Meaning in Life in Earlier and Later Older-Adulthood: Confirmatory Factor Analysis and Correlates of the Meaning in Life Questionnaire

Abstract

Objective: To validate the Meaning in Life Questionnaire (MLQ) in earlier and later older-adulthood, and examine its correlates. Method: Participants in earlier (n = 341, M age = 68.5) and later older-adulthood (n = 341, M age = 78.6) completed the MLQ and other measures. Confirmatory multigroup analysis, correlations, and regression models were conducted. Results: A two-factor (presence and search), eight-item model of the MLQ had a good fit and was age-invariant. Presence and search for meaning were largely unrelated. Meaning was associated with life satisfaction, well-being across a range of domains, and psychological resources. Searching for meaning correlated negatively with these variables, but to a lesser degree in later older-adulthood. Discussion: The MLQ is valid in older-adulthood. Meaning in life is psychologically adaptive in older-adulthood. Searching for meaning appears less important, especially in later older-adulthood. Findings are discussed in the context of aging and psychosocial development.

Keywords

meaning in life older adults factor analysis measurement invariance well-being

Meaning in life may be defined as the purposes for living that provide motivation for goal-directed activity (Feldman & Snyder, 2005). It is considered to be a crucial component in psychological well-being (Ryff & Singer, 1998), and is associated with a host of psychological variables such as increased life satisfaction (Park, Park, & Peterson, 2010; Steger & Kashdan, 2007), self-esteem (Morgan & Farsides, 2009), positive affect (King, Hicks, Krull, & Del Gaiso, 2006), and decreased psychopathology (Moomal, 1999) and death anxiety (Missler et al., 2012; Rappaport, Fossler, Bross, & Gilden, 1993). More recently, it has been suggested that meaning in life represents an integrative causal agent that might account for how and why health and well-being develop and change over time (McKnight & Kashdan, 2009). With a rapidly aging population in Australia (Australian Bureau of Statistics, 2012a), factors such as meaning in life that are implicated in the health and well-being of older adults are attracting increasing research attention. Indeed, older adults are recognized as a priority age group for health interventions by the Australian government (Australian Institute of Health and Welfare: Priority Age Groups, 2016). Given this, there is a need to ensure that meaning in life can be measured in a valid manner in the Australian older adult population.

A number of measures have been developed to assess the meaning in life construct. However, these have differed greatly in their content, dependent on the various conceptions of meaning in life that has been adopted. These measures have also tended to blend together seemingly distinct factors, referring to specific sources of meaning in life, achievements, satisfaction with daily tasks, and the novelty of experiences from day-to-day. For example, the Purpose in Life Test (Crumbauch & Maholick, 1969), developed from Viktor Frankl’s (1985) logotherapy orientation, refers to, among other aspects, having the freedom to make choices in life and viewing oneself as being responsible. Alternatively, items on the Personal Meaning Index (Reker, 1992) reflect, in part, having a consistent understanding of one’s self.

Eschewing the conceptual issues above, The Meaning in Life Questionnaire (MLQ; Steger, Frazier, Oishi, & Kaler, 2006) is a recently developed 10-item self-report tool that offers a parsimonious measurement of meaning in life. The MLQ has two subscales: a Presence subscale, which measures how strongly individuals agree they currently have meaning in their life; and a Search subscale, which measures how strongly they agree that they are currently searching for meaning in their life. These two dimensions of meaning have been found to be only moderately related, with the search for meaning in one’s life not merely a manifestation of a lack of meaning (Steger et al., 2006). The nature of a meaningful life may vary greatly between individuals. The MLQ items account for individual differences in the interpretation of meaning by presenting simple statements whose referent content can be broadly interpreted, for example, “My life has a clear sense of purpose.” The MLQ has been widely adopted, with the brevity of the measure and its reliable factor structure being clear strengths (Steger et al., 2006). It has been found to measure meaning in life as an empirically distinct psychological construct, with superior discriminant validity, relative to other popular measures of meaning in life, with a range of well-being variables (Steger et al., 2006). The psychometric properties of the MLQ have also been found to be consistently strong across various cultural and ethnic groups, including Chinese (Chan, 2014), South African (Temane, Khumalo, & Wissing, 2014), North American (Steger et al., 2006), and Japanese (Steger, Kawabata, Shimai, & Otake, 2008) people.

Despite the wide use of the MLQ, very few studies, and to the authors’ knowledge none in the Australian context, have examined its psychometric validity in older adults. In addition to its strong psychometric properties, the ease with which the MLQ items can be interpreted suggests it may be well suited to older demographics. Furthermore, the subjective way in which meaning can be construed when responding to the items may help to mitigate any cohort effects relating to views on what constitutes meaning in life. To date, the factor structure of the MLQ has been examined only once in older adults, in a small sample aged 65 years and above (Steger, Oishi, & Kashdan, 2009). This sample did not appear to contain reasonable numbers of adults in later stages of older–adulthood, and therefore, it is unknown whether the factor structure of the MLQ would remain consistent for these older adults.

A further motivation of this study is to clarify whether the correlates of meaning in life vary across older-adulthood. Entering later stages of older-adulthood signals the approach of end of life, and may usher in different psychosocial goals. Erikson (1980), in his seminal theory of psychosocial development, suggested that as individuals move through later stages of adulthood, their psychosocial objectives shift from generativity and achievement, to contemplating and integrating the experiences across their life and the meaning they have brought. As adults become older and move through different phases of later adulthood, they may take on different social roles, and also experience changes in their health, level of dependency, and self-concept. This may change the psychological function of meaning in life, as well as the sources from which meaning is derived. Currently, little is known about whether the function and sources of meaning in life change or stay consistent over the course of older-adulthood. Such information is relevant to the provision of health services and the promotion of well-being among Australian older adults.

