The prevalence,types and perceived outcomes of crisis episodes in early adulthood and midlife

Abstract

The objective of the study was to gain data on the prevalence, types and perceived outcomes of crisis episodes in three age decades of adult life: 20–29, 30–39 and 40–49. A further aim was to explore the relationship between crisis occurrence and empathy. A retrospective-autobiographical survey instrument and an empathy questionnaire were administered to 1023 UK-based adults. Prevalence data showed that crisis in the 20–29 decade was reported by 39% of men and 49% of women, while in the 30–39 decade 47% of men and 51% of women reported a crisis, and 46% of men and 59% of women reported a crisis occurring in the 40–49 decade. When prevalence rates were compared by participants’ current age, a recency effect in crisis reporting was apparent. Work-related crises were more common in men, while relationship and family were more common in female crisis. The most common crisis contents across decades and genders were divorce/relationship break-up and debt/financial difficulties. Post-crisis growth was significantly lower in the 40–44 age range in men than in other age ranges for both genders. There was a positive relationship between self-report empathy and number of crisis episodes experienced.

Keywords

autobiographical memory early adulthood midlife crisis quarterlife crisis turning points

The experience of crisis episodes in early adulthood and midlife has received considerable attention in the popular press over recent decades (e.g. Barr, 2004; Gaudette & Courter, 2011; Robbins & Wilner, 2001), but has been somewhat neglected in the scientific literature. The current study aims to be the first to gain prevalence estimates of crisis episodes in early adult and midlife age groups, while also gaining data on the types and perceived outcomes of crisis, within a cross-sectional adult sample. Before describing the methodology in detail, we briefly review the literature on crisis and turning points in early adulthood and midlife, and discuss the issue of autobiographical memory and its role in the formation of crisis episodes.

Autobiographical memory and crisis episodes

Autobiographical memory is comprised of memories from a person’s past that together constitute their personal life story (McAdams, 1993). This narrative of self becomes a key part of identity during adolescence and early adulthood (Conway & Pleydell-Pearce, 2000; Pasupathi, 2001). The central unit of the life story is the ‘nuclear episode’, which is a temporal segment of life that is stored, recalled and told as a unitized whole (McAdams, 2001). A crisis is defined and assessed in the current study as a particular form of nuclear episode – a past period of life that is characterized by difficulty, stress and negative emotion, and which is in retrospect viewed as a turning point or ‘bifurcation point’ in the life story (Robinson, 2008).

A crisis requires a beginning and an end to be stored as a nuclear episode within autobiographical memory and, given this, it exists by definition in hindsight. Crisis episodes are therefore appropriately assessed using retrospective techniques, as the capacity of adults to find thematic coherence in autobiographical memory brings about the division of life into demarked episodes and periods (Pasupathi, 2001). However there are inevitable limitations of a retrospective approach; for example, there is evidence of a recency effect in autobiographical memory (de Vries & Watt, 1996). There is also a tendency for a concentration of life story events in early adulthood – termed the ‘reminiscence bump’ (Kotre, 1995; Rubin & Schulkind, 1997). Furthermore, autobiographical memory is shaped by cultural life scripts – idealized prototypes of when and how life transitions and stages should occur that are normative in a particular culture (Bernsten & Rubin, 2004). Individual life stories thus have cultural values and identifiable dominant narratives woven into them (Bernsten & Bohm, 2009; Hammack & Cohler, 2011). McAdams refers to autobiographical recollections of positive outcomes following difficult life episodes as redemptive narratives, and suggests that they are key to finding meaning within the ups and down of the life story (McAdams, 2013). The redemption narrative form is similar to the rebirth plot that has also been linked to accounts of crisis episodes in early adulthood (Robinson & Smith, 2009). Recollections of post-crisis growth may be in part shaped by the redemptive/rebirth master narrative, and conversely post-crisis decline may be influenced by the tragedy master narrative (Booker, 2005). Thus self-perceptions of post-crisis growth are likely to be shaped both by prospective changes following a crisis and by retrospective narrative reconstruction. In summary, recollections of crisis must be understood as both facilitated by and limited by the contours of autobiographical memory and culture.

Crisis episodes contain stressful and challenging life events, and those events are typically central to the retrospective recognition of that time as a crisis, for they may act as a symbolic turning point in the life story, and readjustment required following the event may be recalled as a key locus of change during the crisis (O’Connor & Wolfe, 1987; Robinson, Wright, & Smith, 2013). The literature on the epidemiology of stressful life events suggests that young adults are more likely to perceive recent life events as stressful and traumatic, and more likely to report having difficulty in adjusting to such events, relative to adults in midlife or old age (Masuda & Holmes, 1978; Norris, 1992; Hatch & Dohrenwend, 2007; Turner, Wheaton, & Lloyd, 1995). Young adults are also more likely to report experiences of abuse or violence than their older counterparts. There is also a normative gender split in the types of events recalled as stressful or traumatic; women more often refer to events involving significant others and to bereavements as stressful, while men report a greater number of overall traumatic events (Turner & Lloyd, 1995; Turner et al., 1995).

A number of theorists have proposed that crises can play a functional role in positive adult development (Denne & Thomson, 1991; Erikson, 1968; Forer, 1963; Levinson, 1978, 1996; King, 2001; Sheehy, 1977). Firstly, crisis episodes can act as a stimulus for change when a person is living in ways that are problematic or unhealthy, by helping to overcome the inertia that adult life commitments exert (Kegan & Lahey, 2009). Secondly, a resolved crisis often leads to self-evaluations of personal strength for having successfully resolved a difficult and emotionally challenging life episode, which in turn can lead to heightened self-esteem and an enhanced sense of autonomy (Robinson, 2008). Thirdly, experiences of crisis may help a person to better empathize with others who are going through difficult times in their lives (Denne & Thomson, 1991; Levinson, 1996). The theoretical rationale for this empathy–crisis link is that affective empathy involves vicariously feeling the distress of another (as opposed to cognitive empathy, which involves simply recognizing an emotion in another), and the capacity for such emotional empathy is enhanced if the person has, at a point previous in life, experienced the feeling in question themselves (Trusty, Ng, & Watts, 2005). Crises contain enduring and intense periods of negative affect (Robinson & Smith, 2010a), and thus provide such an opportunity.

