Which of the following is NOT one of the recent findings in the study of life satisfaction

An essential step before examining the influence of positive factors on mental health and mental illness and promoting mental health is to investigate the constructs themselves, for example their convergent or discriminant validity, stability over time and longitudinal measurement invariance (LMI). A construct has to be shown reliable by at least satisfying psychometric properties. The assessment of constructs is mostly based on rating scales, whose scores are added up to an overall score. Differences in sum-scores over time are often taken to reflect changes in the underlying construct. LMI is a useful tool for interpreting mean differences meaningfully across different measurement points. Simply put, LMI indicates that people ascribe the same meaning to items regardless of time. If LMI of the construct is not given, this means that people change their idea of the construct and rate an item differently over time. Although mean differences of scales are generally considered meaningful, for most scales longitudinal invariance has not been tested yet. LMI is tested standardly in three steps (Vandenberg & Lance, 2000): Configural invariance implies that the structure of the scale is equal across time. Weak measurement invariance refers to equal item loadings across time and implies that structural relationships between latent variables can be meaningfully compared at different measurement occasions. Strong measurement invariance refers to equal intercepts across time. It enables the comparison of relationships between variables, but also comparisons of latent means on a single construct over time (Hirschfeld and von Brachel, 2014, Meredith, 1993). Thus, strong invariance of a construct is the precondition for meaningful comparison of means across time. Another step in testing measurement invariance focuses on the level of residual invariance. If the condition of strict invariance is met, all variations across time are due only to differences in the constructs. However, strict invariance is difficult to establish and not needed for practical applications of scales and is therefore beyond the scope of the present paper.

By establishing weak LMI for all constructs, it is possible to examine the interplay and causal relationships between the constructs over time. In the following section, we describe the constructs of happiness, life satisfaction and positive mental health that are core constructs in the research of mental health. Then, we will review findings regarding their LMI and finally in terms of their longitudinal interplay.

In this study we followed the definition of happiness by Bradburn, 1969, Lyubomirsky et al., 2005 who define happiness as the experience of more frequent positive affective states than negative ones. Argyle, Martin, and Crossland (1989) expand the frequency and degree of the predominance of positive affect by considering the dimension of an average level of satisfaction over a specific period of time. Happiness is a hedonic feeling characterized by moderate levels of arousal and is often used interchangeably with joy (de Rivera et al., 1989, Lazarus, 1991). Thus, happiness is predominately affect-based but different from positive affect (Lü et al., 2014, Singh and Jha, 2008). Lyubomirsky et al. (2005) found that happiness is positively associated with employment and quality of work, income, social relationships and mental and somatic health outcomes. