
these past studies by testing the hypothesis that older adults who have a greater
general tendency to identify “silver linings” or see the positive in a negative
situation will experience regret less frequently. Thus, this study focuses on a
potential individual difference variable that could help people manage regrets.
The framework we used was Rothbaum, Weisz, and Snyder’s (1982) concept of
secondary interpretive control, although related concepts exist such as perceiving
or finding benefit and positive reappraisal (Affleck, Tennen, Croog, & Levine,
1987; Folkman & Moskowitz, 2000; Sears, Stanton, & Dannoff-Burg, 2003;
Wrosch, Heckhausen, & Lachman, 2000). As conceptualized by Rothbaum et al.
(1982), secondary interpretive control is thought to be gained through attempts
to derive meaning from negative outcomes so as to accept them. One way that
individuals may derive positive meaning is to look for what they have gained
from a negative situation. For example, a person may see his/her illness in a
positive light because it resulted in becoming closer to family members. In
general, research has found this tendency to focus on the positive to be beneficial
for people in terms of health and well-being (e.g., Affleck et al., 1987; Davis,
Nolen-Hoeksema, & Larson, 1998; Folkman & Moskowitz, 2000).
Interestingly, this concept can be likened to Gilovich and Medvec’s (1995)
psychological repair strategy of finding a “silver lining” or Torges et al.’s (2005)
notion of “putting the best face on things,” and thus it is possible that through
gaining interpretive control, this may function to help manage regret.
THE PRESENT STUDY: REGRET IN LATER LIFE
The present study provides a descriptive profile of regret in later life by
determining what events cause older individuals to feel regret. Unlike most
studies which have focused on only certain types of regret (e.g., regrets due to
inactions; life regrets), participants were simply asked about their regrets in
REGRET IN LATER LIFE / 267
general. In this way, the study allowed a comprehensive assessment of the content
of regrets of older individuals. Also, unlike past convenience-sample studies,
the present project drew on participants from a representative sample of older
adults, which lends confidence to the generalizability of study results.
The present study also focuses on frequency of experiencing regret among
older adults. This is in contrast to past studies focusing on intensity or number
of regrets. In this way, the present study builds on a study by Chipperfield,
Perry, and Weiner (2003), which examined a variety of discrete positive and
negative emotions among older adults (ages 72+ years), including regret. Zero-
order correlational analyses suggested an association between recent (previous
two days) frequency of experiencing regret and health and life satisfaction. The
present study builds on this study by investigating, in an even older sample
(79+ years), the relationships between regret and health and life satisfaction
more closely. In particular, this study examined the relationships in the context
of sociodemographic variables such as age and this study used a prospective
design, controlling for past life satisfaction and health. It was hypothesized that
experiencing regret more frequently would be associated with poorer health and
life satisfaction, even controlling for past measures of health and life satisfaction.
In addition, as age appears to relate to number of regrets, with older individuals
reporting a fewer number of regrets than younger individuals (Jeffries & Konnert,
2002; Lecci et al., 1994), it is possible that age may relate to frequency of regrets.
Given the small age range of the present study, however, it was difficult to predict
whether any age effects would be found.
Finally, this study examines the relationship between secondary control beliefs
and regret. It was predicted that interpretive secondary control beliefs would
serve as a “repair mechanism” so that those individuals having stronger secon-
dary control beliefs would report experiencing regret less frequently. That is, if
general beliefs translate into a general tendency to use certain strategies (e.g.,
actively looking for the positive in negative situations), then these beliefs may
be associated with less regret and better well-being.
METHOD
The present study included 228 participants who were interviewed in their
homes as part of the Successful Aging Study 2003 (SAS 2003), a follow-up to
SAS 1996 and linked to the Aging in Manitoba (AIM) longitudinal studies. Prior
to describing the SAS 2003 methodology and sample characteristics, a brief
overview of the AIM studies is provided.
Aging in Manitoba (AIM) Study
The Aging in Manitoba (AIM) population-based longitudinal studies have
been ongoing for over 30 years. The initial 1971 survey (N= 4803) was conducted
by the Manitoba Provincial Department of Health and Social Development in
268 / NEWALL ET AL.
order to identify needs of older persons (for more details see Mossey, Havens,
Roos, & Shapiro, 1981). A random sample of Manitobans (65+ years old) was
drawn from the Manitoba Health population registry, stratified by place of
residence (i.e., community or personal care home) and region. Using similar
techniques, two additional cross-sectional samples of older individuals (60+
years old) were selected in 1976 (N= 1302) and again in 1983 (N= 2877). Since
then, there have been several follow-up data collection waves in which socio-
demographic, health, and psychosocial information has been obtained. Specif-
ically, follow-up data were obtained from the survivors of the two initial cross-
sections in 1983 and from survivors of all three cross-sectional samples in 1990,
1996, and 2001, 2005, and 2006.
It was important to the objectives of the AIM study that the study samples
be representative of the broader population of older Manitobans. Thus, it is
important to note that the initial 1971 sample was found to be representative in
terms of age and gender of the Manitoba population aged 65 and over in 1971
(Mossey et al., 1981). Furthermore, the AIM 1990 sample was found to be
comparable to the older Manitoba population in 1990 in terms of gender and
marital status and was generally representative in terms of age (Chipperfield,
Havens, & Doig, 1997).
Successful Aging Study 2003
In 2003, individuals who had participated in the 2001 Aging in Manitoba
follow-up study and who met specific eligibility criteria were interviewed as
part of the Successful Aging Study 2003 (SAS 2003). As described in more
detail elsewhere (Newall, Chipperfield, Blandford, Perry, & Havens, 2004) those
eligible for SAS 2003 met the following criteria:
1. were living in the community (i.e., not institutionalized) in Winnipeg,
Brandon, or Selkirk, in the province of Manitoba;
2. had indicated English as the language of choice for the AIM 2001 interview;
3. had received no or only some assistance from a proxy for the AIM 2001
interview;
4. had fully satisfactory, adequate, or fairly satisfactory comprehension, as
rated by the AIM 2001 interviewer; and
5. had indicated they would be willing to participate in future studies at the
time of the AIM 2001 interview.
