Stability and Change of Well Being: An Experimentally Enhanced Latent State-Trait-Error Analysis

This study uses longitudinal panel data and short-term retest data from the same respondents in the German Socio-economic Panel to estimate the contribution of state and trait variance to the reliable variance in judgments of life satisfaction and domain satisfaction. The key finding is that state and trait variance contribute approximately equally to the reliable variance in well being measures. Most of the occasion specific variance is random measurement error, although occasion-specific variation in state variance makes a reliable contribution for some measures. Moreover, the study shows high similarity in life satisfaction and average domain satisfaction for the stable trait component (r = .97), indicating that these two measures are influenced by the same stable dispositions. In contrast, state variance of the two measures is distinct, although still highly correlated (r = .77). Error variances of the two measures are only weakly correlated, indicating that most of the error component is indeed due to random measurement error.     

Life Satisfaction Across the Lifespan: Findings from Two Nationally Representative Panel Studies

Two large-scale, nationally representative panel studies (the German Socio Economic Panel Study and the British Household Panel Study) were used to assess changes in life satisfaction over the lifespan. The cross-sectional and longitudinal features of these studies were used to isolate age-related changes from confounding factors including instrumentation effects and cohort effects. Although estimated satisfaction trajectories varied somewhat across studies, two consistent findings emerged. First, both studies show that life satisfaction does not decline over much of adulthood. Second, there is a steep decline in life satisfaction among those older than 70. The British data also showed a relatively large increase in satisfaction from the 40s to the early 70s. Thus, age differences in well-being can be quite large and deserve increased empirical and theoretical attention.     

Global Judgments of Subjective Well-Being: Situational Variability and Long-Term Stability

Subjective well-being (SWB) is an important indicator of quality oflife. SWB can be conceptualized as a momentary state (e.g., mood) aswell as a relatively stable trait (e.g., life satisfaction). Thevalidity of self-reported trait aspects of SWB has been questioned byexperimental studies showing that SWB judgments seem to be stronglycontext dependent. Particularly, momentary mood seems to have a stronginfluence on global SWB judgments. To explore the ecological validity ofthese conclusions a non-experimental longitudinal self-reportstudy with three occasions of measurement was conducted(N = 249). The associations between momentarymood ratings and global judgments of SWB (life satisfaction,satisfaction with life domains, frequency and intensity of emotions) aswell as personality ratings (self-esteem, optimism, neuroticism,extraversion) were analyzed in a multistate-multitrait-multiconstructmodel. This model takes (a) measurement error, (b) occasion-specificdeviations, and (c) stable interindividual differences into account. Itis shown that the variability in global SWB judgments and personalityratings is relatively small and much smaller than the variability inmood. Furthermore, the occasion-specific associations between moodstates, on the one hand, and global SWB and personality ratings, on theother hand, are relatively small and inconsistent. All global SWB andpersonality variables are more strongly related to mood on the traitlevel than on the occasion-specific deviation level. Therefore, incontrast to experimental studies, occasion-specific mood effects do notseem to be inherently important in ecological measurement settings.     

Panel Conditioning and Subjective Well-being

The importance of panel, or longitudinal, survey data for analyzing subjective wellbeing, and especially its dynamics, is increasingly recognized. Analyses of such data, however, have to deal with two potential problems: (1) non-random attrition; and (2) panel conditioning. The former is a much researched topic. In contrast, panel conditioning has received much less attention from the research community. In this analysis, longitudinal survey data collected from members of a large national probability sample of households are used to examine whether self-reported measures of psychological well-being exhibit any tendency to change over time in a way that might reflect panel conditioning. Regression models are estimated that control for all time invariant influences as well as a set of time-varying influences. We find very little evidence that mean life satisfaction scores vary with length of time in the panel, especially once non-random attrition is controlled for. In contrast, scores on a measure of mental health do vary with time, and surprisingly men and women exhibit opposing patterns. For men, scores decline over time (though the estimates are not statistically robust), whereas for women the effects are both large and rise with time. Further, for both outcome measures there is a clear narrowing in the dispersion of reported scores over the first few waves of participation. The findings have implications for empirical research employing longitudinal data.     

The Effect of Income on General Life Satisfaction and Dissatisfaction

Increasing evidence from the empirical economic and psychological literature suggests that positive and negative well-being are more than opposite ends of the same phenomenon. Two separate measures of the dependent variable may therefore be needed when analyzing the determinants of subjective well-being. We investigate asymmetries in the effect of income on subjective well-being with a single-item measure of general life satisfaction. Using data from the German Socio-Economic Panel 1984–2004, and a flexible multiple-index ordered probit panel data model with varying thresholds, we find that income has only a minor effect on high satisfaction but significantly reduces dissatisfaction.     