Corresponding to these theoretical functional roles for crisis in development, appraisals of positive post-crisis growth are common (Denne & Thompson, 1991; Robinson & Smith, 2010a). Post-crisis growth is similar to the construct of post-traumatic growth (Tedeschi & Calhoun, 1995). Both involve appraising a negative life experience as having enduring positive effects. The difference between them is that trauma is usually a relatively sudden event such as a natural disaster or a car crash (Tedeschi & Calhoun, 2004), whereas crises are lengthier developmental episodes that typically last several years (O’Connor & Wolfe, 1987; Robinson, 2008). Furthermore, post-traumatic growth is focused on cognitive-affective change, for example increased importance of relationships, greater appreciation of life, and greater awareness of possibilities (Tedeschi & Calhoun, 2004), while post-crisis growth also focuses on external changes in relationships, work and life structure (Robinson, 2008).

Crisis in early adulthood

Early adulthood is the phase of life starting after adolescence and continuing through to approximately the age of 40 (Erikson, 1980; Levinson, 1986). Erikson stated that the experience of crisis is normative at the beginning of this stage, as a result of diverse challenges in the transition from adolescence to adulthood (Erikson, 1980). The first years of adult life are characterized by instability in relationships, identity, social networks and living arrangements (Arnett, 2000), and by a heightened vulnerability to mental illnesses compared with other points in the lifespan (Robinson, 2012). Such crisis in the first decade of early adulthood is frequently referred to in both the popular and academic press as ‘quarterlife crisis’ (Robbins & Wilner, 2001), and we will use that term in this article. Another age range in early adulthood that has been highlighted as a time of transition and possible crisis is the Age Thirty Transition, which runs approximately between the ages of 28 and 33 (Levinson, 1986). This period, which typically involves a re-evaluation of the lifestyle and life structure that defined one’s 20s, can become a crisis if a person finds that the life structure that they developed during their 20s is causing them distress (Levinson, 1986).

A programme of qualitative research has been conducted into the process and experience of crisis in early adulthood (Robinson, 2008; Robinson & Smith, 2010a, 2010b, 2010c; Robinson, Wright, & Smith, 2013). Semi-structured interviews were conducted with individuals who reported experiencing a crisis between the ages of 25 and 39. It was found that crisis episodes in the age group revolved around distressing endings in career or relationship or both, and the corresponding demise of a life structure that has been constructed around these commitments during the first decade of adult life. Crisis episodes showed a common pattern of phases: (1) being locked in to a relationship and/or job, (2) a painful separation from that commitment and taking time out, (3) exploratory searching for a new life structure, and (4) rebuilding a new, more stable life structure.

Crisis in midlife

Midlife crisis is the term given to crisis episodes that occur in the 40–49 age range (Lachman, 2004; O’Connor & Wolfe, 1987). Levinson (1978) suggests that adults at this age experience a juxtaposition of their young self and old self, as the effects of ageing become more apparent while youthful vigour is also very much present. A model of midlife crisis was developed by O’Connor and Wolfe (1987) based on analysis of interviews with people in their 40s. Central features of the model are the presence of strong negative emotions and major changes, showing that midlife crisis is, when it happens, not just a time of internal crisis but also a time of external transition. So, as with early adult crisis, evidence on the contents of midlife crisis suggests it centres on major, tangible changes in life structure, as well as challenges and changes to identity and affect.

Prevalence estimates of crisis are rare, but one research study that has provided a prevalence estimate of midlife crisis is Wethington’s analysis of interview data from the MIDUS (Midlife in the United States) study. In adults over the age of 50, it was found that 34% of men and 36% of women reported a midlife crisis (Wethington, 2000). It was found that more women than men reported that their crisis included problems with family and relationships, while more men than women reported midlife crises including work and career problems.

While crises are by definition difficult and stressful life episodes, turning points include not only negative events, but also positive episodes such as fulfilling a dream or having a child (Wethington, Kessler, & Pixley, 2004). Several studies have concluded that important turning points are reported more in early adulthood than in midlife (Clausen, 1995; Wethington, Kessler & Pixley, 2004). As with Wethington’s data on midlife crisis, men more often report turning points about work than women, while more women than men report turning points about marriage and family (Clausen, 1995; Thurnher, 1983).

Aims and hypotheses

Given the paucity of quantitative research on crises episodes in early adulthood and midlife, much remains unknown about their frequency, contents and outcomes in adult lives. The aim of the study was to gain data on crises occurring between the ages of 20 and 50, using an innovative retrospective-autobiographical survey instrument developed by appropriating techniques from psychiatric epidemiology. Throughout the article, crisis will be referred to in relation to three age decades of occurrence: 20–29 will be referred to as the quarterlife decade, 30–39 will be referred to as the 30s decade, while 40–49 will be referred to as the midlife decade. The rationale for assessing crisis in relation to these three decades stems from several theories of crisis as well as Levinson’s theory of adult life stages. Crisis episodes specific to one’s 20s are now widely referred to as quarterlife crisis, and as such have a theoretical existence on their own terms (Robinson, 2011), and in Levinson’s model were defined as being prior to the ‘Age 30 Transition’. The 30–39 decade is referred to by Levinson as the ‘Settling Down’ decade, and is described as having particular developmental challenges and potential for crisis (Levinson, 1986). Furthermore, midlife is frequently taken for research purposes to refer to the age range of 40 to 59, with the 40–49 decade being ‘early’ midlife and 50–59 being ‘late’ midlife (Lachman, 2004). Midlife transition and crisis most frequently have been theoretically and empirically attributed to the 40–49 early midlife decade (Brim, 1976; Levinson, 1978; O’Conner and Wolfe, 1987), and thus we focus here on that decade too in our analysis of midlife crisis.

Based on existing research, a limited set of predictions could be made. With regards to prevalence, it was tentatively hypothesized that at least a third of adults would report crisis episodes in each decade of occurrence, based on findings from Wethington (2000) on midlife crisis prevalence in a US sample. An interaction between gender and age of crisis was predicted; more women than men would report crisis episodes in early adulthood, and more men than women would report them in midlife. This hypothesis was based on findings from the MIDUS survey reported by Wethington, Kessler, and Pixley (2004), in which it was found that women reported turning points and identity shifts more in the early adult years, while men reported more in the midlife age group.

Based on autobiographical memory theory, a recency effect was predicted in prevalence rates; more crises would be reported during the decades that were recent relative to a participant’s current age compared with those further back in the past. This was expected due to the fact that negative emotional salience of memories fades over time, making more distant memories less emotionally charged and so less likely to be labelled a crisis (Walker, Vogl, & Thompson, 1997)

A list of 23 crisis event types was developed theoretically based on previous research. Given the paucity of previous quantitative research on how the contents of crisis episodes relate to the age of a crisis and to gender, only one hypothesis was put forward, based on the aforementioned findings pertaining to turning points and stressful life events in adulthood: men would report work-related crises more than women, and women would report relationship and family-related problems more than men. Beyond this prediction, our analysis was exploratory.

For post-crisis growth, it was predicted that positive post-crisis growth would be perceived by the majority of both genders in all three decades of occurrence. This hypothesis was based on the functional role that theorists such as Erikson have given to crisis in development. Levels of post-crisis growth across gender and age group were compared in an exploratory manner. Based on the finding that increased empathy with others’ difficulties and distress has been mentioned as a self-reported outcome of personal crisis (Denne & Thompson, 1991), a final outcome-related hypothesis was that those who have been through crises will be more empathic than those who have not.

Method

Participants

1023 UK-based adults over the age of 25 participated in the study, of whom 42% were male and 58% were female. An additional 101 adults commenced the study but did not complete it, indicating a drop-out rate of 9%. The strategies by which participants were recruited are described in the Procedure section below. The proportion of males and females broken down by 5-year age groups and by household income is shown in Table 1. Excluding students in the sample, mean household income for the sample was £24,600, similar to but slightly under the average annual household UK income of £28,400.¹ The sample was 96% white, 1% black, 0.6% Asian, 1% mixed ethnicity and 1.6% not specified. This is a marginally higher proportion of white participants than the UK population more generally (92% white²). With regards to working status, 39.1% were in full employment, 16.4% were in part-time employment, 4.2% were stay-at-home parents, 10.2% were unemployed, 5.7% were students and 24.4% were retired. With regards to marital status, 19.6% were in a relationship or cohabiting, 38.2% were married, 1.7% were in a civil partnership, 12.5% were divorced, 21% were single and 4% were widowed. These working and marital status percentages are within 5% of figures for UK adults as a whole.³ Participants were geographically dispersed across the UK; 16% were based in London and the South East, 12% were based in South West England, 48% were based in Central or Northern regions of England, 10% in Scotland, and 1% in Wales or Northern Ireland.

Table 1.

Sample numbers by gender, age band, and household income split.

	Gender		Household Income
Age bands	Male	Female	£0-20K	£21-40K	£41-60K	£61-80K	£81K+
25–29	47	54	40	35	20	2	4
30–34	42	49	30	25	12	10	14
35–39	32	73	35	34	14	9	13
40–44	47	67	35	31	26	4	18
45–49	56	57	43	27	22	9	12
50–54	55	53	32	26	25	7	18
55–59	40	69	42	36	21	6	4
60–64	40	83	48	48	16	7	4
65–69	31	60	37	31	14	6	3
70+	37	31	33	25	6	1	3

Measures

Adult crisis episode retrospective self-assessment tool (ACERSAT)

To gain structured data about crisis episodes, a self-report tool was developed that was based on retrospective assessment techniques used in psychiatric epidemiology such as the Adult Psychiatric Morbidity Survey (McManus & Bebbington, 2009). The instrument provided data on: (A) the occurrence and timing of crisis episodes, or lack thereof, (B) the events within crisis episodes, and (C) perceived post-crisis growth or decline. The instrument was delivered online, as it requires the skip logic and display logic that online survey tools provide, in order to selectively show only relevant items in parts B and C depending on responses to Part A.

Part A. Occurrence and timing of crises

Participants were given the following simple definition of a crisis episode, based on the theory and research outlined earlier in the literature review: ‘A crisis episode is a period in adult life that is noticeably more difficult, stressful and unstable than normal, and is an important turning point in your life due to changes that occur during it. Crisis episodes typically last for several years, but may be shorter or longer.’ This was followed by a question: ‘Have you experienced any times of crisis in your adult life?’, to which participants responded either ‘Yes’ or ‘No’. If they selected ‘No’, they skipped the remainder of the ACERSAT. If they responded ‘Yes’, they were directed to the second page of the questionnaire, where they were asked to indicate the ages during which the crisis or crises occurred. They were presented with a visual display of 5-year age intervals (the same as shown in Table 1), and were instructed that they could select as many or as few as they like to indicate when crises occurred. Participants were only shown those 5-year age intervals that were either the same or lower than their current age; this was achieved using a display filter from an earlier question requesting the participant’s current age.

Part B. Types of crisis events

For each crisis episode indicated in Part A, an additional page was presented with a list of 23 crisis event types with a check box next to each one. An instruction above the items was given to participants: ‘Please indicate which of the following were present during the episode that occurred at some point between the ages of [age range given]. Select as many or as few of the following as are relevant.’ The list was compiled based on an exhaustive review of the literature on the contents of adult crisis and types of stressful life transitions, and was aimed to conceptually cover challenging or negative events in the key domains of the adult life structure, as contained in classic holistic theories of adult development such as Levinson (1996) and Super (1980). These domains include career, health, family, relationships, sexuality, worldview and identity. The aim was to have a comprehensive list, but not so many that perusing all the items would be too time-consuming for participants. The item wordings are shown in full in Table 2.

Table 2.

Frequencies of response to each item, by gender and age decade of crisis, shown as % of total number of crisis episodes for that age/gender, and results of Chi-square gender difference and age difference tests.