In total, 232 individuals completed SAS 2003 interviews. However, for the
purposes of the present study, the sample included only the 228 individuals who
had complete data for the questions relating to regret. The interview schedule
took on average 1.5 hours to administer and contained questions relevant to the
present study on demographics, beliefs, emotions, health, and life satisfaction.
REGRET IN LATER LIFE / 269
It should be noted that in comparison to the larger AIM 2001 study sample,
the SAS 2003 sample was found to be similar in terms of gender and self-rated
health; and yet they were younger, more physically capable, and better educated
(Newall et al., 2004). The differences between samples are not surprising given
the SAS study focused on a population of individuals residing in urban com-
munities (not institutionalized) who were cognitively and physically capable of
completing interviews with little or no assistance from others. Thus, findings can
be generalized to cognitively competent older adults living in urban communities.
Measures
Measures from the present study were obtained from SAS 2003, or, where
indicated, from AIM 2001. Table 1 provides details of the study measures.
Sociodemographics
Information on age, gender, marital status, and socioeconomic status (i.e.,
education level and annual income) was obtained from the SAS 2003 interview,
and, where appropriate, from AIM 2001. In particular, information on birth dates
and on gender (1 = male,2=female) was obtained from the AIM 2001 interview
(Table 1). Details on current marital status (married, n= 86; widowed, n= 120;
single, n= 15; or divorced/separated, n= 7) were obtained in SAS 2003. For the
purposes of the present study, a dichotomous marital status variable was created
that distinguished between participants who were married or not (1= not married,
2=married). Education level was assessed in AIM 2001 by asking participants
to indicate the number of years or grades completed in school. One missing
case was replaced with the mean. Participants were also asked in AIM 2001 to
indicate an income range that best corresponded to their annual income before
deductions (categories ranged from 1 = <$5,000 to 8 = >$39,999). Using a method
outlined by Tabachnick and Fidell (2001), 18 missing values were replaced
using predicted values from a regression equation in which income was regressed
onto education level, marital status, gender and age.
Regrets
To assess frequency of regret, participants were asked to think about their
feelings of regret over the past couple of years and to indicate on a 7-point
response scale how often they felt regret (0 = never,3=sometimes,6=almost
always). The raw frequency scores were: 78 participants gave a score of “0”
or never; 2 participants a score of “1”; 8 participants a score of “2”; 97 partici-
pants a score of “3” or sometimes; 27 participants a score of “4,” 10 participants
a score of “5,” and 6 participants a score of “6” or almost always. Because of
the small number of people who reported certain values (e.g., only 10 people
reported responses of “1” or “2”), responses for this variable were subsequently
270 / NEWALL ET AL.
re-coded into a 3-point scale ranging from “never” to “often.” The cut points
used were: 1 = never (i.e., raw scores of zero, n= 78); 2 = sometimes (i.e., raw
scores ranging from 1-3, n= 107), and 3 = often (i.e., raw scores of 4-6, n= 43).
Next, to determine the content of causes of regret, those individuals who
indicated that they had felt regret (n= 150) were then asked, “Can you tell
me what has caused you to feel regret?” Interviewers recorded up to two
causes of regret for each participant. Note that based on this information, we
were able to create a number of reported regrets variable which ranged from
zero to two reported regrets (Table 1). This variable was used in subsequent
supplementary analyses.
Secondary Interpretive Control Beliefs
Secondary interpretive control beliefs were measured using five items designed
to assess a tendency to see the positive in negative experiences (“I’m a believer
in the idea that ‘every cloud has a silver lining’”; “Negative experiences can
often be a ‘blessing in disguise’”; I often tell myself I should ‘count my
blessings’”; “Things will work out in the end”; “patience is a virtue”). Also
termed reinterpretive folk beliefs (Swift, Bailis, Chipperfield, Ruthig, & Newall,
2008), participants were asked about their agreement with the statements
(1 = strongly disagree;to6=strongly agree).
A principle components analysis of the five items was performed using
responses from the 228 individuals and pair-wise deletion of missing values. As
expected, one component emerged, accounting for 52% of the total variance
(eigenvalue = 2.62; all five items loading above .68). A Cronbach’s alpha value
of .76 indicated acceptable reliability for the scale. Thus, all five items were
used to create a composite mean score for each participant in which higher
scores reflected stronger interpretive control beliefs. Swift et al. (2008) showed
evidence for the distinctness of this construct from both positive and negative
affect. As further evidence for the validity of this secondary interpretive control
beliefs scale, this measure showed expected moderately strong correlations with
perceived control over life in general (r= .35, p< .01), optimism (Scheier
& Carver, 1985) (r= .38, p< .01), and positive reappraisal beliefs (Wrosch et al.,
2000) (r= .49, p< .01).1
Health
A measure of total number of health conditions was obtained by asking
individuals to report whether in the past year they had experienced any of 21
REGRET IN LATER LIFE / 271
1We were able to perform this correlation between interpretive control beliefs and posi-
tive reappraisal in a sample of older prticipants (N= 261) who took part in a later 2006
data collection (see Newall, Chipperfield, Swift, Haynes, & Chuchmach, 2008) in which
both constructs were assessed.
Table 1. Description of Study Measures
Measures Anchors
No. of
items MSDSkewness Kurtosis Range
Age (yrs)
Gender
Marital status
Education (yrs)
Annual income
Frequency of regret
Number of reported regrets
—
1 = men (38%)
2 = women (62%)
1 = not married (62%)
2 = married (38%)
—
1 = <5k; 8 = >40k
1 = never
3 = often
0 = no reported regrets
2 = two reported regrets
1
1
1
1
1
1
1
84.99
—
—
10.47
4.61
1.85
1.00
4.29
—
—
2.63
1.91
.71
.83
.74
—
—
.44
.30
.23
–.00
–.15
—
—
1.69
–.99
–1.01
–1.56
79-98
—
—
3-21
1-8
1-3
0-2
272 / NEWALL ET AL.