The Present Study

The present study had two, related aims. The first was to assess the factorial structure of the MLQ among two groups of older adults and to determine whether the measurement is invariant across these two age groups. The second aim was to examine correlates of the presence and search for meaning in life among Australian older adults, and whether these differed between earlier and later stages of older-adulthood. A range of variables previously found to be associated with meaning in life were used for this purpose, including life satisfaction (Zika & Chamberlain, 1992), psychological resources such as self-esteem (Morgan & Farsides, 2009) and optimism (Ho, Cheung, & Cheung, 2010), personality traits (Schnell & Becker, 2006), and income (Steger & Samman, 2012). We were also interested in which of these variables might uniquely contribute to variance in the presence or search for meaning in life in older adults. In accordance with the theory presented above, it was predicted that searching for meaning would be less strongly related to well-being and psychological resources in later older-adulthood, relative to earlier older-adulthood, given the decline in the importance of generativity and achievement that is proposed to normatively accompany older age (Erikson, 1980).

Method

Participants

Participants were recruited as part of the 2011 longitudinal follow-up survey of the Australian Unity Wellbeing Index (AUWI). The cross-sectional AUWI is an annual survey that monitors the subjective well-being of the Australian population. Each survey sample is representative of the geographic distribution and gender composition of the Australian population. The present sample comprised a follow-up of participants from previous surveys, with measures relevant to the present study added to the normal battery of measures. Of the 3,000 questionnaires mailed out, 1,653 were returned completed. Given this attrition from the original sample, the derived sample is no longer systematically stratified by location. The representativeness of the gender ratio has been retained.

We adopted the definition of older-adulthood used by the Australian Bureau of Statistics (2012b), and only participants aged 65 years or more who provided valid response sets were retained for analysis (n = 682). The sample was then median split by age, creating a group of adults in earlier older-adulthood (n = 341, M age = 68.5, SD = 2.3, range = 65-73) and later older-adulthood (n = 341, M age = 78.6, SD = 4.5, range = 74-92). It is noted that the age range of the older group is relatively larger, and that this may bias correlation-based statistics. However, the characteristics of the two groups are shown in Tables 1 and 2, where it can be seen that the variance for all measured variables is very similar for both groups, with no systematic bias due to age.

Table 1.

Characteristics of the Earlier and Later Older-Adulthood Groups.

Characteristic	Earlier older-adulthood (n = 341)	Later older-adulthood (n = 341)
Age	68.5 (2.3)	78.6 (4.5)
Percentage females	50.4%	50.1%
Living with others	71.6%	59.8%
Relationship
Never married	4.4%	5%
Separated not divorced	1.8%	1.8%
Married	64.8%	53.7%
Divorced	10.9%	7.9%
De facto or living together	4.1%	1.5%
Widowed	14.1%	30.2%
Work/study
Full-time paid employment	4.3%	1.5%
Full-time home or family care	1.2%	2.1%
Full-time retired	78.8%	89.1%
Full-time study	0.3%	0.3%
Semi-retired	14.4%	5%
Unemployed	0.3%	1.5%
Full-time volunteer	0.6%	0.6%

Table 2.

Means and Standard Deviations for Earlier and Later Older-Adulthood Groups.

Variable	Earlier older-adulthood M (SD)	Later older-adulthood M (SD)	Effect size
Meaning in life—Presence	7.2 (1.9)	7.2 (2.0)
Meaning in life—Search	3.9 (2.8)	4.2 (2.7)^†	d = .10
Global life satisfaction	8.1 (1.2)	8.1 (1.2)
Overall PWI	7.8 (1.1)	7.8 (1.2)
PWI—standard of living	8.3 (1.3)	8.3 (1.3)
PWI—health	7.1 (1.9)	7.2 (1.8)
PWI—achievement in life	7.5 (1.6)	7.3 (1.8)
PWI—personal relationships	8.2 (1.7)	8.2 (1.5)
PWI—safety	8.2 (1.4)	8.2 (1.4)
PWI—community	7.7 (1.7)	7.7 (1.7)
PWI—future security	7.7 (1.7)	7.8 (1.5)
Self-esteem	8.3 (1.2)*	8.0 (1.2)*	d = .25
Optimism	7.6 (1.5)	7.6 (1.5)
Perceived control	7.5 (1.5)	7.3 (1.5)
Neuroticism	3.1 (1.7)	3.1 (1.5)
Extraversion	6.1 (1.8)*	5.8 (1.7)*	d = .17
Income before tax	1.9 (1.1)**	1.7 (1.1)**	d = .18

Note. The possible scores on all variables range from 0 to 10, except for income before tax (1-8). PWI = Personal Well-Being Index.

†

p < .10. *p < .05. **p < .01.

Materials

Meaning in life

The MLQ is a 10-item self-report measure consisting of two, five-item subscales (Steger et al., 2006). The Presence subscale assesses the extent to which respondents feel their lives are currently meaningful, and the Search subscale assesses the extent to which they are searching for meaning. Initial (Steger et al., 2006) and subsequent (Steger et al., 2009) validations of the MLQ have provided evidence of its good psychometric properties, and that it measures a construct distinct from other well-being variables. In this study, participants responded to the MLQ items using an 11-point, end-defined scale ranging from 0 (absolutely untrue) to 10 (absolutely true). This response format was used due to the improved scale sensitivity that it offers relative to a 7-point point format adopted in the original scale (Cummins & Gullone, 2000). One item on the Presence subscale is worded negatively, and was reverse-coded prior to analyses. The responses to items on each subscale were averaged together, with higher scores reflective of a stronger agreement that one has meaning in life or that one is searching for meaning in life. Previous studies have reported Cronbach’s alphas ranging from α = .82 to α = .86 (Steger et al., 2006). In the present study, the internal reliabilities for the total sample were good (Presence subscale α = .86, Search subscale α = .92).

Global life satisfaction (GLS)

GLS was measured using the single item, “Thinking about your own life and personal circumstances, how satisfied are you with your life as a whole?” Participants responded using an 11-point, end-defined scale ranging from 0 (completely dissatisfied) to 10 (completely satisfied).

Subjective well-being

This was measured using the Personal Well-Being Index (PWI; International Wellbeing Group, 2006). The PWI consists of seven items that measure satisfaction with standard of living, health, achievement, relationships, safety, community connectedness and future security. As well as being of interest in isolation, these items can be averaged to form an overall measure of subjective well-being. Participants responded using an 11-point, end-defined scale ranging from 0 (completely dissatisfied) to 10 (completely satisfied). Previous studies have shown these items and the total scale to have good psychometric properties in the Australian population (Cummins, Eckersley, Pallant, Van Vugt, & Misajon, 2003). In the present study, the internal reliability was good (α = .85).