	Female				Male
	20–29	30–39	40–49	χ - age diff	20-29	30-39	40-49	χ - age diff
Feeling trapped in a job you didn’t want to be in any more	22.5	21.1	13.4	ns	33.3	28.7	28.7 ^g	ns
Experiencing a high level of stress and pressure in your job	21.7	25.4	22.6	ns	30.3	37.5	41.6 ^g	ns
Resigning from a job	11.1	12.0	9.2	ns	15.6	14.4	13.5	ns
Being fired from a job	5.7	3.7	2.5	ns	10.0	9.7	7.9	ns
An unwanted period of unemployment	13.7	13.4	12.0	ns	29.9 ^g	25.5 ^g	23.6 ^g	ns
A change of career	11.1	15.1	14.1	ns	18.6	18.5	16.9	ns
Debt or financial difficulties	29.2	31.1	28.3	ns	29.4	26.4	33.7	ns
Being locked into in a relationship that you no longer wanted to be in	27.6	22.6	20.5	ns	16.9	22.7	14.6	ns
Being in an abusive relationship	23.3 ^g	16.3	13.1	10.31*	10.4	7.4	7.9	ns
Divorce or relationship break-up	31.3	40.6	26.9	9.54*	26.8	30.6	27.0	ns
Death of a person close to you	23.5	18.9	34.3	ns	23.4	22.7	27.0	ns
The birth of a child	15.8	11.7	3.2	24.13*	5.2	8.3	2.2	ns
Difficulties with your child’s behaviour	5.7 ^g	9.1	17.7	23.66*	1.3	2.3	4.5	ns
Family conflict or dispute	25.1	21.1	18.7 ^g	ns	20.3	18.1	13.5	ns
Confusion over sexual preference	3.1	1.1	0.4	ns	6.1	3.2	1.7	ns
‘Coming out’: Making sexual preference known to friends and family	1.3	0.0	1.1	ns	5.2	1.9	0.0	ns
Uncertainty or confusion in your own beliefs or worldview	15.8	9.1	4.2	21.78*	19.9	17.1	18.5	ns
Alcohol/substance abuse or addiction	10.6	7.1	4.2	ns	14.7	9.7	11.8 ^g	ns
Physical illness or injury to yourself	18.3	18.0	24.0	ns	13.9	21.8	19.1	ns
Physical illness or injury to someone close to you	9.3	9.4	15.5	ns	7.4	5.1	10.7	ns
Caring for an ill or disabled loved one	7.2	8.3	17.7	19.49*	3.9	3.7	9.0	ns
Being diagnosed with a mental illness	15.5	10.0	7.4	10.28*	15.2	15.3	15.2	ns
A person close to you being diagnosed with mental illness	4.7	3.4	8.5	ns	3.5	3.7	2.2	ns

Notes. Bold: Score in top quartile of frequencies (over 22.6%). ^g + italics = Figure is significantly higher than matched figure for other gender. NB. Corrected significance: 0.05 / 23 = p < 0.002.

Box: Item is in top quartile of frequencies in all three crisis decades across both genders. Grey: Most prevalent item in column.

Part C. Perceived post-crisis growth

For each crisis reported in Part A of the ACERSAT, participants were asked to rate their agreement with four items that described negative or positive post-crisis change, on a 5-point Likert Scale from ‘Strongly Disagree’ to ‘Agree’. Two of the items pertain to changes in the self (‘As a result of this episode I became a stronger person in the long-term’ and ‘As a result of this episode my self-esteem decreased in the long-term’), while two items pertain to changes in lifestyle and life structure (‘The result of this episode was that my quality of life decreased in the long-term’ and ‘During this episode I made changes to my life that had a long-term positive effect’). The negative items were reversed, and post-crisis growth was calculated as the mean of the scores on the four items. The four items showed strong internal consistency (Cronbach α = .82), supporting the internal reliability of post-crisis growth as a singular construct. For each participant, an overall post-crisis growth score was calculated as the mean of all post-crisis growth scores.

Toronto empathy questionnaire (TEQ)

Empathy was measured using the Toronto Empathy Questionnaire (TEQ) (Spreng, McKinnon, Mar, & Levine, 2009). This 16-item questionnaire assesses behavioural, emotional, cognitive and physiological facets of empathy. Responses are given on a 5-point scale from Never (0) to Always (5). Example items include ‘When someone else is feeling excited, I tend to get excited too’, ‘I have concerned feelings for people less fortunate than me’ and ‘I can tell when others are sad even when they do not say anything’. In the current sample, Cronbach alpha coefficients demonstrated acceptably high internal consistency, α = 0.78.

Procedure

Recruitment of participants was achieved by publicizing the study online and using a postal strategy. The postal strategy involved distributing 1500 letters by hand to neighbourhoods around London. A mix of neighbourhoods was chosen to include a mix of low-income, middle-income and high-income areas. The letter gave a brief summary of the study and a short web link to access it online.

The online publicity outlets were: (1) A feature on the study in a major science magazine blog, (2) A paid advertisement on a major social networking site, (3) A free feature in an online newsletter distributed to adults in retirement, and (4) Inclusion in an online SONA ‘participation portal’ at an adult education college. These were aimed at locating adults of both genders and across different demographic groups.

After accessing the online survey, participants were first given information about the study and the requirements of participation. Participants were informed that they could choose after completing the survey to enter a raffle for one of 20 x £50 vouchers. Having confirmed their consent to participate, questionnaires were presented on screen.

Results

Crisis prevalence and timing for males and females

It was tentatively hypothesized, based on the past research on crisis prevalence, that at least one third of adults would report a crisis in each of the three decades: 20–29 (quarterlife decade), 30–39 (30s decade) and 40–49 (midlife decade). All adults older than the decade of occurrence were included in the prevalence calculation. The resulting prevalence percentages are shown in Figure 1: 39% of men and 50% of women over the age of 30 reported a quarterlife crisis, 47% of men and 51% of women over the age of 40 reported a 30s crisis, while 46% of men and 59% of women over the age of 50 reported a midlife crisis. This supported our prediction that a minimum of 33% would report crises within each decade.

Figure 1.

Percentages of males and females (over the age of relevant crisis decade) retrospectively reporting a crisis in the chronological age decades of 20–29 (N = 922), 30–39 (N = 726) and 40–49 (N = 499).

To test whether the presence of crisis differed in prevalence by decade of occurrence, a 3x1 Chi-square goodness of fit test was run within each gender, and expected frequencies were weighted according to the sample sizes contributing to prevalence estimates. The outcome was that no significant association between decade of occurrence and crisis prevalence was found for either gender.

To test whether men and women differed significantly in the number of quarterlife, 30s and midlife crises, 2x2 Chi-square tests were run for each decade with (a) gender and (b) crisis/no crisis as two-level variables. Significant differences were observed for all three respectively, (quarterlife crisis χ² ₍₁₎ = 10.133, p < .001, 30s crisis χ² ₍₁₎ = 1.352, p < .05, midlife crisis χ² ₍₁₎ = 7.96, p < .01), indicating that all three types of crisis event were more prevalent amongst females than males, which meant that our hypothesis that gender would interact with prevalence across the three decades (women higher in early adulthood, men higher in midlife), was not supported.

Recency effect

We hypothesised that recency would positively relate to crisis prevalence. In order to test this prediction, percentages of crises reported in the age decades were compared by participants’ current age. The results are shown in Figure 2; for all crises occurring less than one decade ago, estimates are between 62% and 72%, while for decades occurring more than ten years in the past, prevalence estimates are between 33% and 49%.