Interpretive secondary
control (mean)a
Health conditions (sum)b
Life satisfaction (mean)a
2001 Health conditions
(sum)b
2001 Life satisfaction
2001 Depressiona
1 = strongly disagree
6 = strongly agree
0=no
1 = yes
0 = disagree
1 = agree
0=no
1 = yes
1 = fair
3 = excellent
0 = rarely
4 = most or all of the time
5
21
20
21
1
10
4.64
5.26
.71
4.41
2.05
.57
.80
2.77
.20
2.65
.57
.44
–.29
.49
–.59
.73
.01
.88
–.14
–.11
–.17
.08
.11
.80
1.8-6
0-14
.12-1
0-12
1-3
0-2.1
aMean scores; bsum scores.
REGRET IN LATER LIFE / 273
specific health problems (e.g., diabetes, cancer, incontinence) (0 = no, 1 = yes).
A composite measure was created by summing the number of health conditions
(possible range 0-21). Although not taking into account the severity or seriousness
of each health problem, a count of the number of health problems is a useful
index of the burden of illness (Morbidity and Mortality Weekly Report, 1989).
Life Satisfaction
Older individuals’ life satisfaction was assessed using the Life Satisfaction
Index A (Neugarten, Havighurst, & Tobin, 1961). This scale contained 20 items
(e.g., I am just as happy as when I was younger) and response options ranged
from 0 (disagree)to1(agree). After reverse-coding the appropriate negative
statements, mean composite scores were calculated for each participant (possible
range 0-1). The Cronbach’s alpha for the measure was .73.
Past Health, Life Satisfaction, and Depression (2001)
It was possible to obtain information on participants’ past health and life
satisfaction as assessed approximately 2 years earlier in AIM 2001 (Table 1).
As will be seen in the results section, this permitted an assessment of whether
certain variables predicted outcomes even after controlling for prior levels of
health and life satisfaction. Specifically, a measure of individuals’ 2001 health
conditions was created in the same way as the previously described SAS 2003
measure. An assessment of individuals’ 2001 general life satisfaction was
obtained by asking individuals, “How would you describe your satisfaction with
life in general at present?” (1 = excellent to 4 = poor). As there were only two
individuals who answered “poor,” these were combined with “fair” responses.
Responses were also reverse-coded so that higher scores reflected more satis-
faction (possible range 1-3).
In addition, in certain supplemental analyses we included a measure of past
depression into the models. Participants were assessed on depression in 2001
using 10 items from the Center for Epidemiology Studies-Depression scale
(Radloff, 1977; Radloff & Teri, 1986) (e.g., how often in the past week a
respondent “felt depressed”; “could not get going”) (0 = rarely;4=most or all of
the time). Specifically, a mean score was calculated for each participant after
reverse-coding the appropriate items such that higher scores indicated greater
depression (Cronbach’s alpha = .74).
RESULTS
Descriptive Profile of Regret in Older Individuals
Regret Content Analysis
To address the first research question pertaining to what older individuals
regret, a content analysis as described below was used to classify the reported
274 / NEWALL ET AL.
causes of regrets. Interviewers recorded up to two regrets per participant. Out of
the 228 participants, 150 reported experiencing regret in the past couple of years.
Of these, 140 individuals indicated a specific cause of regret (seven individuals
could not identify a specific cause and three individuals refused to provide one).
An additional 75 individuals specified a second cause of regret (and four refused to
provide the second specific cause). Thus, a total of 215 regrets were reported.
Regrets reported first (N= 140) were categorized separately from regrets
reported second (N= 75) to allow for the possibility that they might differ in
certain ways. Focusing on first reported regrets, two raters independently sorted
the regrets into groups on the basis of similarities in content. Although it was
possible for raters to come up with very different categories given this method-
ology, on the first categorization attempt, raters identified 11 similar categories.
Moreover, they classified 67% (94/140) of regrets into the same categories.
After discussing the cases in which there was disagreement, raters created six
more categories and were able to agree on how to categorize the remaining
responses. Thus, first regrets were classified into 18 categories, including the
“other” category.
The same method was used to classify the regrets reported second. On the first
categorization attempt, when the two raters independently grouped the regrets,
they agreed on the categorization of 60% (45/75) of the regrets. After discussing
those cases in which there was disagreement, the raters were able to agree on how
to categorize the remaining responses. Because the identification of categories for
second regrets did not differ from the first regrets, it was decided to combine the
regrets. As a final step, to help ensure reliability of ratings, a third rater examined
the regrets and categorized them into the 18 previously-established categories.
Interrater-reliability was good, kappa = .80, p< .001, increasing confidence in the
reliability of established categories.
Ultimately, as shown in Table 2, all of the regret categories relating to “things
not done” were subsumed into one category. Similarly, all regrets relating to
death (e.g., spouse or family/friends) and health (self or other) were subsumed
into these two categories. This resulted in a final total of 12 categories plus one
“other” category. Table 2 summarizes all 215 regrets, with categories listed in
order of most to least commonly mentioned. By far the most commonly men-
tioned regrets resulted from things not done (n= 42) and death (n= 40). The
next most commonly reported regrets resulted from health (n= 27), lack of
connection with others (n= 19) and being limited (n= 18). Although retained as
separate categories, many regrets were the result of interpersonal problems,
e.g., “hurting others” (n= 12), “family conflicts” (n= 8), and “other’s behaviors”
(n= 7). This interpersonal theme is congruent with other research findings on
the causes of discrete emotions (Chipperfield, Perry, Weiner, & Newall, in press;
Magai, Consedine, Krivoshekova, Kudadjie, & McPherson, 2006). And the inter-
personal theme as a cause of regret is further implied in the categories of “lack
of connection with others” and the two subcategories within “things not done”
REGRET IN LATER LIFE / 275
276 / NEWALL ET AL.