Validity and Reliability of Subjective Social Indicators: The effect of different measures of association

In an international research project, a meta analysis of multitrait-multimethod matrices was used to evaluate the effects of characteristics of survey measurement instruments on the quality of survey data. The analyses in this project were done on the basis of polychoric/polyserial correlations. The result of such a study can be used to correct for measurement errors if the correlation in the substantive study is also a polychoric/polyserial correlation. It is, however, doubtful if these quality estimates of survey measurement instruments are appropriate for studies based on the most frequently used Pearson correlations. In this paper, the general approach will be discussed first. After that, the results with respect to validity and reliability estimates, obtained using two different measures of association, will be compared: the polychoric/polyserial correlations and the Pearson correlations. This study suggests that the differences between estimates of the reliability for scales with different numbers of categories are quite large, while all other effects remain approximately the same. The reasons for the differences will also be given. All the results are based on experiments containing questions with respect to satisfaction with life and aspects of life. Thus the study provides estimates of validity and reliability on measures of satisfaction with life and aspects of life which can be used in practice for the prediction and optimization of data quality and for correction of the data with regard to measurement error.     

Reliability reporting practices in youth life satisfaction research

Due to the emergence of positive psychology in recent years, a growing line of research has focused on aspects of psychological wellness rather than psychopathology. Within the context of positive psychology, life satisfaction has emerged as a key variable of study in relation to adult and youth populations. Accurate measurement of life satisfaction and subsequent interpretation of results requires that the assessment devices yield strong psychometric properties. In terms of consistent measurement, reliability estimates must be calculated with each administration, as changes in sample characteristics may alter the scale’s ability to generate reliable scores. Therefore, the present study was designed to address reliability reporting practices for three youth life satisfaction measures. Seventy-three occasions of use were identified in the literature for the three scales examined in this study. Results suggest that greater than half of the life satisfaction studies calculated reliability coefficients from their own data. A discussion of reliability from classical test theory is provided, examples of correct reliability reporting practices are given, and implications for reporting reliability are discussed.     

Environmental Influences on Well-Being: A Dyadic Latent Panel Analysis of Spousal Similarity

This article uses dyadic latent panel analysis (DLPA) to examine environmental influences on well-being. DLPA requires longitudinal dyadic data. It decomposes the observed variance of both members of a dyad into a trait, state, and an error component. Furthermore, state variance is decomposed into initial and new state variance. Total observed similarity between members of a dyad is decomposed into trait similarity, initial state similarity, new state similarity, and error similarity. Dyadic similarity in new state variance reveals that both members of a dyad change in the same direction, which is a strong indication of environmental effects. DLPA is used to examine environmental influences on life satisfaction and domain satisfaction based on 22 annual assessments of married couples in the German Socio-Economic Panel Study (N’s = 607–740). The results show high similarity in new state variance for life satisfaction and objectively identical domains (household income, housing), and less similarity for objectively less similar domains (recreation, health). This finding provides strong evidence for environmental influences on well-being. In addition, the results show high trait similarity. The implications of the latter finding for interpretation of behavioral genetics studies of well-being are discussed.     

The affective and cognitive context of self-reported measures of subjective well-being

Researchers attempting to understand the experience of subjective well-being have relied heavily on self-report measurement. Recent research focused on this method has demonstrated that a number of factors, such as the current mood of the respondent and the cognitive and social context surrounding the response, can significantly influence response to items inquiring about global subjective well-being or satisfaction with life. In the present study, several measurement strategies (e.g., single-item measures, multiple-item scales, and memory search tasks) were compared with regard to their susceptibility to such influences. Although some evidence for effects due to item-placement or transient mood were found, all of the global measures of subjective well-being and life satisfaction has significant convergence with peer-reports, and the single-item measures showed good temporal reliability across a one-month interval. The data provide evidence for a significant degree of stability in subjective well-being and life satisfaction.     

A Hybrid Model for Research on Subjective Well-being: Examining Common- and Component-specific Sources of Variance in Life Satisfaction,Positive Affect,and Negative Affect

The primary components of subjective well-being (SWB) include life satisfaction (LS), positive affect (PA), and negative affect (NA). There is little consensus, however, concerning how these components form a model of SWB. In this paper, six longitudinal studies varying in demographic characteristics, length of time between assessment periods, number of assessment periods, and type of measures employed were used to test a hybrid model of SWB comprising both common and component-specific sources of variance. The majority of explained variance in LS was attributable to temporally consistent common variance, rather than component-specific sources. For PA and NA, both common and component-specific sources were substantive and consistent over time. Implications for the conceptualization and operationalizations of SWB are discussed.     