Self-esteem

A short-form, five-item version of the Rosenberg Self-Esteem Scale (Rosenberg, 1965), using only the positively worded items, was used to assess self-esteem. This short-form version retains the measurement fidelity of the longer version (Gray-Little, Williams, & Hancock, 1997) and has good psychometric properties (Hallford, Mellor, & Cummins, 2013). Participants responded using an 11-point, end-defined scale ranging from 0 (do not agree at all) to 10 (agree completely). The items were averaged together, with higher scores indicating higher levels of perceived self-worth. In the present study, the internal reliability of the items was good (α = .91).

Optimism

Optimism was assessed using a shortened version of The Life Orientation Test–Revised (Scheier & Carver, 2003), which consists of the three positively worded items. Participants responded using an 11-point, end-defined scale ranging from 0 (do not agree at all) to 10 (agree completely). The items responses were averaged together, with higher scores indicating a more positive future orientation. Previous research has reported a Cronbach’s α = .85 for this three-item shortened version of the scale (Lai & Cummins, 2013). In the present study, the internal reliability was good (α = .86).

Perceived control

To assess perceived control, that is, the perceived ability to change the environment to achieve desired outcomes (Folkman, 1984), a six-item scale using questions developed by Holloway (2003) was used. Participants responded using an 11-point, end-defined scale ranging from 0 (do not agree at all) to 10 (agree completely). The responses to the six items were averaged together, with higher scores reflecting a stronger perception of control over problems in life. This scale has good psychometric properties (Hollway, 2003), with previous research reporting a Cronbach’s α = .82 (Lai & Cummins, 2013). In the present study, the internal reliability was good (α = .86).

Personality

The neuroticism and extraversion items from the 10-Item Personality Inventory (Gosling, Rentfrow, & Swann, 2003) were used to measures these respective personality traits. This widely used, ultrabrief personality assessment has good psychometric properties (Gosling et al., 2003), and both scales showed acceptable internal reliability in the present study (neuroticism α = .67, extraversion α = .74).

Income

Participants were asked to indicate their household’s annual income before tax using one item, with response options ranging from 1 (less than AU$15,000) to 8 (more than AU$500,000).

Procedure

Ethics approval was granted from the university human research ethics committee prior to commencement of the study (DUHREC2006-266). As noted above, AUWI respondents who were agreeable to receiving follow-up surveys were mailed a package consisting of a plain language statement detailing the aims of the present study, as well as the paper-based questionnaire. Consent was implied by return of the completed questionnaire in a reply paid envelope. No incentives were offered for participation.

Statistical Analyses

All data were cleaned and analyzed using SPSS 22.0 and AMOS 22.0. Descriptive statistics were generated to report the demographic characteristics of the groups and means and standard deviations of all the study variables. A multivariate analysis of variance (MANOVA) was conducted to assess group mean differences on the study variables, with follow-up t tests and effect sizes reported as Cohen’s d, whereby .2, .5, and .8 represent small, medium, and large effects, respectively (Cohen, 1992). Cronbach’s alphas were used to assess the internal reliability of scales. Pearson correlations were used to assess associations between the MLQ and continuous study variables. Fischer’s z-transformation approach was used to assess whether any differences between the two groups in the magnitude of correlation coefficients were significant. Multiple linear regressions, using the Presence and Search subscales as the dependent variable, were conducted for each group to assess which variables might predict unique variance in the presence or search for meaning. Given the large number of correlations, and therefore inflated risk of a type 1 error, the alpha level was adjusted to .01 for univariate and multivariate analyses.

All confirmatory factor analyses were conducted through structural equation modeling with maximum likelihood estimation. A range of indices were used to assess how well the data fit the proposed model. These were the chi-square value and corresponding p value, the relative chi-square statistic (CMIN/DF), the root mean square error of approximation (RMSEA), the standardized root mean square residual (SRMR), and the comparative fit index (CFI). Widely adopted guidelines to gauge the fits of a model to the data are RMSEA ≤ .06, SRMR ≤ .09, and a CFI ≥ .95 (Hu & Bentler, 1999). However, these guidelines do not represent cutoff scores for decision making, but rather a broad benchmark of how well, or poorly, a model fits to the data (Marsh, Hau, & Wen, 2004).

To establish factorial invariance across the older adult groups, multigroup confirmatory factor analyses were conducted (Byrne, 2004). Baseline models of the MLQ were first estimated in each group to ascertain appropriate factor structures. Factorial invariance was then assessed by conducting multigroup CFA with no constraints on parameters. This model was then compared with progressively more constrained models (nested within less restricted models) in which factor loadings were held equal across the groups, then intercepts, then variances of the latent variables, and finally the item residuals. A measure is deemed to be invariant across the groups if the fit of the constrained models are not significantly worse than the fit of the unconstrained model. In the present study, a difference exceeding .01 on the CFI was adopted as the criteria for difference in the model fit, as this goodness-of-fit index is independent of model complexity and sample size in assessing model invariance (Cheung & Rensvold, 2002).

Results

Table 2 presents the means and standard deviations for the study variables. A multivariate effect was found for differences between the groups, F(16, 665) = 2.4, p = .002, with follow-up t tests indicating there were small differences between the groups on self-esteem, extraversion, and income. As a check on data integrity, the values for GLS and PWI are slightly higher than population norms (Cummins et al., 2013), as is normal for older groups in Australia.

Factorial Validity and Measurement Invariance

Baseline models were first estimated in each group. Tests of the original factor structure of the MLQ in the earlier older-adulthood group indicated that the model did not fit the data particularly well (see Table 3, for indices of model fit). The reverse-coded item on the Presence subscale (“My life has no clear purpose”) had a significantly (p < .05) lower factor loading onto the latent variable (.54) relative to the other items on this subscale (.73-.93). The modification indices also strongly indicated an improved model fit if the error term of this item was co-varied with items on the Search subscale (modification indices: 77-85). Given the relative poor loading and apparent lack of discriminant validity, this item was removed and the model was retested. This resulted in a significantly improved model fit, Δ χ2 (Δ df = 8) = 104.7, p < .001. Inspection of the factor loadings indicated that the last item on the Search subscale (“I am searching for meaning in life”) had a significantly (p < .05) lower factor loading onto the latent variable (.71) relative to the other items on this subscale (.85-.93). Furthermore, modification indices indicated its error term to be co-varied with the Presence subscale (modification index: 22). This item was removed, and the retested model showed significantly improved fit, Δ χ2 (Δ df = 7) = 32, p < .001. The fit indices for this final model indicated a good fit to the data, and all items loaded strongly onto their respective latent variables (see Figure 1). Although the RMSEA for this baseline model did exceed the recommended cutoff, recent simulation research indicates that this may occur in otherwise good-fitting models due to a small number of degrees of freedom (Kenny, Kaniskan, & McCoach, 2015).