Table 2. Summary of Content of Regrets
Regrets
Regret categories Examples No. %
Things not done:
General
Being better to others
Doing things for others
who are now dead
Getting better education
Death-related:
Spouse
Family/friend
Health-related:
Self
Others
Lack of connection with
others
Being limited
Moving/relocating
Hurting others
Family conflicts
Other’s behaviors
Unmet expectations
Getting older
Finance-related
Other
“Something I should have done” (14)
“Regretted not helping someone” (14)
“Did not visit [friend] before she died” (6)
“Regret not getting better education” (8)
“That my husband passed away” (15)
“Loss of my brother and sister” (25)
“My heart attack” (10)
“Husband’s illness” (17)
“My children don’t live here”
“Living by myself”
“Can’t do the physical things I used to do”
“Regret I can’t go to work”
“Regret moving to the city”
“Regret we went back to the country”
“I yelled at people”
“A little misunderstanding with family”
“Regret the way my son is running farm”
“When things don’t go the right way”
“Things not going my way”
“I regret getting old and I can’t do
anything about it”
“Family estate not going as it should”
“Indecision in general”
42
40
27
19
18
14
12
8
7
5
6
4
13
19.5
18.6
12.6
8.8
8.4
6.5
5.6
3.7
3.3
2.8
2.3
1.9
6
Total 215 100
(being better to others/doing things for others). Also noteworthy, at least four
of the categories reflected possible age-related regrets such as “getting older”
(n= 6), “being limited” (n= 18), “death” (n= 40), and “health” (n= 22), together
representing a large proportion of the total regrets (86/215 = 40%).
Frequency of Regret
A second research question related to the frequency that older individuals
experience regret. A mean of 1.85 (Table 1) suggests that, on average, individuals
experience regret fairly often. On the other hand, over one-third (78/228 = 34%)
of participants answered “never” when asked how frequently in the past couple
of years they had experienced regret. This finding is similar to Torges et al.’s
(2005) study results in which 48% of the sample of 60-year-olds reported
having no “missed opportunities” or lifestyle change regrets.
Sociodemographics and Regret
Pearson (for continuous variables) and Spearman (for dichotomous variables)
correlations were used to examine whether certain sociodemographic variables
related to how frequently individuals experience regret. These correlations are
embedded within Table 3 which also shows the correlations between all study
variables. Interestingly, none of the sociodemographic variables significantly
correlated with frequency of regret. This suggests that the frequency with which
older individuals experience regret is similar for men and women, for those
married and not married, and so forth.
Regret Predicting Health and Life Satisfaction
It was hypothesized that frequency of regrets would independently predict
the outcomes of health and life satisfaction even after controlling for the back-
ground sociodemographic and past health/life satisfaction variables. Addressing
this question involved performing separate multiple regression analyses in which
frequency of regret and the background variables were included as predictors
of each of the outcome variables. Table 4 shows the results of these regression
analyses.
Focusing first on life satisfaction, as expected, even after controlling for socio-
demographics and past life satisfaction and health, frequency of regret predicted
current life satisfaction (B= –.23, p< .01). As hypothesized, individuals who
reported experiencing regret more frequently were less satisfied with their
life. Note that with the addition of frequency of regret, the adjusted R2value
significantly increased from R2= .08 to R2= .13 (F= 13.36, p< .05). Indeed regret
was a stronger predictor of current life satisfaction than was previous life satis-
faction (B= .18), although the strength of the association between current and past
REGRET IN LATER LIFE / 277
Table 3. Correlations between Study Variables
Variables 12345678910111213
1. Age
2. Gender
3. Marital status
4. Education level
5. Annual income
6. Frequency of regret
7. Number of reported regrets
8. Interpretive control
9. Health conditions
10. Life satisfaction
11. Past health conditions
12. Past life satisfaction
13. Past depression
—
.02
–.23**
–.09
–.13*
–.08
–.09
.02
.02
–.08
.09
–.05
–.01
—
–.42**
.05
–.37**
–.00
–.03
.10
.04
–.04
.06
–.10
.11
—
.10
.04
–.01
.12
–.05
–.07
.07
–.05
.18**
–.13*
—
.40**
–.01
.04
.01
–.10
.13
–.12
.09
–.19**
—
–.01
.02
–.07
.03
.12
–.02
.13*
–.17*
—
.77**
–.19**
.18**
–.26**
.09
–.11
.23**
—
–.10
.15*
–.17*
.10
–.02
.23**
—
–.21**
.32**
–.11
.11
–.15*
—
–.24**
.77**
–.12
.32**
—
–.23**
.25**
–.43**
—
–.16*
.31**
—
–.30** —
Notes: Pearson correlations used for all variables with the exception that Spearman correlations were used for the dichotomous variables of
gender and marital status. N= 228 for all correlations.
*p< .05. **p< .01.
278 / NEWALL ET AL.
life satisfaction may have been stronger had the same measures been used at
both points in time.
Turning to the hypothesis that frequency of regret would predict health, results
indicated that having a higher frequency of regret predicted having more health
conditions (B= .11, p< .01), even after controlling for sociodemographic factors
and past health. This result is particularly compelling given the strong association
between past health and current health (B= .76, p< .01). Again note that with
the addition of frequency of regret, the adjusted R2value significantly increased
from R2= .60 to R2= .61 (F= 7.03, p< .05).
In summary, as expected, frequency of regret predicted life satisfaction and
number of health conditions, findings which are particularly remarkable given that
demographic factors and prior measures of life satisfaction and health conditions
were included as covariates.
Supplementary Analyses
We performed two supplementary analyses related to the research question of
regret predicting health and life satisfaction. In our first supplementary analysis,
we addressed the possibility that it may simply be older adults’ level of depressed
mood which may affect life satisfaction, health, and regret. Thus, we re-ran the
same regression analyses described above in which the outcome variables were
REGRET IN LATER LIFE / 279
Table 4. Regression Beta Weights for Regret Predicting
Life Satisfaction and Health
Variables
Current life satisfaction
B
Current health conditions
B
Covariates
Age
Gender
Marital status
Education level
Annual income
Past life satisfaction
Past health conditions
Independent variable
Frequency of regret
–.07
–.03
–.01
.07
.05
.18**
–.16*
–.23**
–.04
–.03
–.04
–.03
.03
.01
.76**
.11**
F
adj R2
5.36**
.13
42.85**
.60
*p£.05. **p£.01.
regressed onto the background covariates, past health and life satisfaction and
frequency of regret, but statistically controlling for depression. Results (not
shown in Table 4) showed that regret continued to be a significant predictor of
life satisfaction (B= –.18, p< .01) and health (B= .09, p< .05) even when con-
trolling for depression.