Figure 2.

Crisis prevalence percentages for the three crisis decades, stratified by current age of participant.

Chi-square tests were conducted to test the relationship between prevalence rates and current age. For quarterlife crisis a 2x5 Chi-square test indicated that there was a significant relationship between current age and prevalence of crisis (χ² ₍₄₎ = 73.157, p <.001), while a 2x4 analysis conducted for 30s crisis also revealed a significant result (χ² ₍₃₎ = 30.549, p <.001), and a significant relationship was observed for a 2x3 analysis of data for midlife crisis (χ² ₍₂₎ = 12.608, p = .0018). This supports the hypothesis that recent episodes of crisis in all three decades would be more commonly reported than those in decades occurring less recently.

Types of crisis events

The frequency with which crisis event types were endorsed is given in Table 2, given as a percentage of total crisis episodes for each particular age decade and gender. For the sake of parsimony, the descriptive account of these findings below only refers to those items that featured in the top quartile of frequencies (22.6% or higher), and are thus most indicative of crisis.

The most frequent type of quarterlife crisis for female participants was divorce or relationship break-up, and the following were also in the top quartile: debt/financial difficulties, being locked into in a relationship that you no longer wanted to be in, family conflict or dispute, being in an abusive relationship and death of a person close to you. For 30s crisis in women, the first two most frequent items were also divorce/break-up and debt/financial difficulties, while job stress/pressure and being locked into a relationship were also in the top quartile. For women in midlife, death of a person close to you is the most common feature of crisis, and debt/financial difficulties, divorce/relationship break-up, physical illness or injury and job stress/pressure are also in the top quartile. Therefore the most prevalent feature of crisis in all three decades is relationship-orientated.

For male participants, crisis in the quarterlife decade most frequently contained feeling trapped in a job, followed by job pressure/stress, unemployment, debt/financial difficulties, divorce/relationship break-up and bereavement. In 30s crisis, the most commonly cited type of crisis is job stress/pressure, followed by divorce/relationship break-up, feeling trapped in a job, debt/financial difficulties, unemployment and bereavement. In midlife, job stress/pressure is also the most frequent item, followed by debt/financial difficulties, feeling trapped in a job, death of a person close to you and divorce/relationship break-up. Therefore the most frequently cited feature of crisis for men in all three decades is work-related.

When comparing findings across genders and decades, two themes appear in the top quartile for both genders in every crisis type, and are shown in boxes on Table 2: divorce/relationship break-up, and debt/financial difficulties. These are the most consistently present contents of crisis across gender and age decade.

Gender differences in the data were tested using Chi-square tests, and a significant difference is indicated in Table 2 as ‘^g’ next to the higher of the two figures. Given the number of tests required, a Bonferroni correction was used to lower the p-value for each set of tests run. Tests were calculated in sets of 23, thus the corrected p value was 0.05/23 = 0.002. Unemployment as a type of crisis was higher in men than women in all age groups, while feeling trapped in a job and job stress/pressure was significantly higher in men than women in midlife. Being in an abusive relationship was higher in women than men in the quarterlife decade. Other gender differences that do not refer to a figure within the top quartile can be observed in Table 2.

Chi-square tests were employed to establish the significance of age differences for all crisis event types. The same Bonferroni correction was used as with the gender difference tests. When female prevalence figures were compared by age decade, seven emerged as significantly different (being in an abusive relationship, divorce/break-up, birth of child, difficulties with child’s behaviour, uncertainty in beliefs/worldview, caring for ill/disabled loved one, being diagnosed with mental illness), yet no male prevalence figures differed by age.

Perceived post-crisis growth

Figure 3 shows the percentage of adults for whom three out of four post-crisis growth items were rated as positive, therefore for whom positive post-crisis growth outweighed decline. Five-year age ranges were used for this analysis, rather than the ten-year groups used in the prevalence and crisis-type analysis, as the continuous data and parametric statistics used with this data allowed for a more fine-grained grouping variable than the frequency data permitted. It was predicted that the majority of adults would report positive post-crisis growth. The descriptive data provide partial support for this: between 48% and 56% of men and women appraise positive post-crisis growth, with the exception of crisis occurring for men in the 40–44 age range, of which only 31.2% are reported as positive growth events.

Figure 3.

Percentage of crises rated as having positive post-crisis growth, broken down by five-year age interval at which crisis occurred.

Figure 3 illustrates that crisis occurring between the ages of 40 and 44 in men had lower post-crisis growth scores than other age ranges in either gender. In order to test whether differences between gender and age-of-occurrence groups were significant at the mean level, a two-way ANOVA was conducted with gender and age of crisis as independent variables and post-crisis growth as the dependent variable. It was found that the effects of both age (F _(5,1635) = 2.26, p < 0.05) and sex (F _(1,1635) = 8.31, p < 0.005) emerged as a significant effect within the ANOVA model, but the interaction between them did not. Post-hoc tests established that the source of the significant difference between the age groups was the difference between the 40–44 age group and other age groups.

Empathy and presence of past crisis

It was hypothesized that individuals who have experienced adult crisis episodes would have increased empathy for others, compared with those who had not. The rationale for this was that individuals who had experienced difficult and stressful crises in their lives would find it easier to imagine what others would experience if they go through similarly hard times. In order to test this, adults over the age of 50 were sampled, and the number of crises that they had experienced between the ages of 20 and 49 was computed. Mean empathy levels were then compared across the four groups of: no crises (N = 90), 1 crisis (N = 177), 2 crises (N = 142) and 3 or more crisis (N = 90). The results are shown in Figure 4.

Figure 4.

Mean empathy levels across four groups defined by number of crises experienced between ages 20 and 49: 0, 1, 2 or 3+ crises.

An ANOVA was conducted with number of crises as the IV and empathy as the DV, and the difference between the groups was statistically significant, F _(3,495) = 8.11, p < 0.001. Post-hoc tests showed the source of this effect to be between the 3+ group and all the other groups. In order to control for gender as a possible confound, the ANOVA was also run with gender as a covariate, and the effect of empathy remained significant, F _(3,495) = 6.35, p < 0.001.