Table 3.

Fit Indices for Baseline Models of the MLQ for the Earlier and Later Older-Adulthood Groups.

	χ²	df	χ²/df	p value	CFI	RMSEA	SRMR
Earlier older-adulthood
Original model	242.2	34	7.1	<.001	.92	.13 (.12-.15)	.10
Reverse-coded item removed	137.5	26	5.3	<.001	.95	.11 (.09-.13)	.06
Final model	105.5	19	5.5	<.001	.96	.11 (.09-.14)	.04
Later older-adulthood
Original model	251.3	34	7.4	<.001	.91	.14 (.12-.15)	.12
Reverse-coded item removed	155.7	26	5.9	<.001	.95	.12 (.10-.14)	.07
Final model	110.9	19	5.8	<.001	.96	.12 (.09-.14)	.03

Note. MLQ = Meaning in Life Questionnaire; CFI = Comparative Fit Index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual.

Figure 1.

Confirmatory factor analysis of eight-item MLQ in the earlier older-adulthood group.

The original factor structure of the MLQ in the later older-adulthood group was also not a good fit to the data. Changes to the baseline model mirrored those in the earlier adulthood group. The reverse-coded item on the Presence subscale had a significantly (p < .05) lower factor loading onto the latent variable (.39) relative to the other items on this subscale (.80-.89), and modification indices strongly indicated this item shared variance with the Search subscale latent variable (modification index: 42). Removal of this item and retesting of the model resulted in a significantly improved model fit, Δ χ2 (Δ df = 8) = 95.6, p < .001. The last item on the Search subscale also had a significantly (p < .05) lower factor loading onto the latent variable (.64) relative to the other items on this subscale (.84-.89), and modification indices indicated its error term to be co-varied with the Presence subscale (modification indices: 7-36). This item was also removed and the resulting model showed significantly improved fit, Δ χ2 (Δ df = 7) = 44.8, p < .001. For this final model, the fit indices indicated a good fit to the data, and all items loaded strongly onto the respective latent variables (see Figure 2). Consistent with the earlier older-adulthood baseline model, the RMSEA exceeded recommend cutoffs (Kenny et al., 2015). All subsequent analyses were conducted using the eight-item version of the MLQ, as per the findings of the baseline CFAs. The internal reliability of the subscales was high, with alphas of .87 and .93 for the Presence and Search subscales in the earlier older-adulthood group, respectively, and .86 and .92 for the later older-adulthood group. As seen in Figures 1 and 2, the Presence and Search subscales had a weak negative correlation in the early older-adulthood group, and a weak positive correlation in the later older-adulthood group.

Figure 2.

Confirmatory factor analysis of eight-item MLQ in the later older-adulthood group.

Multigroup comparisons were then conducted to assess for measurement invariance across the older adult groups. Table 4 presents each of the five models, which progressively restrained more estimates. As indicated, none of the constrained models differed from the unconstrained model, as assessed by change in the CFI. This included the last model with residuals constrained, a strict, conservative test of invariance. Therefore, these findings supported full invariance of the MLQ for adults in earlier and later older-adulthood.

Table 4.

Fit Indices for Multigroup Comparisons.

	χ²	df	χ²/df	p value	CFI	RMSEA [90% CI]	SRMR
Unconstrained model	216.5	38	5.7	<.001	.96	.08 [.07, .09]	.04
Factor loadings constrained	229.0	44	5.2	<.001	.96	.08 [.07, .09]	.04
Intercepts constrained	240.9	52	4.6	<.001	.96	.07 [.06, .08]	.04
Structural covariance constrained	250.0	55	4.5	<.001	.96	.07 [.06, .08]	.06
Residuals constrained	278.4	63	4.4	<.001	.96	.07 [.06, .08]	.06

Note. Each row represents a different model for multigroup comparison, with descending rows becoming progressively more constrained in their estimates. CFI = Comparative Fit Index; RMSEA = root mean square error of approximation; CI = confidence interval; SRMR = standardized root mean square residual.

Correlates of the MLQ Across Earlier and Later Older-Adulthood

Tables 5 and 6 show the bivariate correlations between the MLQ subscales and the other study variables in both older adult groups. The presence of meaning in life was significantly associated with all study variables in both groups, with the exception of income. In both groups, higher meaning in life was related to higher levels of GLS, personal well-being (as an overall score and on individual dimensions), self-esteem, optimism, perceived control, and the personality trait of extraversion. Neuroticism was related to lower meaning in life, and this relationship was significantly stronger for the earlier adulthood group compared with the later older-adulthood group. With the exception of neuroticism in the later older-adulthood group, these correlations were generally moderate in strength.

Table 5.

Correlations Between the MLQ Presence of Meaning Subscale and Other Study Variables.

Variable	Earlier older-adulthood	Later older-adulthood
Global life satisfaction	.41***	.40***
Overall personal well-being	.43***	.43***
PWI—standard of living	.26***	.33***
PWI—health	.25***	.32***
PWI—achievement in life	.44***	.39***
PWI—personal relationships	.28***	.30***
PWI—safety	.25***	.29***
PWI—community	.36***	.34***
PWI—future security	.30***	.30***
Self-esteem	.48***	.45***
Optimism	.55***	.49***
Perceived control	.42***	.42***
Neuroticism	−.29***^a	−.13***^a
Extraversion	.35***	.26***
Income before tax	−.12	.00

Note. MLQ = Meaning in Life Questionnaire; PWI = Personal Well-Being Index.

Correlations differ at the p < .01 level.

***

p < .001.

Table 6.

Correlations Between the MLQ Search for Meaning Subscale and Other Study Variables.

