| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Aging Research Center, Karolinska Institute and Stockholm University, Sweden.
2 Vårdal Institute, Lund University, Sweden.
Address correspondence to Neda Agahi, Aging Research Center, Gävlegatan 16, SE-113 30 Stockholm, Sweden. E-Mail: neda.agahi{at}ki.se
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Methods. We used longitudinal data from three waves of an interview survey that followed 495 individuals from 1968 to 2002. Individuals were aged 43–65 in 1968 and 77–99 in 2002. We conducted logistic regression analyses on each of the leisure activities.
Results. For the panel followed, a decline in participation rates was the most common pattern over time. Analyses at the individual level showed that late-life participation was generally preceded by participation earlier in life. Previous participation, both 10 and 34 years earlier, predicted late-life participation. The modifying effect of functional status in late life was small.
Discussion. In accordance with the continuity theory of aging, leisure participation in old age is often a continuation of previous participation. However, there is considerable variation among both activities and individuals.
STUDIES have repeatedly found continuous patterns of participation in several domains of life, such as social relationships (Antonucci, 1990
), friendships (Field, 1999), activities (Singleton, Forbes, & Agwani, 1993), and identity (Troll & Skaff, 1997). However, very few studies have investigated patterns of leisure participation over an extended period of time. In this study we explored the influence of earlier leisure participation on late-life participation in different kinds of leisure activities over a period of three decades.
According to the continuity theory of aging (Atchley, 1989, 1999a, 1999b), individuals do not really change as they age; they just become "more" of what they have always been. Observed continuity and consistency as individuals age in patterns of thought, activities, and habits, despite considerable changes in health and social conditions, have given rise to the continuity theory. Continuity theory deals with processes of adaptation that individuals experience and implement as they age. Although aging does not necessarily lead to declines in health and function, it usually involves many transitions. Retirement, bereavement, and restrictions in the body's reserve capacity all affect the pursuits of daily life, social network, priorities, and time use. According to the continuity theory, maintaining patterns of thought, activities, and habits is the most common strategy for adaptation, although not the only one. Continuity over time does not necessarily entail successful aging; people who do not display continuous patterns may very well experience successful aging and vice versa. Nor does the continuity theory dispute the fact that aging is usually a period of great changes; rather, it makes a point of the coexistence of continuity and change and the various degrees of these (Atchley, 1999b).
The theory of selective optimization with compensation (Baltes & Baltes, 1990) also emphasizes continuity and adaptation. According to this theory, individuals use three strategies in order to maintain continuity in activities and habits as they get older. Selection refers to the choices and priorities individuals make when they must restrict their activity due to losses in function or resources. Optimization refers to engagement in activities that stimulate the body and mind in order to increase the reserve capacity needed to continue participation. Finally, the individual uses compensating strategies in order to continue engaging in the chosen activities despite reductions in or losses of capacities. Because of the extensive interindividual differences in late life, the application of these strategies varies greatly for different individuals.
Longitudinal age-related changes in leisure participation may be investigated by describing changes within a group (interindividual level) or by describing changes within individuals (intra-individual level) over time. On the interindividual level, declining participation patterns over time reflect age-related changes in the group; because individuals fluctuate (i.e., start or cease participation), declines in participation rates due to increasing age may be primarily visible on the group level. However, interindividual participation patterns may mask various intra-individual patterns, and Collins has therefore suggested that the most appropriate level for studying change over time is the intra-individual level (Collins, 1996).
Longitudinal studies of age-related changes in participation do not always distinguish between the two levels of analysis. On the group level, a general decrease in leisure participation with age is common. At the same time, researchers have also found continuous intra-individual participation patterns with increasing age (Armstrong & Morgan, 1998; Bijnen, Feskens, Caspersen, Mosterd, & Kromhout, 1998; Bukov, Maas, & Lampert, 2002; Strain, Grabusic, Searle, & Dunn, 2002; Verbrugge, Gruber-Baldini, & Fozard, 1996). A study investigating leisure activities among older individuals (aged 60–85 years) over an 8-year period found that most individuals did continue participation, although the likelihood of continuing participation decreased with higher age. The most common activities, such as watching television, reading, or shopping, were also the ones participants were most likely to continue (Strain et al., 2002). A study of social participation over a 4-year period among individuals aged 70 and older revealed that participation levels for the group as a whole remained relatively unchanged although individuals showed variation, with some individuals increasing and others decreasing participation (Bukov et al., 2002). Similarly, a study of physical activities over an 8-year period showed that for the whole group (aged 65 and older) general activity levels decreased, although there were individuals who increased the amount of time spent walking and/or shopping (Armstrong & Morgan, 1998). Another study of physical activity also reported decreasing overall activity levels among a group of men (aged 65 and older) over a 10-year period (Bijnen et al., 1998). In yet another study, with a follow-up time of up to 34 years, the results suggested that individuals displayed rather continuous participation patterns as they got older, although the variety of activities decreased (Verbrugge et al., 1996).