In our second supplementary analysis, we investigated the distinction between
frequency of regret and the number of regrets reported by participants and which
variable may be the stronger predictor of health and life satisfaction.2Recall that in
total, 78 people reported no regrets, 71 people reported only one regret, and 79
people reported two regrets. Thus, we created a variable which reflected people’s
number of reported regrets, ranging from zero to two.
Not surprisingly, the correlation between number and frequency of regret was
quite strong (r= .77, p< .01). Note also that number of regrets correlated
significantly with both health (r= .15, p< .05) and life satisfaction (r= –.17,
p< .05) (Table 3). In order to examine whether frequency of regrets is a stronger
predictor of health and life satisfaction, we included number of regrets as a
predictor in the regression analysis in which the outcome variables were regressed
onto the background covariates, past health and life satisfaction, and frequency
of regrets. When we performed this supplementary analysis, we found that
although frequency of regret continued to significantly predict the outcomes,
number of regrets was not a significant predictor of life satisfaction (B= .05,
p> .05) or chronic conditions (B= –.02, p> . 05). This suggests that it is fre-
quency and not the number of reported regrets that is more important for health
and life satisfaction.
Secondary Interpretive Control Beliefs
Predicting Regret
To test the hypothesis that individuals with stronger secondary control beliefs
would report experiencing regret less frequently, correlations (Table 3) as well
as multiple regression analyses were performed. As expected, having stronger
interpretive control beliefs was associated with less frequent regret (r(226) =
–.19, p< .01 ) (Table 3). This is consistent with the proposed logic that possessing
a general tendency to strive to see the positive in a negative situation (or find a
“silver lining”) will serve to prevent or deactivate feelings of regret.
Next, we were interested in examining this relationship in the context of
other relevant variables. We performed a regression analysis in which secondary
interpretive control beliefs predicted frequency of regret, controlling for socio-
demographic variables, and past health, life satisfaction, and depression. Results
indicated that even controlling for all these variables, secondary interpretive
control beliefs continued to be a predictor of frequency of regret (B= –.16, p< .05;
280 / NEWALL ET AL.
2We thank anonymous reviewers for this suggestion.
F(9, 217) = 2.23, p< .05). Note that with the addition of secondary control beliefs,
the adjusted R2value significantly increased from adjR2= .03 to adjR2= .05
(F= 5.49, p< .05).
Supplementary Analyses: Secondary Control
Predicting Health and Life Satisfaction
Correlational analyses showed in the present study that secondary interpretive
control beliefs are associated with better life satisfaction and health. This was
expected, given past research which shows that secondary interpretive beliefs,
or related constructs such as finding benefit, tend to be associated with better
health and well-being (e.g., Affleck et al., 1987). Because in the present study
we additionally found that secondary interpretive control beliefs relate to regret,
and that regret relates to health and life satisfaction, in a set of supplementary
analyses, we examined the possibility that secondary interpretive control beliefs
may partially influence health and life satisfaction through its effect on regret.
As outlined by Baron and Kenny (1986), we first calculated the direct effects
of secondary control beliefs onto health and life satisfaction, accounting for the
covariates and past life satisfaction, health, and depression. We then controlled
for the effects of frequency of regret and estimated the significance of the media-
tional effect using Sobel tests. Our regression analysis showed that secondary
interpretive beliefs directly predicted life satisfaction (B= .26) in the context of
sociodemographic variables as well as controlling for the past outcomes of life
satisfaction, health, and depression (F(9, 217) = 9.30, p< .01). Note that with
the addition of secondary control beliefs, the adjusted R2value significantly
increased from adjR2= .19 to adjR2= .25 (F= 19.26, p< .01).When we repeated
this analysis with the addition of regret into the model, secondary interpretive
control beliefs (B= .24) along with regret (B= –.14) predicted life satisfaction.
Although this slight drop in the strength of association between beliefs and life
satisfaction (B= .26 to B= .24) suggests a partial mediation effect of regret,
Sobel’s test failed to support this mediation effect (z=1.69, p>.05).
Turning to health, our regression analysis showed that secondary interpretive
beliefs directly predicted health (B= –.12) in the context of sociodemographic
variables as well as controlling for past outcomes of life satisfaction, health, and
depression (F(9, 217) = 39.90, p< .01). Note that with the addition of secondary
control beliefs, the adjusted R2value significantly increased from adjR2= .59 to
adjR2= .61 (F= 7.93, p< .01).When we repeated this analysis with the addition of
regret into the model, secondary control beliefs (B= –.11) continued to predict
health; however, regret was only marginally associated with health (B= .08,
p= .08), which therefore fails to support a mediation effect. Altogether, these
analyses show the importance of secondary interpretive control beliefs in directly
predicting health and life satisfaction. However, little support was found for the
idea that secondary control beliefs may effect health and life satisfaction through
effects on regret.
REGRET IN LATER LIFE / 281
DISCUSSION
This study began with the question: Is old age “a regret”? It appears it need
not be, as approximately one-third of this sample of 79- to 98-year-olds reported
feeling no regret in the past couple of years. By the same token, approximately
two-thirds of participants indicated experiencing regret, some more frequently
than others. What factors can explain why certain individuals feel regret more
frequently than others? Based on results from the present study, sociodemographic
variables appear to have little bearing on regret. That is, being male or female,
married vs. unmarried, or having a higher vs. lower annual income or level of
education did not relate to how frequently older individuals experienced regret.
The absence of a gender difference is consistent with the study by Lecci et al.
(1994) showing that number of reported regrets was unrelated to being male or
female. Furthermore, again in response to the question of whether age relates to
regret, contrary perhaps to stereotypes, being older was not associated with
experiencing regret more frequently, although it is possible that no age differences
were found because of the relatively restricted age range of the present study
sample. Lecci et al.’s (1994) study in which they found that the older middle-
aged adults reported a fewer number of regrets than the younger adults, further
questions this stereotype, however.