Variable	Earlier older-adulthood	Later older-adulthood
Global life satisfaction	−.18***	−.08
Overall personal well-being	−.24***	−.08
PWI—standard of living	−.18***	−.04
PWI—health	−.14**	−.01
PWI—achievement in life	−.20***	−.15**
PWI—personal relationships	−.21***	−.06
PWI—safety	−.15**	−.03
PWI—community	−.12	−.10
PWI—future security	−.18**	−.04
Self-esteem	−.20***	−.08
Optimism	−.14	.02
Perceived control	.01	.07
Neuroticism	.29**	.32***
Extraversion	−.18***	.00
Income before tax	−.12	.04

p < .01. ***p < .001.

In the earlier older-adulthood group, the Search for Meaning subscale was negatively correlated with GLS, overall personal well-being, and each of its dimensions (with the exception of community connectedness), self-esteem, and neuroticism, and positively correlated with extraversion. These correlations were generally of a small magnitude. In the later older-adulthood group, searching for meaning in life was only associated with lower perceived achievement in life and higher levels of neuroticism.

For the multivariate analyses, only variables that had significant univariate associations with the subscales were entered into the regression models. The overall PWI was excluded from multivariate analyses due to its conceptual and empirical overlap (a bivariate correlation in this study of r = .74) with GLS, and risk of causing unstable coefficient estimates in the regression models. Inspection of scatterplots indicated that the predictors were linearly related to the MLQ subscales in both groups. Inspection of histograms indicated that all regression variables (except for the personality traits which were normally distributed) had a mild negative skew. Given that regression analyses in large samples (more than n = 100) are known to be robust in the face of mild degrees of nonnormality (Lumley, Diehr, Emerson, & Chen, 2002), no data transformations were conducted. Durbin–Watson statistics fell between 1.5 and 2.5, indicative of independence of observations. With respect to multicollinearity, the variance inflation factors (VIF) indicated that variables were only weakly-to-moderately related in regression models in the early (VIF range = 1.3-2.6) and later (VIF range = 1.2-2.6) older-adulthood groups. The results (shown in Table 7) show that only optimism and perceived control predicted unique variance of the presence of meaning in life in both older adult groups. For Search for Meaning, only neuroticism predicted unique variance in both older adult groups. In combination, the variables predicted around a third of the variance in the Presence subscale in both groups, and around 10% in the Search for Meaning subscale in both groups.

Table 7.

Summary of Regression Analyses of MLQ Subscales With Other Study Variables (all n = 341).

Predictor	Presence subscale						Search subscale
	Earlier older-adulthood			Later older-adulthood			Earlier older-adulthood			Later older-adulthood
	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β
Global life satisfaction	.10	.11	.06	.15	.12	.10	.11	.18	.05
PWI—standard of living	−.08	.09	−.05	.00	.10	.00	−.21	.16	−.10
PWI—health	−.01	.06	−.01	.03	.07	.02	−.01	.09	−.01
PWI—achievement in life	.19	.09	.15	.12	.34	.10	−.07	.14	−.04	−.07	.09	−.04
PWI—personal relationships	−.05	.07	−.04	−.01	.08	−.01	−.19	.11	−.11
PWI—safety	−.10	.09	−.07	−.01	.10	−.01	.05	.14	.02
PWI—community	.08	.07	.07	.05	.07	.04
PWI—future security	.08	.08	.07	−.06	.10	−.05	−.12	.13	−.07
Self-esteem	.08	.12	.05	.20	.12	.11	−.09	.18	.04
Optimism	.36	.09	.29***	.37	.09	.27***
Perceived control	.18	.07	.13**	.23	.07	.18***
Neuroticism	−.03	.06	−.03	.13	.07	.10	.40	.10	.24***	.54	.10	.30***
Extraversion	.13	.05	.12	.04	.06	.03	−.16	.09	−.14
F			15.4***			12.9***			4.7***			19.2***
R²			.38			.34			.13			.10
Adjusted R²			.36			.31			.10			.10

Note. MLQ = Meaning in Life Questionnaire; PWI = Personal Well-Being Index.

p < .01. ***p < .001.

Discussion

This study aimed to validate the factor structure of the MLQ among Australian older adults, to test for measurement invariance among adults in earlier and later stages of older-adulthood, and to examine the correlates of the presence and search for meaning in life in these two groups.

In both older adult groups, the initial model fit of the MLQ was poorer than previously reported for this age demographic (Steger et al., 2009). However, trimming one poorer-fitting, item from each of the MLQ subscales improved the overall fit of the model in both groups. Given the large conceptual overlap in item content, and otherwise good psychometric properties, four items per subscale was deemed sufficient for measurement, in accordance with the advice of (Raubenheimer, 2004). This eight-item MLQ had measurement invariance in earlier and later stages of older-adulthood, indicating that any differences in means or correlates of the MLQ were not an artifact of the items being answered in different ways across the groups.

Somewhat unexpectedly, the Presence and Search for Meaning subscales were only very weakly correlated with one another, and in different directions for each older adult group. Steger et al. (2009) have previously reported inverse correlations between the Presence and Search subscales, with the magnitude of this correlation increasing as a function of increasing age. They reported a significant negative correlation of r = −.44 between these subscales among older adults, indicating that the search for meaning in life increased as the reported presence of meaning decreased. In contrast to this, Chan (2016) has reported a strong significant positive correlation between the Presence and Search subscale in a sample of Chinese older adults living in residential care homes (r = .40). He suggests that this correlation may be due to the interaction of cultural and contextual factors, in that Asian cultures view the search for meaning as a responsibility and that stress related to institutionalization may make the centrality of this duty even more salient. Results from the present study suggest, to the contrary, that these factors are largely independent among this sample of community-living, older adults in Australia, and that they may seek meaning in life regardless of their current sense of meaning and purpose. As discussed earlier, individuals are likely to experience significant changes in various domains, such as social, occupational, and vocational, as they enter older-adulthood. This may prompt the search for new and varied sources of meaning that are independent of existing meaning they already have in their lives. For example, as older adults retire from the workforce (as many in this sample had), they may retain the sense of meaning derived from a history of work experience and achievements, but seek further occupational pursuits from which to derive new meaning and purpose. Notwithstanding this hypothesis, future research may help clarify the disparity in findings on the correlation of the MLQ subscales. In terms of absolute scores on these scales, both groups of older adults tended to agree that they had meaning in their life, and tended slightly toward disagreeing that they were searching for meaning.