In addition, researchers have reported that factors such as gender, socioeconomic status, and living situation influence the kind and level of activities individuals engage in (Agahi & Parker, 2005; Bennett, 1998; Bukov et al., 2002; Cutler & Hendricks, 1990; Perren, Arber, & Davidson, 2003; Satariano, Haight, & Tager, 2002; Zaranek & Chapleski, 2005). However, in longitudinal studies of participation patterns over time, gender and socioeconomic status seem to be of less importance for whether individuals continue to participate (Bennett, 1998; Bukov et al., 2002; Strain et al., 2002), whereas worsening health or reduced income may be more important (Moody, 2002; Strain et al., 2002).
Studies have repeatedly reported the relation between leisure participation and health. This relationship is not a simple one, as causal links probably work both ways. Activities promote health, but at the same time participation in activities is dependent upon health factors. In a cross-sectional study, better self-reported health was associated with higher frequency of participation as well as participation in a greater number of activities (Searle & Iso-Ahola, 1988). This association was primarily found for physical leisure activities. Longitudinal studies suggest that changes in health and functional status may have greater effects on participation than health and functional status per se. For example, a worsening of health decreased the odds of continued attendance at theater, movies, and spectator sports; walking; shopping; gardening; and attendance at church services and activities over an 8-year period (Strain et al., 2002). In another study, individuals with new chronic conditions decreased their religious service attendance, whereas their use of religious media remained unaffected (Benjamins, Musick, Gold, & George, 2003). The results can be interpreted as support for the interaction and adaptation process suggested by the continuity theory as well as the theory of selective optimization with compensation. Thus, both previous participation and current health factors seem to influence current participation.
Most studies examining patterns in leisure participation with increasing age have a short period of follow-up time (usually less than 10 years). The present study followed individuals over 34 years in order to explore the extent to which late-life participation is affected by participation earlier in life. We also investigated the influence of functional status in late life and whether age, gender, and education affected participation patterns over time.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Professional interviewers conducted structured interviews with participants in their homes or institutions (referred to as direct interviews). Interviewers conducted some telephone interviews, mainly due to the unwillingness of a participant to receive someone in the home. In cases where the respondent could not be interviewed (mainly due to poor cognition), researchers interviewed a close relative or caregiver (referred to as indirect interviews). Descriptions of the LNU data collection and sample are available elsewhere (Erikson & Åberg, 1987). Further details about the SWEOLD in 1992, which was essentially the same as that in 2002, are also available elsewhere (Lundberg & Thorslund, 1996).
The LNU interviews included questions about working life, family relations, health, leisure participation, and other living conditions. The SWEOLD interviews covered these domains as well as asked additional questions about health and functional status. In order to maintain the longitudinal perspective, most questions have remained unchanged since the first survey in 1968.
Study Population
Survivors from the original sample in 1968 comprise the present study sample. We included in the sample those who had been interviewed in the LNU in 1968, the LNU in 1991 or the SWEOLD in 1992 (we combined these two data collections, allowing for a wide age span), and the SWEOLD in 2002. This left us with a panel of 495 individuals who had participated in all three waves of data collection. The nonresponse rate in 1968 was 11.3% for the relevant age groups (still alive in 2002), and attrition in the subsequent studies was 23.1%.
At baseline (1968), the individuals in the panel were 43–65 years old. In the combined LNU/SWEOLD of 1991/1992, they were aged 66–89; in the 2002 SWEOLD, 77–99.
Measures
Nine items concerning leisure activities were identical in all waves of data collection. We took eight of the nine leisure activities—reading books, hobby activities, gardening, cultural activities, fishing or hunting, restaurant visits, study circles and courses, and dancing—from a list of activities with the question "Do you usually engage in any of these activities?". Respondents could answer no; yes, sometimes; or yes, often. We dichotomized these into participation (yes, sometimes; and yes, often) and nonparticipation (no). The ninth activity, attending religious services, was measured separately with six possible answers: no, about once a year, a few times a year, about once a month, a few times a month, and once a week or more. We coded the first three answers as nonparticipation and the last three as participation.
Researchers measured mobility in 2002 as an index of the self-reported ability to stand without support, to rise from a chair, to walk 100 meters without difficulty, and to go up and down stairs without problems. We dichotomized this index into good (no limitations) and impaired (at least one limitation).