Secondary Control and Regret
It appears that one individual difference factor that influences how frequently
older individuals experience regret is beliefs in finding the positive in negative
situations and the strategies and sense of secondary interpretive control that
presumably follow from these beliefs. As hypothesized, more strongly endorsing
secondary interpretive control beliefs was associated with lower frequency of
regret, and this relationship emerged even when accounting for sociodemographic
characteristics and past health, life satisfaction, and depression. This finding,
which complements past research (e.g., Torges et al., 2005), suggests “finding a
silver lining” may indeed act as a psychological repair mechanism as theorized
by Gilovich and Medvec (1995). Moreover, the finding is compelling as it
suggests that interventions designed to enhance secondary interpretive control
through helping people find meaning from negative outcomes, perhaps through
the development of interventions similar to Wrosch et al.’s (2007) writing
exercise interventions, may aid individuals in managing their negative emotions
like regrets.
Our findings also suggest that secondary interpretive control beliefs not only
have the potential to influence how frequently individuals experience emotions
such as regret, but also have important implications for the health and life
satisfaction of older individuals. This demonstrates the general beneficial effects
of looking for the positive or finding benefit in negative situations, as found
in past research in specific contexts (e.g., Affleck et al., 1987; Swift et al., 2008).
282 / NEWALL ET AL.
Note that supplementary regression analyses showed little support for the idea
that secondary interpretive control beliefs may not only influence health and
life satisfaction directly but also indirectly through effects on regret. Thus, the
processes through which secondary interpretive control, regret, and health are
connected remain unclear. Of future interest would be to examine what other
negative or even positive emotions might be affected by interpretive control
beliefs. It may be that for strong mediation effects to be observed, a more global
measure of affect would need to be incorporated into a model.
Regret, Health, and Life Satisfaction
The importance of designing interventions to help people manage their regret
is underscored by the study findings that regret frequency was related to life
satisfaction and health. In particular, as hypothesized, frequency of regret pre-
dicted poorer life satisfaction and number of health conditions even when
controlling for these same variables measured 2 years prior. These findings
correspond to past studies (using different measures of regret) which have found
an association between regret and life satisfaction and depression (DeGenova,
1992; Lecci et al., 1994; Wrosch et al., 2005). The findings also add to the larger
body of theoretical and empirical studies suggesting linkages between negative
emotions or negative affect and physical and psychological well-being (e.g.,
McKeen, Chipperfield, & Campbell, 2004; Meeks & Murrell, 2001; Salovey,
Rothman, Detweiler, & Steward, 2000). However, the mechanisms (e.g., physio-
logical, behavioral) through which negative emotions such as regret may affect
health deserve further study.
Content of Regret
In simply asking about people’s regrets either in terms of frequency or content,
the present study allowed participants themselves to define regret. That is, the
definition of regret was not prescribed by the researcher as being, for example, a
goal that they had never achieved (Lecci et al., 1994), neither were participants
told to report on life regrets (e.g., Wrosch et al., 2007). As such, the present
study may have tapped into regrets not typically captured in past studies, making
it difficult to compare across past studies. For example, contrary to past research
which has shown that people commonly report regretting education-related
inactions (Roese & Summerville, 2005) participants in the present study cited
only a small number of regrettable inactions relating to education (n= 8). It is
possible that had participants been asked to think back over their entire lives and
identify their biggest regrettable inaction, a larger number of people may have
reported education-related regrets. It is also possible that, in correspondence to
the “opportunity principle” (Roese & Summerville, 2005), these older adults had
different opportunities other than education to focus on and thus had different
regrets. Altogether, the results from this study suggest that, as researchers continue
REGRET IN LATER LIFE / 283
to refine the measurement of regret and differentiate it from related constructs and
emotions, there is some utility in beginning with a broad conception of regret.
Results surrounding the content of regrets are particularly intriguing in light
of the idea that they can be indicators of what people value or perhaps what
events people are facing at a given time in their life. When asked about the
causes or content of their regrets, individuals in the present study reported a
wide variety of regrets that fell into 12 qualitatively different categories. Inter-
estingly, there were several regret categories elucidated that seemed particularly
unique to an older population. These included such categories as “death of a loved
one,” “getting older,” “being limited,” and arguably even “health” and “lack of
connection with others.” As people get older and face difficult events such as
deaths of friends or age-related changes in health or capabilities, it is perhaps not
surprising that the content of regrets would focus around these events, reflecting
perhaps age-related developmental contexts (Jokisaari, 2004). This contrasts with
the career-related regrets faced at midlife (Stewart & Vandewater, 1999).
Using an approach that incorporated a broad definition of regret revealed
that many regrets older people identified were not necessarily goal-related or
associated with a sense of responsibility over an act or failure to act. Researchers
studying regrets in the context of decision-making have concluded that regret
stems from making counterfactual thoughts that focus on a personally controllable
feature of the event (Zeelenberg et al., 1998). Roese (1997) has discussed this “if
only” thinking as having causal inference effects on motivation and emotions as
guided by attribution theory. However, one could speculate that a contrast effect
(Roese, 1997), rather than a causal inference effect, could explain regrets that do
not reflect unattained goals (e.g., the regret someone feels about the death of a
loved one or about getting older). That is, the emotion may be arising out of a
simple comparison of reality with “what was” or “what could be.” Constraining
the definition of regret as relating to those events that result either from
controllable inactions or actions, for example, ignores other regrets that appear
to be commonly experienced by older individuals.
STUDY LIMITATIONS AND CONCLUSIONS
It is important to state that, as with all correlational and cross-sectional study
designs, causality could not be demonstrated in the present study; it was only
possible to determine how concepts varied in relation with one another and not
whether one variable caused changes in another. For example, although it was
argued that frequency of regret predicted health, it is also possible that the reverse
pattern holds true and that regret is a consequence and not a cause of poor health.
Although the present study design does not preclude this possible interpretation,
the study findings as presented are consistent with several longitudinal studies
that have shown negative affect to predict health outcomes measured at a later
time (Barefoot, Dahlstrom, & Williams, 1983; Pollard & Schwartz, 2003; and
284 / NEWALL ET AL.
see Taylor, 1990). Moreover, by controlling for past health and life satisfaction
outcomes, we were able to determine whether regret would be prospectively
associated with well-being.