In both groups of Australian older adults, the MLQ subscales correlated with the other study variables in directions that indicated convergent and discriminant validity. Meaning in life was related to higher satisfaction with life overall, and higher well-being in relation to standard of living, health, achieving in life, personal relationships, safety, community connectedness, and future security. Meaning in life was only moderately correlated with achieving in life, indicating that older adults derive only part of their meaning from satisfaction with having achieved goals. This is consistent with the notion that meaning may be derived from numerous sources, including those not be amenable to being “achieved” in a concrete sense, such as living in accordance with one’s values. Meaning in life was also related to more adaptive psychological resources, namely, viewing oneself favorably and as possessing intrinsic worth, having more positive expectations about the future, and a stronger conviction regarding perceived control. Optimism and perceived control were found to contribute unique variance to meaning, independent of increased satisfaction with life, well-being, or self-esteem. Thus, although meaning is associated with identifying that life and self are good and worthwhile, it appears to be uniquely associated with positive expectancies for the future and a sense of ability to control the environment and overcome challenges. These latter two styles of thinking that are likely to indicate agency and adaptivity in the context of maintaining health and well-being.

Consistent with previous findings (Steger et al., 2006), more meaning in life was associated with being more extraverted, and individuals with more meaning were less likely to experience neuroticism. For neuroticism, this relationship was stronger for adults in earlier older-adulthood relative to those in later older-adulthood. This finding is consistent with the notion that emotional instability tends to reduce as individuals age and may play a less influential role in psychological functioning (Soto, John, Gosling, & Potter, 2011).

For the search for meaning, differences between the earlier and later stages of older-adulthood were more apparent. Adults in earlier older-adulthood who agreed more strongly they were searching for meaning in life reported less life satisfaction and well-being (for all domains with the exception of community connectedness), and lower self-worth. Among adults in later older-adulthood, only achievement in life was inversely related to searching for meaning. Search for meaning in life therefore appears to play more of a role in well-being and psychological resources in earlier stages of older-adulthood relative to later stages. The idea of searching for meaning may suggest a lack of meaning in life, and potentially lower satisfaction with one’s life. In this sample, this was true to a small extent for adults in earlier older-adulthood, but only true with respect to achievement in life for those in later older-adulthood. These findings are congruent with predictions from Erikson’s (1980) theory of psychosocial development, in that the well-being and self-worth of individuals in later stages of older-adulthood may be less dependent on the active search for further purpose, but rather a consolidation of their current meaning and reflection on the experiences across their lifetime. Personality-wise, searching for meaning was associated with the tendency to be more neurotic in both groups, and less extraverted in the earlier older-adulthood group.

Contrary to previous findings in a Chilean sample (Steger & Samman, 2012), reported income was not related to either MLQ subscale. This indicates that the amount of money that Australian older adults receive as income does not directly influence their perception of having meaning or searching for more meaning.

Implications and Limitations

The primary implications of this study are twofold. First, the reduced version of the MLQ appears to be a psychometrically acceptable measure for use among older Australian adults. It is also invariant in its measurement of meaning and search for meaning among people in earlier and later stages of older-adulthood. This increases the confidence of using the MLQ in samples of very old people, and facilitates meaningful comparisons across older age groups. Second, and consistently across earlier and later stages of older-adulthood, those who identify as having meaning and purpose in their life report greater life satisfaction and well-being, and possess stronger psychological resources.

This latter relationship is notable, as psychological resources may act as buffers in the context of stressors, and are prophylactic for mental (Taylor & Stanton, 2007) and physical illness (Taylor, Kemeny, Reed, Bower, & Gruenewald, 2000). For example, optimism is related to better psychological adjustment in times of difficulty and more proactive positive health behaviors (Carver, Scheier, & Segerstrom, 2010), while self-esteem is a prospective factor in greater happiness and less depression (Baumeister, Campbell, Krueger, & Vohs, 2003; Orth, Robins, & Widaman, 2012). In contrast, older adults who do not identify meaning are more likely to feel poorly about their life and about themselves, be pessimistic, and have a reduced sense of being able to cope with difficulties.

The significant shifts in roles, self-concept, and living context that can occur during older-adulthood may necessitate recalibration of the sources of meaning, and may represent potential challenges to this task. For example, challenges faced might include loss of independence, the death of a partner and age-related peers, retirement from work, and reduced physical functioning and capacity for occupational engagement. Indeed, meta-analytic findings have shown a small age-related decline in purpose in life, which is stronger in older age (Pinquart, 2002). In particular, older adults who move to residential facilities and care homes may struggle to maintain sources of meaning in these new and potentially restrictive environments. Ensuring what they do is crucial for maintaining good mental and physical health (McKnight & Kashdan, 2009).

Strengthening the perception that one has lived a full and productive life, referred to by Erikson (1980) as achieving ego integrity, is also likely to be an effective means through which to increase meaning. Indeed, reminiscence-based interventions that focus on recalling and reflecting on past experiences are one effective method of achieving this (Pinquart & Forstmeier, 2012). While searching for meaning in life was not highly related to personal well-being or psychological resources in older adults, the fact that this was largely orthogonal to presence of meaning suggests that newer sources of meaning in older-adulthood, in addition to those already identified, may be adaptive. This highlights the need for opportunities for older adults to create new meaning in their lives, for example, through novel and stimulating social or occupational engagement.

In regard to limitations, the cross-sectional nature of this study does not facilitate elucidation of causal relationships between the variables studied. However, it seems possible that at least some of these relationships would be causally bidirectional. For example, higher perceived control is likely to lead to overcoming challenges and achieving meaningful goals, thus increasing meaning in life, while the sense of purpose may also drive individuals to approach challenges and cope directly with them. The present sample, although large, was not randomly sampled or known to be demographically representative of the Australian population. Future studies may establish the validity of the findings in this context, and whether the findings are generalizable to older adults of other nationalities. Future studies may also further explore the correlates of meaning among older adults by including measures not used in this study (e.g., measures of social contact or physical activity). Furthermore, sampling older adults from specific contexts, such as residential facilities, may clarify whether these environments act as moderators for the correlates of meaning in life.