The stability and validity of quality of life measures

Effective social indicators must be stable when individual or societal characteristics are unchanged and dynamic when circumstances alter. Highly reliable measures may be poor indicators because they are insensitive to change. Little evidence is available on the sensitivity or validity of objective and subjective indicators. A lack of panel data has restricted the assessment of the stability of subjective measures. This paper examines longitudinal data on a representative sample of 2162 Canadians interviewed in 1977 and again in 1979. Test-retest correlations of approximately 0.50 were obtained for satisfaction and self-anchoring ladder measures among respondents who reported no significant changes in their lives during the past two years. Correlations were substantially lower, as expected, for those reporting life changes. Comparisons of the absolute values of these subjective indicators show that very little change in quality of life measures occurs when stable circumstances are reported but the indicators rise or fall significantly when situations change with downward adjustments being more dramatic than upward modifications. Positive and negative life events had little effect on overall evaluations of life quality. In general, these findings provide very strong evidence for the stability and validity of subjective indicators over time. These measures, with one exception, were constant in unchanging situations and sensitive to change when it occurred.     

The domains of life satisfaction: An attempt to order chaos

While life satisfaction is commonly measured as an aggregate of individual life domains, the characterisation of such domains is uncertain. This study attempts to group 173 different domains names derived from the literature under seven headings as used by the Comprehensive Quality of Life Scale (ComQol). It was found that 68% could be classified in this way. Moreover, due to the repeated use of some domain names, the ComQol classification included 83% of the total reported data. The ComQol domain data did not differ from single-item global measures of life satisfaction and the within-study variance was lower using the ComQol rather than the original domains. A hierarchy of domain satisfaction was found which was dominated by the domain of intimacy. The other ComQol domains were quite tightly clustered within a range of 1.08 standard deviations. No difference was found between normative data and data gathered from people with a chronic medical condition, but people selected on psychiatric criteria had a lower life quality, most particularly in the domain of intimacy. It is concluded that life satisfaction, and therefore subjective well-being, can be economically and validly measured through the seven ComQol domains.     

Inducing Score Reliability from Previous Reports: An Examination of Life Satisfaction Studies

The sample specific nature of reliability is often overlooked in substantive research, with investigators frequently referencing reliability estimates from previous studies as the basis for presuming the score integrity for their data. The purpose of the current study was to examine this practice, known as reliability induction, across studies using a popular life satisfaction (LS) measure, the Satisfaction with Life Scale. A comprehensive literature search was conducted, yielding a final sample of 196 studies using the instrument. Fifty-two studies involving 76 samples induced reliability coefficients from other sources. These studies were coded in terms of sample characteristics and variability estimates to derive comparisons in these two areas between the induced reports and the original studies. Results suggest that none of the studies inducing reliability had sufficient justification for induction. It was also found that estimating reliability coefficients for the data in hand has increased relative to induced reports and studies omitting reliability information. Practical implications for substantive researchers are discussed and suggestions for additional research are provided.     

Relationships between overall and life facet satisfaction: A multitrait-multimethod (MTMM) study

We extended research on the construct validity of overall and life facet satisfaction measures by (a) estimating relations among satisfaction constructs disattenuated for measurement error, and (b) controlling for spurious effects of common method variance, using confirmatory factor analysis (CFA) of multitrait-multimethod (MTMM) data. Results indicated strong support for convergent and discriminant validity, as well as for significant method effects. Results also indicated that corrected correlations among latent satisfaction variables were similar to their observed counterparts, suggesting that the attenuating effects of measurement error tend to balance spuriousness interjected by common method variance. Suggestions for future research include determining boundary variables which influence the direction of the relationship between overall and life facet satisfaction, and for identifying subgroups of individuals within which these relationships are homogeneous.     

Estimating Increment-Decrement Life Tables with Multiple Covariates from Panel Data: The Case of Active Life Expectancy*

A fundamental limitation of current multistate life table methodology-evident in recent estimates of active life expectancy for the elderly-is the inability to estimate tables from data on small longitudinal panels in the presence of multiple covariates (such as sex, race, and socioeconomic status). This paper presents an approach to such an estimation based on an isomorphism between the structure of the stochastic model underlying a conventional specification of the increment-decrement life table and that of Markov panel regression models for simple state spaces. We argue that Markov panel regression procedures can be used to provide smoothed or graduated group-specific estimates of transition probabilities that are more stable across short age intervals than those computed directly from sample data. We then join these estimates with increment-decrement life table methods to compute group-specific total, active, and dependent life expectancy estimates. To illustrate the methods, we describe an empirical application to the estimation of such life expectancies specific to sex, race, and education (years of school completed) for a longitudinal panel of elderly persons. We find that education extends both total life expectancy and active life expectancy. Education thus may serve as a powerful social protective mechanism delaying the onset of health problems at older ages.     