Discussion

This study aimed to explore the recollection of crisis episodes in early adulthood and midlife. We predicted that at least one-third of adults would report a crisis in each decade, and this was indeed the case for both genders. On average, women reported significantly more crises than men, which was contrary to our hypothesis that more women than men would report quarterlife crisis, but that more men than women would report midlife crises. The finding of higher prevalence rates in women could be interpreted in a number of ways. It may suggest that women are more likely to admit to crisis than men, which would fit with research showing higher levels of self-disclosure in women than men (Dindia & Allen, 1992). Alternatively, it may reflect the fact that there are contradictory demands on women in contemporary society that make crisis genuinely more likely (Hochschild & Machung, 1990), or that women are more prone to retrospectively perceiving crisis in adulthood due to some dispositional factor such as higher average levels of neuroticism compared with men (Srivastava, John, Gosling, & Potter, 2003), or the higher prevalence of depression in women (Nolen-Hoeksema, 1987), which may affect the likelihood of recalled crisis.

As expected, evidence of a recency effect in the reporting of crisis episodes was also found – there was higher prevalence within the last ten years of life than in earlier decades. This supports the existence of a recency effect in autobiographical memory (de Vries & Watt, 1996), and extends the empirical understanding of this effect by showing that it applies to crisis episodes within the life story. An alternative interpretation, however, is that events over the ten years of history prior to the study have made adult crises more likely, and that what appears to be a recency memory effect is in fact a ‘time-of-measurement effect’ caused by recent historical events (Schaie & Willis, 2002). A higher proportion of young adults reported crises in the past ten years compared with older adults, which supports previous findings on stressful life events – that younger adults are more likely to retrospectively appraise events as stressful or traumatic than older adults (Norris, 1992; Hatch & Dohrenwend, 2007).

Types of crisis event were found to differ both by gender and decade. Women’s crises focused more on family and relationship problems than men (e.g. divorce, break-up, abusive relationships, family disputes and difficulties with children), while male crises focused more on work-related problems than women (e.g. feeling trapped in a job, job stress/pressure and unemployment). These gender differences support previous findings on stressful life events (Turner et al., 1995) and may reflect a different valence that men and women attach to work and relationship-related life events, or to the autobiographical memories of those events. It supports the view that the life story and autobiographical memory are created, interpreted and combined within cultural frameworks that have gender expectations embedded within them (Fivush, 2011). Seven types of crisis were differentiated by age in the female sample, but none were age-graded in men, suggesting that for women, the contents of crisis are more age-specific than for men.

Crisis events found in the quarterlife decade in this study are congruent with the model of early adult crisis developed by Robinson and Smith (2010b) from qualitative data. The model suggests that a feeling of being ‘trapped’ or ‘locked into’ a life commitment is a common aspect of early adult crisis, and in the current study ‘feeling trapped in a job you didn’t want to be in’ was the most frequently endorsed item for men in this decade, while ‘being locked into a relationship that you no longer wanted to be in’ was the third most frequently endorsed quarterlife item for women.

A feature of midlife crisis was the presence of bereavement as the most prevalent feature amongst women, and the fourth most prevalent in men. This supports the original theory of midlife crisis, put forward by Jaques (1965), which stated that midlife crisis stems from a growing awareness of death and mortality. It may be that bereavement is a feature of crisis episodes in midlife because it brings about greater awareness of one’s own mortality and the challenges of ageing. Given that our data do not signify the relationships of the bereaved to the participants, more research needs to be done to explore the nature of this link between bereavement and midlife crisis, to understand whether it is parental bereavement that is the most common and, if so, what interpersonal, cognitive and emotional processes mediate it as a crisis event.

Positive post-crisis growth was experienced by approximately half (48.1% to 55.8%) of the participants in the study, with the exception of men in the 40–44 age range, for whom 31.2% reported positive post-crisis growth. This may relate to the original link between midlife crisis and men (Jaques, 1965). While in our sample men did not experience crisis more than women in midlife, they found it harder to see episodes in this period of life in a positive light, and this may be the manifestation of why men have been traditionally linked to the phenomenon of midlife crisis. Across the age ranges, approximately half the participants did not experience positive post-crisis growth – this is an important finding that counteracts findings from earlier qualitative research in which the overwhelming majority of crises were described as leading to positive outcomes (Robinson, Wright, & Smith, 2013). It may be that when participating in an interview about a crisis, there is a selection bias whereby mainly those with positive stories volunteer to tell their story at length (Bauer & McAdams, 2004). In contrast, a survey design such as the current study may gain a higher proportion of those with negative outcomes, as reflection on the events is far briefer. Speculatively, it may also be that the proportion of negative outcomes is also shaped by the cultural context of the sample – the redemptive master narrative is arguably less defining of the British cultural script than in the US (McAdams, 2013). It may be that if this research were to be replicated in the US, retrospective post-crisis growth would be more prevalent.

An additional finding was that empathy is related to the recall of crises. The relationship between crisis and empathy was found to be cumulative – those who have been through one or two crises are more empathic on average than those who have not, and those who report three or more are more empathic on average than those who have been through one or two. This finding suggests that a positive outcome of crisis is the capacity to understand and respond appropriately to others who are going through a difficult or distressing time in their own life. Experiences of having successfully navigated a personal crisis may act as important learning experience for individuals, which may in turn lead to better performance in roles or jobs where empathy is necessary. This is redolent of Jung’s notion of the ‘wounded healer’ (Jung, 1993). An alternative interpretation of the relationship between crisis and empathy is that more empathic individuals are more prone to crisis as they are sensitive to other people’s problems and distress and this can vicariously induce stressful life episodes and even crises in themselves.

Limitations and future directions for research

The study used a retrospective self-report survey method to assess crisis episodes in a person’s adult life as autobiographical episodes. Such a method provides insightful data on the demarcation of crisis episodes within the adult life story but has inherent limitations, for autobiographical memories are selective, constructive and prone to loss or fading over time (Pasupathi, 2001). Limitations also relate to the effect of a person’s current state on their memory recall – emotional state and personality traits may influence what is and what is not perceived to be a crisis in autobiographical memory (Bower, 1981). It is possible that certain individuals in the study experienced the same challenging life events but that only some reported them as a crisis, while others did not, due to dispositional differences or current emotional state differences. The method employed here meant that those who did not report a crisis did not report life events in that age range. Therefore it is not possible to establish from these data whether autobiographical crisis differs from periods of non-crisis in terms of the number of stressful events. This could be addressed in future studies by including a checklist of life events for past age ranges for completion by all participants, and then comparing the number and intensity of events reported by those who retrospectively report crises and those who don’t.

A further recall-related issue that could be assessed in future research is whether crises that are less recent in time are on average more severe than recent crises. Evidence shows that emotional intensity of events predicts recall (Talarico, LaBar, & Rubin, 2004), and it may therefore be that crisis intensity is measurably different in recent and not-recent crises. The presence of less severe crises in ratings of crisis within the previous ten years would help to explain the recency effect shown in our data. This could be explored in future research by way of rating crisis severity, and comparing crises that occurred at different elapsed time periods in the past.

The types of crisis events were assessed by way of participants endorsing items on a theoretically-derived list of events. The checklist of 23 items is not exhaustive, and can be increased or improved upon in future studies in order to gain data on more kinds of life events. For example, different types of bereavement were not distinguished in the current set, and this could be included in the future. Another more complex option for future research would be to request information on when events occurred within a crisis episode relative to each other, which could be used as evidence of the time flow of events within crisis. Another option for a future addition would be to assess the presence of emotions within crisis as well as events, in order to explore the affective contents of crisis.

In the current study, post-crisis growth was related to gender and age of crisis, but not to crisis event type. The different event contents of a crisis may be differentially predictive of growth and decline following a crisis, and the link between crisis type and growth/decline is an important avenue for future research. In order to conduct such an analysis, crisis type would ideally need to be assessed as discrete categories within a single nominal variable, rather than using the overlapping event-checklist approach that was taken here.

The recruitment method used in the study was one that employed both online and real-world strategies. The resulting sample has a slightly lower household income and percentage of married individuals than the UK population more generally, and has a higher proportion of whites than the UK population, but overall the sample is one that is geographically and demographically diverse and shows clear parallels to the composition of the UK more generally. However, it is possible that differences between age groups or genders are an artefact of sampling and online data collection, as there are biases in collecting data online that are endemic to that method, while response rates cannot be calculated as they can with a postal survey. In the future, research using a nationally representative sample and a postal method would be a strong step forward, as would research in different cultures and with ethnic minorities. It would also be of interest in future studies to compare geographical regions within the UK and other countries to explore the spatial distribution of crisis at a macro-level.

For reasons of parsimony and fit with our objectives, only crises between the ages of 20 and 50 are presented in this study. It will be important to gain data on crisis in later decades as well, in terms of prevalence, contents and post-crisis growth, to understand what occurs in the second decade of midlife and in later life. In final summary, there is a long way to go before a clear and comprehensive understanding of crisis episodes in adult life is gained in the literature, but this study is an important exploratory step on that path that we hope will be followed by other researchers.

References

Arnett

J. J.

(2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469–480.

Barr

(2004). Get it together: A guide to surviving your quarterlife crisis. London: Hodder & Stoughton Ltd.

Bauer

McAdams

D. P.

(2004). Personal growth in adults’ stories of life transitions. Journal of Personality, 72, 573–602.

Berntsen

Bohn

(2009). Cultural life scripts and individual life stories. In Boyer

Wertsch

(Eds.), Memory in mind and culture (pp. 62–82). Cambridge: Cambridge University Press.

Bernsten

Rubin

D. C.

(2002). Emotionally charged autobiographical memories across the life span: The recall of happy, sad, traumatic and involuntary memories. Psychology and Aging, 17, 636–652.

Bernsten

Rubin

D. C.

(2004). Cultural life scripts structure recall from autobiographical memory. Memory and Cognition, 32, 427–442.

Booker

(2005). The seven basic plots: Why we tell stories. London: Continuum.

Bower

G. H.

(1981). Mood and memory. American Psychologist, 36, 129–148.

Brim

O. G.

(1976). Theories of the male midlife crisis. The Counseling Psychologist, 6, 2–29.

10.

Clausen

J. A.

(1995). Gender, contexts, and turning points in adults’ lives. In Moen

P. E.

Elder

Luscher

(Eds.), Examining lives in context: Perspectives on the ecology of human development (pp. 365–389). Washington, DC: American Psychological Association.

11.

Conway

M. A.

Pleydell-Pearce

C. W.

(2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288.

12.

Denne

J. M.

Thompson

N. L.

(1991). The experience of transition to meaning and purpose in life. Journal of Phenomenological Psychology, 22, 109–133.

13.

de Vries

Watt

(1996). A lifetime of events: Age and gender variations in the life story. The International Journal of Aging and Human Development, 42, 81–102.

14.

Dindia

Allen

(1992). Sex differences in self-disclosure: a meta-analysis. Psychological Bulletin, 112, 106–24.

15.

Erikson

E. H.

(1968). Identity, youth and crisis. London: Faber & Faber.

16.

Erikson

E. H.

(1980). Identity and the life cycle. London: W. W. Norton & Co.

17.

Fivush

(2011). The development of autobiographical memory. Annual Review of Psychology, 62, 559–582.

18.

Forer

B. R.

(1963). The therapeutic value of crisis. Psychological Reports, 13, 275–281.

19.

Gaudette

Courter

(2011). How to survive your husband’s midlife crisis. London: Home & Leisure Publishing.

20.

Hammack

P. L.

Cohler

B. J.

(2011). Narrative, identity, and the politics of exclusion: Social change and the gay and lesbian life course. Sexuality Research and Social Policy, 8, 162–182.

21.

Hatch

S. L.

Dohrenwend

B. P.

(2007). Distribution of traumatic and other stressful life events by race/ethnicity, gender, SES and age: A review of the research. American Journal of Community Psychology, 40, 313–332.

22.

Hochschild

A. R.

Machung

(1990). The second shift: Working parents and the revolution at home. London: Piatkus.

23.

Jaques

(1965). Death and the midlife crisis. International Journal of Psychoanalysis, 46, 502–514.

24.

Jung

C. G.

(1993). The practice of psychotherapy. London: Routledge.

25.

Kegan

Lahey

L. L.

(2009). Immunity to change: How to overcome it and unlock the potential in yourself and your organization. Cambridge, MA: Harvard Business School Press.

26.

King

L. A.

(2001). The hard road to the good life: The happy, mature person. Journal of Humanistic Psychology, 41, 51–72.

27.

Kotre

(1995). White gloves: How we create ourselves through memory. New York: The Free Press.

28.

Lachman

M. E.

(2004). Development in midlife. Annual Review of Psychology, 55, 305–331.

29.

Levinson

D. J.

(1978). The seasons of a man’s life. New York: Ballantine Books.