Conclusion

Meaning in life is a construct that has significant relevance as people enter old age. A greater understanding of its measurement and association with other aspects of life is important for those providing services to this growing population. The MLQ with the modifications reported here is a suitable reliable and valid measurement instrument for both the presence and search for meaning in life.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

D. J. Hallford is a clinical psychologist and research fellow engaging in clinical and research work across the adult life span. He has published more than 15 refereed articles with a focus on aging-related issues. His present research focuses on autobiographical memory and mental health.

D. Mellor is a professor of psychology at Deakin University and has published widely in aging-related areas including depression, dementia, respect for the elderly, and systemic issues in aged-care provision.

R. A. Cummins is Emeritus professor of psychology at Deakin University and has published more than 200 books, book chapters, articles, and reports. His special area of interest is research into quality of life, and he is regarded as an international authority in this field.

M. P. McCabe is head of the Institute for Health and Ageing and has published more than 400 refereed articles. She conducts theoretical and applied research in the areas of body image disorders, aging, depression, sexual health, and the mental health and well-being of Indigenous people across the life span.

References

Australian Bureau of Statistics. (2012a). Population projections. Retrieved from http://www.ausstats.abs.gov.au/ausstats/subscriber.nsf/0/13D196FB0DBECC3BCA257C2E00173FAD/$File/32220_2012%20%28base%29%20to%202101.pdf

Australian Bureau of Statistics. (2012b). Who are Australia’s older people? Reflecting a nation: Stories from the 2011 Census. Retrieved from http://www.abs.gov.au/ausstats/abs@.nsf/Lookup/2071.0main+features602012-2013

Australian Institute of Health and Welfare: Priority Age Groups. (2016). Retrieved from http://www.aihw.gov.au/priority-age-groups/

Baumeister

R. F.

Campbell

J. D.

Krueger

J. I.

Vohs

K. D.

(2003). Does high self-esteem cause better performance, interpersonal success, happiness, or healthier lifestyles? Psychological Science in the Public Interest, 4, 1-44. doi:10.1111/1529-1006.01431

Byrne

B. M.

(2004). Testing for multigroup invariance using AMOS graphics: A road less traveled. Structural Equation Modeling, 11, 272-300. doi:10.1207/s15328007sem1102_8

Carver

C. S.

Scheier

M. F.

Segerstrom

S. C.

(2010). Optimism. Clinical Psychology Review, 30, 879-889. doi:10.1016/j.cpr.2010.01.006

Chan

W. C. H.

(2014). Factor structure of the Chinese Version of the Meaning in Life Questionnaire among Hong Kong Chinese caregivers. Health & Social Work, 39, 135-143. doi:10.1093/hsw/hlu025

Chan

W. C. H.

(2016). Assessing meaning in life in social work practice: Validation of the Meaning in Life Questionnaire among clinical samples. British Journal of Social Work, 47, 9-27. doi:10.1093/bjsw/bcv144

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255. doi:10.1207/s15328007sem0902_5

10.

Cohen

(1992). A power primer. Psychological Bulletin, 112, 155-159. doi:10.1037//0033-2909.112.1.155

11.

Crumbauch

Maholick

L. T.

(1969). Manual of instructions for the purpose in life test. Munster, Ireland: Psychometric Affiliates.

12.

Cummins

R. A.

Eckersley

Pallant

Van Vugt

Misajon

(2003). Developing a national index of subjective wellbeing: The Australian Unity Wellbeing Index. Social Indicators Research, 64, 159-190. doi:10.1023/A:1024704320683

13.

Cummins

R. A.

Gullone

(2000, March). Why we should not use 5-point Likert Scales: The case for subjective quality of life measurement. In Paper presented at the proceedings of the second international conference on quality of life in cities (pp. 74-93). Singapore: National University of Singapore.

14.

Cummins

R. A.

Woerner

Weinberg

Collard

Hartley-Clark

Horfiniak

(2013). Australian Unity Wellbeing Index: -Report 30.0-The Wellbeing of Australians: Social media, personal achievement, and work. Melbourne: Australian Centre on Quality of Life, School of Psychology, Deakin University. Retrieved from http://www.deakin.edu.au/research/acqol/reports/auwbi.php

15.

Erikson

E. H.

(1980). Identity and the life cycle (Vol. 1). New York, New York: W.W. Norton.

16.

Feldman

D. B.

Snyder

C. R.

(2005). Hope and the meaningful life: Theoretical and empirical associations between goal-directed thinking and life meaning. Journal of Social & Clinical Psychology, 24, 401-421. doi:10.1521/jscp.24.3.401.65616

17.

Folkman

(1984). Personal control and stress and coping processes: A theoretical analysis. Journal of Personality and Social Psychology, 46, 839-852. doi:10.1037//0022-3514.46.4.839

18.

Frankl

V. E.

(1984). Man’s search for meaning. New York, NY: Washington Square Press/Pocket Books.

19.

Gosling

S. D.

Rentfrow

P. J.

Swann

W. B.

(2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37, 504-528. doi:10.1016/s0092-6566(03)00046-1

20.

Gray-Little

Williams

V. S.

Hancock

T. D.

(1997). An item response theory analysis of the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 23, 443-451. doi:10.1177/0146167297235001

21.

Hallford

Mellor

Cummins

(2013). Adaptive autobiographical memory in younger and older adults: The indirect association of integrative and instrumental reminiscence with depressive symptoms. Memory, 21, 444-457.

22.

M. Y.

Cheung

F. M.

Cheung

S. F.

(2010). The role of meaning in life and optimism in promoting well-being. Personality and Individual Differences, 48, 658-663. doi:10.1016/j.paid.2010.01.008

23.

Hollway

(2003). The relationship of perceived control to neighbourhood wellbeing (Unpublished honours thesis). Deakin University, Burwood, Victoria, Australia.

24.

L.-t.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55. doi:10.1080/10705519909540118

25.

International Wellbeing Group. (2006). Personal wellbeing index: Adult, manual (4th ed.). Melbourne: Australian Centre on Quality of Life, Deakin University. Retrieved from http://www.deakin.edu.au/research/acqol/instruments/wellbeing_index.htm

26.