Importance Weighting in Client Satisfaction Measures: Lessons from the Life Satisfaction Literature

The linearisation approach to approximating variance of complex non-linear statistics is a well-established procedure. The basis of this approach is to reduce non-linear statistics to a linear form, justified on the basis of asymptotic properties of large populations and samples. For diverse cross-sectional measures of inequality such linearised forms are available, though the derivations involved can be complex. Replication methods based on repeated resampling of the parent sample provide an alternative approach to variance estimation of complex statistics from complex samples. These procedures can be computationally demanding but tend to be straightforward technically. Perhaps the simplest and the best established among these is the Jackknife Repeated Replication (JRR) method. Recently the JRR method has been shown to produce comparable variance for cross-sectional poverty measures (Verma and Betti in J Appl Stat 38(8):1549–1576, 2011); and it has also been extended to estimate the variance of longitudinal poverty measures for which Taylor approximation is not currently available, or at least cannot be easily derived. This paper extends the JRR methodology further to the estimation of variance of differences and averages of inequality measures. It illustrates the application of JRR methodology using data from four waves of the EU-SILC for Spain. For cross-sectional measures design effect can be decomposed into the effect of clustering and stratification, and that of weighting under both methodologies. For differences and averages of these poverty measures JRR method is applied to compute variance and three separate components of the design effect—effect of clustering and stratification, effect of weighting, and an additional effect due to correlation of different cross-sections from panel data—combining these the overall design effect can be estimated.     

The Link Between Political Participation and Life Satisfaction: A Three Wave Causal Analysis of the German SOEP Household Panel

Is there a relationship between political participation and individual life satisfaction? The idea that political participation makes people more satisfied with their lives has long been debated. However, the existing empirical research has not been very successful in demonstrating that such a relationship exists while some studies show that instead it is individual life satisfaction that impacts political participation. This paper aims to shed some light on the issue of causality between political participation and individual life satisfaction. Unlike former studies, we resort to panel data and apply a three wave model which allows us great flexibility to test several hypotheses. Also unlike previous studies, after correcting for measurement error, our analysis shows no compelling evidence of a causal relationship between political participation and life satisfaction.     

Some sources of error and their effect on census statistics

Often the reliability of survey data is examined only in relationship to sampling variances, excluding many other potential sources of error. If the sampling variance dominates the mean-square error, then few mistakes result by considering sampling variance only; however, if sampling variance is only a small part of the mean-square error, serious mistakes in inference could be made. The Bureau of the Census has developed a model describing the joint effect of sampling and nonsampling errors on census statistics. This article shows how a study of the components of error may lead to methods of improving the accuracy and reliability of survey data.     

The impact of life events and changes in domain satisfactions on well-being

Previous research has shown that QOL measures are reasonably reliable and valid for static analysis. This article is concerned with the validity of the measures for use in panel studies/dynamic analysis. Are they sensitive enough to register changes (i.e., enhanced satisfaction and well-being) of the kind one would expect to follow favourable life events, and also changes (i.e., reduced satisfaction and well-being) due to adverse life events? Also, do changes in satisfaction with specific domains of life produce concomitant changes in the general sense of well-being? These questions are addressed by means of a small panel study (N=184) in which respondents were interviewed first in November 1978 and a second time in January 1981. The questionnaire included a more detailed life events inventory (Dohrenwendet al., 1978) and one which contained a larger number of favourable events than inventories included in previous QOL panel studies (Campbellet al., 1976; Atkinson, 1982). Unlike previous researchers, we found some statistically significant relationships between aggregate life events scores and changes in well-being. Changes in relatively affective measures of well-being were greater than in relatively cognitive measures. We also found strong, predictable relationships between changes in domain satisfactions and changes in the general sense of well-being.     

A first glance into the black box of life satisfaction surrounding childbearing

The vast majority of studies looking into the relationship between childbearing and subjective well-being use overall measures where respondents either report their general level of happiness or their life satisfaction, leaving substantial doubt about the underlying mechanisms. However, life satisfaction and happiness are intuitively multidimensional concepts, simply because there cannot be only one aspect that affects individuals' well-being. In this study, by considering seventeen specific life satisfaction domains, these features come out very clearly. Whereas all the domains considered matter for the overall life satisfaction, only three of them, namely satisfaction with leisure, health and satisfaction with the partnership, change dramatically surrounding childbearing events. Even though we cannot generalise (since these results stem from one particular panel survey, i.e., Household Income and Labour Dynamics in Australia data), it appears that the typical anticipation and post-child decrease of life satisfaction, so often found in existing studies, stems from changes in these three domains.