30.

Levinson

D. J.

(1986). A conception of adult development. American Psychologist, 41, 3–13.

31.

Levinson

D. J.

(1996). The seasons of a woman’s life. New York: Ballantine Books.

32.

Masuda

Holmes

T. H.

(1978). Life events: Perceptions and frequencies. Psychosomatic Medicine, 40, 236–261.

33.

McAdams

D.P.

(1993). The stories we live by: Personal myths and the making of the self. New York: Guilford.

34.

McAdams

D. P.

(2001). The psychology of life stories. Review of General Psychology, 5, 100–122.

35.

McAdams

D. P.

(2013). The redemptive self: Stories Americans live by – revised and expanded edition. New York: Oxford University Press.

36.

McManus

Bebbington

(2009). Adult psychiatric morbidity in England, 2007: Results of a household survey. Leicester: National Centre for Social Research.

37.

Nolen-Hoeksema

(1987). Sex differences in unipolar depression: Evidence and theory. Psychological Bulletin, 101, 259–282.

38.

Norris

F. H.

(1992). Epidemiology of trauma: Frequency and impact of different potentially traumatic events on different demographic groups. Journal of Consulting and Clinical Psychology, 60, 409–418.

39.

O’Connor

D. J.

Wolfe

D. M.

(1987). On managing midlife transitions in career and family. Human Relations, 40, 12, 799–816.

40.

Pasupathi

(2001). The social construction of the personal past and its implications for adult development. Psychological Bulletin, 127, 651–672.

41.

Robbins

Wilner

(2001). Quarter-life crisis. London: Bloomsbury.

42.

Robinson

O. C.

(2008). Developmental crisis and the self in early adulthood: A composite qualitative analysis (Doctoral dissertation), Birkbeck College, London.

43.

Robinson

O. C.

(2011). Quarterlife crisis and emerging adulthood: Towards an understanding of their relation. Paper presented at the 5th Conference on Emerging Adulthood, Providence, RI.

44.

Robinson

O. C.

(2012). Development through adulthood: An integrative sourcebook. London: Palgrave Macmillan.

45.

Robinson

O. C.

Smith

J. A.

(2009). Metaphors and metamorphoses: Narratives of identity during times of crisis. In Robinson

Fisher

Yeadon-Lee

Robinson

S. J.

Woodcock

(Eds.) Narrative, memory and identities (pp. 85–94). Huddersfield: University of Huddersfield Press.

46.

Robinson

O. C.

Smith

J. A.

(2010a). A case study of pre-midlife crisis: An extraordinary transformational episode in an ordinary life. In Robinson

. (Eds.) Narrative, memory and ordinary lives (pp. 145–157). Huddersfield: University of Huddersfield Press.

47.

Robinson

O. C.

Smith

J. A.

(2010b). Investigating the form and dynamics of crisis episodes in early adulthood: The application of a composite qualitative method. Qualitative Research in Psychology, 7, 170–191.

48.

Robinson

O. C.

Smith

J. A.

(2010c). The stormy search for self in early adulthood: Developmental crisis and the dissolution of dysfunctional personae. The Humanistic Psychologist, 38, 120–145.

49.

Robinson

O. C.

Wright

G. R. T.

Smith

J. A.

(2013). The holistic phase model of early adult crisis. Journal of Adult Development.

50.

Rubin

D. C.

Schulkind

(1997). The distribution of autobiographical memories across the lifespan. Memory & Cognition, 25, 859–866.

51.

Schaie

K. W.

Willis

S. L.

(2002). Adult development and aging (5th edn). New York: Prentice Hall.

52.

Sheehy

(1977). Passages: Predictable crises of adult life. New York: Bantam Books.

53.

Spreng

R. N.

McKinnon

M. C.

Mar

R. A.

Levine

(2009).The Toronto Empathy Questionnaire: Scale development and initial validation of a factor-analytic solution to multiple empathy measures. Journal of Personality Assessment, 91, 62–71.

54.

Srivastava

John

O. P.

Gosling

S. D.

Potter

(2003). Development of personality in early and middle adulthood: Set like plaster or persistent change? Journal of Personality and Social Psychology, 84, 1041–1053.

55.

Super

D. E.

(1980). A life-span, life-space approach to career development. Journal of Vocational Behavior, 16, 282–298.

56.

Talarico

J. M.

LaBar

K. S.

Rubin

D. C.

(2004). Emotional intensity predicts autobiographical memory experience. Memory and Cognition, 32, 1118–1132.

57.

Tedeschi

R. G.

Calhoun

L. G.

(1995). Trauma and transformation: Growing in the aftermath of suffering. London: SAGE.

58.

Tedeschi

R. G.

Calhoun

L. G.

(2004). Posttraumatic growth: Conceptual foundations and empirical evidence. Psychological Inquiry, 15, 1–18.

59.

Tennant

Hiller

Fishwick

Platt

Joseph

Weich

Parkinson

Secker

Stewart-Brown

(2007). The Warwick-Edinburgh Mental Well-Being Scale (WEMWBS): Development and UK validation. Health and Quality of Life Outcomes, 5(63).

60.

Thurnher

(1983). Turning points and developmental change: Subjective and ‘objective’ assessments. Journal of Orthopsychiatry, 53, 52–60.

61.

Trusty

K.-M.

Watts

R. E.

(2005). Model of effects of adult attachment on emotional empathy of counseling students. Journal of Counseling & Development, 83, 66–77.

62.

Turner

R. J.

Lloyd

D. A.

(1995). Lifetime traumas and mental health: The significance of cumulative adversity. Journal of Health and Social Behavior, 36, 360–376.

63.

Turner

R. J.

Wheaton

Lloyd

D. A.

(1995). The epidemiology of social stress. American Sociological Review, 60, 104–125.

64.

Walker

W. R.

Vogl

R. J.

Thompson

C. P.

(1997). Autobiographical memory: Unpleasantness fades faster than pleasantness over time. Applied Cognitive Psychology, 11, 399–413.

65.

Wethington

(2000). Expecting stress: Americans and the ‘midlife crisis’. Motivation and Emotion, 24, 85–103.

66.

Wethington

Kessler

R. C.

Pixley

J. E.

(2004). Turning points in adulthood. In Brim

O. G.

Ryff

C. D.

Kessler

R. C.

(Eds.), How healthy are we? A national study of well-being at midlife (pp. 586–613). London: University of Chicago Press.