Kenny

D. A.

Kaniskan

McCoach

D. B.

(2015). The performance of RMSEA in models with small degrees of freedom. Sociological Methods & Research, 44, 486-507. doi:10.1177/0049124114543236

27.

King

L. A.

Hicks

J. A.

Krull

J. L.

Del Gaiso

A. K.

(2006). Positive affect and the experience of meaning in life. Journal of Personality and Social Psychology, 90, 179-196. doi:10.1037/0022-3514.90.1.179

28.

Lai

L. C.

Cummins

R. A.

(2013). The contribution of job and partner satisfaction to the homeostatic defense of subjective wellbeing. Social Indicators Research, 111, 203-217. doi:10.1007/s11205-011-9991-6

29.

Lumley

Diehr

Emerson

Chen

(2002). The importance of the normality assumption in large public health data sets. Annual Review of Public Health, 23, 151-169.

30.

Marsh

H. W.

Hau

K.-T.

Wen

(2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling, 11, 320-341. doi:10.1207/s15328007sem1103_2

31.

McKnight

P. E.

Kashdan

T. B.

(2009). Purpose in life as a system that creates and sustains health and well-being: An integrative, testable theory. Review of General Psychology, 13, 242-251. doi:10.1037/a0017152

32.

Missler

Stroebe

Geurtsen

Mastenbroek

Chmoun

Van Der Houwen

(2012). Exploring death anxiety among elderly people: A literature review and empirical investigation. Omega: Journal of Death and Dying, 64, 357-379. doi:10.2190/om.64.4.e

33.

Moomal

(1999). The relationship between meaning in life and mental well-being. South African Journal of Psychology, 29, 36-41. doi:10.1177/008124639902900106

34.

Morgan

Farsides

(2009). Psychometric evaluation of the meaningful life measure. Journal of Happiness Studies, 10, 351-366. doi:10.1007/s10902-008-9093-6

35.

Orth

Robins

R. W.

Widaman

K. F.

(2012). Life-span development of self-esteem and its effects on important life outcomes. Journal of Personality and Social Psychology, 102, 1271-1288. doi:10.1037/a0025558

36.

Park

Peterson

(2010). When is the search for meaning related to life satisfaction? Applied Psychology: Health and Well-Being, 2, 1-13. doi:10.1111/j.1758-0854.2009.01024.x

37.

Pinquart

(2002). Creating and maintaining purpose in life in old age: A meta-analysis. Ageing International, 27, 90-114. doi:10.1007/s12126-002-1004-2

38.

Pinquart

Forstmeier

(2012). Effects of reminiscence interventions on psychosocial outcomes: A meta-analysis. Aging & Mental Health, 16, 541-558.

39.

Rappaport

Fossler

R. J.

Bross

L. S.

Gilden

(1993). Future time, death anxiety, and life purpose among older adults. Death Studies, 17, 369-379. doi:10.1080/07481189308252631

40.

Raubenheimer

(2004). An item selection procedure to maximize scale reliability and validity. SA Journal of Industrial Psychology, 30(4), 59-64. doi:10.4102/sajip.v30i4.168

41.

Rosenberg

(1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

42.

Reker

(1992). Manual of the life attitude profile-revised. Peterborough, Ontario, Canada: Student Psychologists.

43.

Ryff

C. D.

Singer

(1998). The contours of positive human health. Psychological Inquiry, 9, 1-28. doi:10.1207/s15327965pli0901_1

44.

Scheier

Carver

(2003). Optimism. Encyclopedia of Stress, 3, 99-102.

45.

Schnell

Becker

(2006). Personality and meaning in life. Personality and Individual Differences, 41, 117-129. doi:10.1016/j.paid.2005.11.030

46.

Soto

C. J.

John

O. P.

Gosling

S. D.

Potter

(2011). Age differences in personality traits from 10 to 65: Big Five domains and facets in a large cross-sectional sample. Journal of Personality and Social Psychology, 100, 330-348. doi:10.1037/a0021717

47.

Steger

M. F.

Frazier

Oishi

Kaler

(2006). The Meaning in Life Questionnaire: Assessing the presence of and search for meaning in life. Journal of Counseling Psychology, 53, 80-93. doi:10.1037/0022-0167.53.1.80

48.

Steger

M. F.

Kashdan

T. B.

(2007). Stability and specificity of meaning in life and life satisfaction over one year. Journal of Happiness Studies, 8, 161-179. doi:10.1007/s10902-006-9011-8

49.

Steger

M. F.

Kawabata

Shimai

Otake

(2008). The meaningful life in Japan and the United States: Levels and correlates of meaning in life. Journal of Research in Personality, 42, 660-678. doi:10.1016/j.jrp.2007.09.003

50.

Steger

M. F.

Oishi

Kashdan

T. B.

(2009). Meaning in life across the life span: Levels and correlates of meaning in life from emerging adulthood to older adulthood. The Journal of Positive Psychology, 4, 43-52. doi:10.1080/17439760802303127

51.

Steger

M. F.

Samman

(2012). Assessing meaning in life on an international scale: Psychometric evidence for the Meaning in Life Questionnaire-Short Form among Chilean households. International Journal of Wellbeing, 2, 182-195. doi:10.5502/ijw.v2i.i3.2

52.

Taylor

S. E.

Kemeny

M. E.

Reed

G. M.

Bower

J. E.

Gruenewald

T. L.

(2000). Psychological resources, positive illusions, and health. American Psychologist, 55, 99-109. doi:10.1037//0003-066x.55.1.99

53.

Taylor

S. E.

Stanton

A. L.

(2007). Coping resources, coping processes, and mental health. Annual Review of Clinical Psychology, 3, 377-401. doi:10.1146/annurev.clinpsy.3.022806.091520

54.

Temane

Khumalo

I. P.

Wissing

M. P.

(2014). Validation of the Meaning in Life Questionnaire in a South African context. Journal of Psychology in Africa, 24, 51-60. doi:10.1080/14330237.2014.904088

55.

Zika

Chamberlain

(1992). On the relation between meaning in life and psychological well-being. British Journal of Psychology, 83, 133-145. doi:10.1111/j.2044-8295.1992.tb02429.x