Limitations surrounding the measure of regret are important to point out. For
example, the study could be criticized for the use of a one-item measure of
frequency of regret. By their nature, one-item measures cannot be assessed in
terms of their internal reliability; and it could be argued that complex constructs
cannot be measured with only one item. However, many studies focusing on
people’s regret have used one-item measures. Moreover, an item that asks people
about the frequency of which they experienced regret, certainly has good face
validity as a self-reported measure of frequency of regret. In future studies,
however, it would be interesting to use diary or time sampling methodologies to
assess frequency of regret.
In conclusion, this study adds to the small but growing body of literature on
regret in older individuals. More generally, it adds to research on the experience
of discrete emotions in later life. The results point to the idea that older indi-
viduals may be experiencing age-related regrets (e.g., related to health, loss of
a loved one) and even regrets about becoming older which appear to differ
from goal-related regrets. In addition, it must be highlighted that study results
point to the important role that secondary interpretive control may play not
only in helping to manage regrets but in enhancing or maintaining older indi-
viduals’ health and life satisfaction. In light of the potential to modify people’s
beliefs or use of strategies, these findings have practical implications as a basis
for designing interventions.
REFERENCES
Affleck, G., Tennen, H., Croog, S., & Levine, S. (1987). Causal attribution, perceived
control, and recovery from a heart attack. Journal of Social and Clinical Psychology,
5(3), 339-355.
Barefoot, J. C., Dahlstrom, W. G., & Williams, R. B. (1983). Hostility, CHD incidence, and
total mortality: A 25-yr follow-up study of 255 physicians. Psychosomatic Medicine,
45(1), 59-63.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1173-1182.
Chipperfield, J. G., Havens, B., & Doig, W. (1997). Method and description of the Aging
in Manitoba project: A 20-year longitudinal study. Canadian Journal on Aging, 16(4),
606-625.
Chipperfield, J. G., Perry, R. P., & Menec, V. H. (1999). Primary and secondary control
enhancing strategies: Implications for health in later life. Journal of Aging and Health,
11, 517-539.
Chipperfield, J. G., Perry, R. P., & Weiner, B. (2003). Discrete emotions in later life.
Journals of Gerontology: Psychological Sciences, 58B, P1-P12.
REGRET IN LATER LIFE / 285
Chipperfield, J. G., Perry, R. P., Weiner, B., & Newall, N. E. (in press). Causal antece-
dents of discrete emotions in late life. International Journal of Aging and Human
Development.
Davis, C. G., Nolen-Hoeksema, S., & Larson, J. (1998). Making sense of loss and
benefiting from the experience: Two construals of meaning. Journal of Personality
and Social Psychology, 75(2), 561-574.
DeGenova, M. K. (1992). If you had to live your life over again: What would you
do differently? International Journal of Aging and Human Development, 34(2),
135-143.
Disraeli, B. (1911). Coningsby. London: J. M. Dent and Sons, Ltd.
Folkman, S., & Moskowitz, J. T. (2000). Positive affect and the other side of coping.
American Psychologist, 55(6), 647-654.
Gilovich, T., & Medvec, V. H. (1995). The experience of regret: What, when and why.
Psychological Review, 102(2), 379-395.
Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review.
Review of General Psychology, 2(3), 271-299.
Gross, J. J., Carstensen, L. L., Tsai, J., Skorpen, C. G., & Hsu, A. Y. C. (1997). Emotion
and aging: Experience, expression, and control. Psychology and Aging, 12(4),
590-599.
Jeffries, S., & Konnert, C. (2002). Regret and psychological well-being among voluntarily
and involuntarily childless women and mothers. International Journal of Aging and
Human Development, 54(2), 89-106.
Jokisaari, M. (2003). Regret appraisals, age, and subjective well-being. Journal of
Research in Personality, 37, 487-503.
Jokisaari, M. (2004). Regrets and subjective well-being: A life course approach. Journal
of Adult Development, 11, 281-288.
Kahneman, D., & Tversky, A. (1982). The simulation heuristic. In D. Kahneman,
P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases
(pp. 201-208). New York: Cambridge University Press.
Kinnier, R. T., & Metha, A. T. (1989). Regrets and priorities at three stages of life.
Counseling and Values, 33, 182-193.
Landman, J. (1987). Regret: A theoretical and conceptual analysis. Journal for the
Theory of Social Behaviour, 17(2), 135-160.
Landman, J. (1988). Regret and elation following action and inaction: Affective responses
to positive versus negative outcomes. Personality and Social Psychology Bulletin,
13(4), 524-536.
Lecci, L., Okun, M., & Karoly, P. (1994). Life regrets and current goals as predictors
of psychological adjustment. Journal of Personality and Social Psychology, 66(4),
731-741.
Magai, C., Consedine, N. S., Krivoshekova, Y. S., Kudadjie, G. E., & McPherson, R.
(2006). Emotion experience and expression across the adult life span: Insights from
a multimodal assessment study. Psychology and Aging, 21(2), 303-317.
Mandel, D. R. (2003). Counterfactuals, emotions and context. Cognition and Emotion, 17,
139-159.
Markman, K. D., Gavanski, I., Sherman, S. J., & McMullen, M. N. (1993). The mental
simulation of better and worse possible worlds. Journal of Experimental Social
Psychology, 29, 87-109.
286 / NEWALL ET AL.
Mayne, T. J. (2001). Emotions and health. In T. J. Mayne & G. A. Bonnani (Eds.),
Emotions: Current issues and future directions. New York: Guilford Press.
McKeen, N. A., Chipperfield, J. G., & Campbell, D. W. (2004). A longitudinal analysis of
discrete negative emotions and health-services use in elderly individuals. Journal
of Aging and Health, 16(2), 204-227.
Meeks, S., & Murrell, S. A. (2001). Contribution of education to health and life satisfaction
in older adults mediated by negative affect. Journal of Aging and Health, 13(1),
92-119.