In 2002, researchers measured vision subjectively with the question "Can you read a newspaper without difficulties?". We dichotomized the answers into good vision (yes, with or without glasses) and impaired vision (no, have certain difficulties; and no, not at all).
Researchers measured cognition in 2002 with a validated short version of the Mini-Mental State Examination (MMSE; Parker, Gatz, & Thorslund, 1996). The range of the shortened scale was 0–18, dichotomized into good cognition (equal to or higher than the cut-off of 12) and impaired cognition (lower than 12). This cut-off is equivalent to the typical MMSE cut-off of 23 or lower for cognitive impairment (Folstein, Folstein, & McHugh, 1975; Gatz et al., 2005). Due to the inclusion of indirect and telephone interviews, some individuals lacked data on the MMSE. Interview notes explaining the reason for proxy interviews revealed that many missing cases were due to cognitive impairment. In order to minimize the number of internal missing, we created a separate category for individuals for whom we had no information about cognitive status. We therefore divided cognition into three categories: good cognition, impaired cognition, and no information on cognition.
We dichotomized the level of education into no more than grade school (corresponding to about 6–8 years of education) or beyond grade school.
We gathered information on age and gender from population registers and confirmed this during the interviews.
Statistical Analyses
We analyzed data by using SPSS 12.0.1 for Windows (SPSS, Chicago, IL). We conducted tests of significance for interindividual differences between survey waves by using chi-square tests.
We analyzed intra-individual data with logistic regressions in order to investigate factors related to participation in the different leisure activities in 2002. The independent variables were participation in 1968 and 1991/1992, mobility, vision, and cognition. In addition, we investigated whether participation patterns differed by gender, age, and education. We used McFadden pseudo R2 in order to describe the explained variance in the different models (DeMaris, 2002).
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
|
Model 3 added three measures of current functional status (mobility, vision, and cognition) to the regressions. Impaired mobility was significantly related to lower odds ratios for all activities except fishing or hunting and religious services. Impaired vision was significantly related to lower odds for reading books, hobby activities, and gardening. Impaired cognition was associated with lower odds for reading books, hobby activities, gardening, cultural activities, and restaurant visits. The addition of functional status into the model increased the explained variance for a few of the activities but in general had very little effect on the strength of the relationship between previous and current participation.
We know that gender, age, and education influence the rates and frequency of leisure participation. A secondary focus of this study was to investigate whether these factors also influenced the participation pattern over time (i.e., whether participation in 1968 and/or 1991/1992 predicted participation in 2002 differently with regard to gender, age, and education). We analyzed these variables in separate models (not shown). Men and women did not differ in their participation patterns over time, nor did the younger (aged 77–84 in 2002) and older (aged 85 and older in 2002) cohorts. We found differences in participation patterns between individuals with high and low educational levels for one activity: Gardening in 1968 was significantly related to gardening in 2002 (controlling for participation in 1991/1992) only among individuals with high educational levels (beyond grade school).
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
; Collins, 1996). According to these results, leisure participation in old age is often a continuation of participation earlier in life. We found that participation in 1968 and/or 1991/1992 predicted participation in 2002 for all of the nine leisure activities studied. The recurring predictions over both 34 and 10 years in the different activities indicate a pattern of continuous participation with increasing age. Overall, results from middle age into old age seem to support the theory of continuity.
When examining the results on the interindividual and intra-individual level for the different activities, it becomes evident that activities that show similar patterns on one level do not necessarily show the same patterns on the other level. For example, restaurant visits and religious services both had very stable rates of participation with increasing age on the interindividual level, whereas on the intra-individual level the odds ratio of previous participation predicting late-life participation was very much greater for religious services than for restaurant visits. We found that about half of those who attended religious services in late life did so both 10 and 34 years earlier, and a great majority of those who did not attend services in late life did not do so earlier in life either. In the case of religious services, corresponding patterns on the two levels may seem evident because one somehow expects religiousness to be stable in late life. However, a recent study showed that religious allegiance is declining among older individuals in Britain (Coleman, Ivani-Chalian, & Robinson, 2004). Restaurant visits show a different intra-individual pattern. Among those reporting restaurant visits in old age, there were equal percentages of reported and no reported previous restaurant visits both 10 and 34 years earlier (i.e., many people started eating out at restaurants in late life). One reason may be that there are many more restaurants today compared to a decade ago, and eating out at restaurants is much more common both in general and among more recent cohorts of elders (Agahi & Parker, 2005). Another explanation could be that many individuals find it more difficult to prepare food as they reach old age and begin to eat out more often.