Morbidity and Mortality Weekly Report. (1989). Current trends comorbidity of chronic
conditions and disability among older persons—United States, 1984. Morbidity and
Mortality Weekly Report, 38(46), 788-791.
Mossey, J. M., Havens, B., Roos, N. P., & Shapiro, E. (1981). The Manitoba Longitudinal
Study on Aging: Description and methods. The Gerontologist, 21(5), 551-558.
Newall, N. E., Chipperfield, J. G., Blandford, A., Perry, R. P., & Havens, B. (2004).
The Successful Aging Study (2003): Methods and description. Technical Report
#HLHPRI104. Health, Leisure, & Human Performance Research Institute, University
of Manitoba, Winnipeg, Manitoba.
Newall, N. E., Chipperfield, J. G., Swift, A., Haynes, T. L., & Chuchmach, L. P. (2008).
The Successful Aging Study (2006): Methods and description. Technical Report.
Health, Leisure, & Human Performance Research Institute, University of Manitoba,
Winnipeg, Manitoba.
Neugarten, B. L., Havighurst, R. J., & Tobin, S. S. (1961). The measurement of life
satisfaction. Journal of Gerontology, 16, 134-143.
Ordonez, L. D., & Connolly, T. (2000). Regret and responsibility: A reply to Zeelenberg
et al. (1998). Organizational Behavior and Human Decision Processes, 81(1), 132-142.
Pollard, T., M., & Schwartz, J. E. (2003). Are changes in blood pressure and total
cholesterol related to changes in mood? An 18-month study of men and women. Health
Psychology, 22(1), 47-53.
Radloff, L. S. (1977). The CES-D scale: A self-report depression score for research in the
general population. Applied Psychological Measurement, 1, 385-401.
Radloff, L. S., & Teri, L. (1986). Use of the Center for Epidemiology Studies–Depression
Scale with older adults. Clinical Gerontologist, 5, 119-136.
Roese, N. J. (1997). Counterfactual thinking. Psychological Bulletin, 121(1), 133-148.
Roese, N. J., & Summerville, A. (2005). What we regret most . . . and why. Personality and
Social Psychology Bulletin, 31, 1273-1285.
Rothbaum, F., Weisz, J. R., & Snyder, S. S. (1982). Changing the world and changing the
self: A two-process model of perceived control. Journal of Personality and Social
Psychology, 42, 5-27.
Salovey, P., Rothman, A. J., Detweiler, J. B., & Steward, W. T. (2000). Emotional states
and physical health. American Psychologist, 55(1), 110-121.
Scheier, M. F., & Carver, C. S. (1985). Optimism, coping, and health: Assessment and
implications of generalized outcome expectancies. Health Psychology, 4, 219-247.
Sears, S. R., Stanton, A. L., & Danoff-Burg, S. (2003). The yellow brick road and the
emerald city: Benefit finding, positive reappraisal coping and posttraumatic growth
in women with early-stage breast cancer. Health Psychology, 22(5), 487-497.
Shimanoff, S. B. (1984). Commonly named emotions in everyday conversations. Per-
ceptual and Motor Skills, 58, 514.
REGRET IN LATER LIFE / 287
Stewart, A. J., & Vandewater, E. A. (1999). “If I had it to do over again . . .”: Midlife
review, midcourse corrections, and women’s well-being in midlife. Journal of
Personality and Social Psychology, 76(2), 270-283.
Swift, A. U. A. A., Bailis, D. S., Chipperfield, J. G., Ruthig, J. C., & Newall, N. E. (2008).
Gender differences in the adaptive influence of folk beliefs: A longitudinal study of
life satisfaction in aging. Canadian Journal of Behavioural Sciences, 40(2), 104-112.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). New
York: HarperCollins.
Taylor, S. E. (1990). Health psychology: The science and the field. American Psychologist,
45(1), 40-50.
Torges, C. M., Stewart, A. J., & Miner-Rubino, K. (2005). Personality after the prime of
life: Men and women coming to terms with regrets. Journal of Research in Personality,
39(1), 148-165.
Tucker, J. S., & Friedman, H. S. (1996). Emotion, personality and health. In C. Magai &
S. H. McFadden (Eds.), Handbook of emotion, adult development, and aging. Toronto:
Academic Press.
Weiner, B. (1985). An attributional theory of achievement motivation and emotion.
Psychological Review, 92(4), 548-573.
Wrosch, C., Bauer, I., Miller, G. E., & Lupien, S. (2007). Regret intensity, diurnal cortisol
secretion, and physical health in older individuals: Evidence for directional effects
and protective factors. Psychology and Aging, 22(2), 319-330.
Wrosch, C., Bauer, I., & Scheier, M. F. (2005). Regret and quality of life across the adult
life span: The influence of disengagement and available future goals. Psychology and
Aging, 20(4), 657-670.
Wrosch, C., & Heckhausen, J. (2002). Perceived control of life regrets: Good for young
and bad for old adults. Psychology and Aging, 17(2), 340-350.
Wrosch, C., Heckhausen, J., & Lachman, M. E. (2000). Primary and secondary control
strategies for managing health and financial stress across adulthood. Psychology and
Aging, 15(3), 387-399.
Zeelenberg, M., van den Bos, K., van Dijk, E., & Pieters, R. (2002). The inaction effect
in the psychology of regret. Journal of Personality and Social Psychology, 82(3),
314-327.
Zeelenberg, M., Van Dijk, W. W., & Manstead, A. S. R. (2000). Regret and responsibility
resolved? Evaluating Ordonez and Connolly’s (2000) conclusions. Organizational
Behavior and Human Decision Processes, 81(1), 143-154.
Zeelenberg, M., van Dijk, W. W., van der Pligt, J., Manstead, A. S. R., Empelen, P., &
Reinderman, D. (1998). Emotional reactions to the outcomes of decisions: The role of
counterfactual thought in the experience of regret and disappointment. Organizational
Behavior and Human Decision Processes, 75, 117-141.
Direct reprint requests to:
Nancy E. Newall
Psychology Department
University of Manitoba
Winnipeg, Manitoba R3T 2N2 Canada
e-mail: N_Newall@umanitoba.ca
288 / NEWALL ET AL.