Book reading was a popular activity and maintained its popularity over time, although the percentage of book readers declined as cohorts entered old age. Most people who read books in late life also read books 10 and 34 years earlier. Hobby activities, gardening, and cultural activities displayed patterns similar to those of book reading. Fishing or hunting is an activity where previous participation strongly predicted late-life participation but where participation rates showed a sharp decline with age on the group level. Patterns of previous participation among late-life fishermen or hunters and late-life book readers were very similar, although the two activities differed when it came to rates of participation and change over time for the group as a whole.
Dancing had yet another participation pattern. Intra-individual results for dancing suggest a turnover: Those who danced in middle age (1968) were not the same individuals as those who danced in old age (2002). About one third of those who danced in old age took it up in late life. It may be that the meaning of dancing changes with increasing age; at younger ages, dancing may entail going to nightclubs, whereas in late life it may be more of a physical activity or an organized social event. Participation in study circles was also an activity that individuals seemed to take up in late life. About 20% of the late-life study circle participants had never before reported participating in study circles.
To sum up, there were both differences and similarities between the different activities, although all of them displayed continuity over either a short or a longer period of time. Engagement in an activity earlier in life is a strong predictor of participation in late life. Book reading, gardening, and hobby activities are common activities that individuals often continue as they age. Functional status, such as impairments in mobility, vision, or cognition, had significant effects on late-life participation in all activities except fishing or hunting and attending religious services. However, these limitations did not have much influence on estimates in the regressions, suggesting that previous participation influences late-life participation irrespective of functional status in late life.
Gender, age, and education are all associated with rates of leisure participation among the oldest old (Agahi & Parker, 2005). They do not, however, make much difference for the participation pattern over time, which is consistent with previous research (Bennett, 1998; Bukov et al., 2002; Strain et al., 2002). Men and women display similar participation patterns with increasing age, as do the young old and the old old. Educational level influenced the participation pattern only for gardening; for individuals with high educational levels, but not low, gardening in 1968 predicted gardening in 2002.
Although previous participation predicts late-life participation, and many individuals display continuous participation patterns as they age, this study also found individuals who gave up or started new activities. Elderly people may start new activities after retirement or take advantage of activities organized by retirement organizations. They may give up activities due to poor functional status. In this study, poor functional status (measured as impaired mobility, vision, or cognition) was related to nonparticipation in most of the activities in old age. Loss of interest and environmental barriers are other potential reasons for decreased participation. Many studies have reported declining participation rates with increasing age on the group level; nevertheless on the individual level, a continuous pattern of participation is most common. In the Ohio Longitudinal Study of Aging and Adaptation there were also many different participation patterns over the 20-year study period, but the most prevalent was that of continuous participation (Atchley, 1999a). This variety in participation patterns does not necessarily contradict the continuity theory. The main proposition of the theory is that the most common adaptation process upon reaching old age is maintaining continuity in different life domains, but other kinds of adaptation processes may also occur. This variety of participation patterns reflects the rich diversity that can be found among older individuals.
As always in longitudinal studies, there are some issues that warrant caution. The individuals included in the sample of this study were in some ways a selected group; they had all survived to very old age and had also taken part in all three waves of data collection. This may have led to higher reported levels of participation because recurrent survey respondents can be assumed to be active in other domains of life as well. The high response rates in the LNU and SWEOLD surveys minimize this effect. Other longitudinal studies investigating leisure participation in older populations have found that those who remain in the study at follow-up are usually more active at baseline (Bukov et al., 2002). Although there may be an overestimation of the participation levels, it is not likely that the associations between previous and current participation would be very different. In a study by Bukov and colleagues (2002), correcting for selection did not influence the strong association found between previous and current participation.
An important issue when studying leisure participation over an extended period of time is the effect of societal changes. These changes may affect social roles, as well as opportunities for leisure participation, thereby influencing the individual's alternatives and choices (Cutler & Hendricks, 1990). Although our intention was to study age-related changes in leisure participation, this study cannot statistically separate age effects from period effects (Glenn, 1977).
The results found in this study confirm previous studies showing continuity and consistency in individuals' leisure participation as they age (Atchley, 1999a; Strain et al., 2002; Verbrugge et al., 1996). This points in the direction of the continuity theory of aging and its proposition that aging does not change the individual and his or her habits and activities. People seem to adjust their activities rather than changing them. A practical implication that can be drawn from these results is that in order to uphold an active older population, it may be more important to facilitate participation in activities that elderly cohorts engaged in earlier in life than to start new activities. Important exceptions are restaurant visits, dancing, and study circles; many people seem to be open to starting these activities even in late life regardless of previous participation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication November 3, 2005. Accepted for publication February 18, 2006.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  N. C. Franklin and C. A. Tate Lifestyle and Successful Aging: An Overview American Journal of Lifestyle Medicine, January 1, 2009; 3(1): 6 - 11. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
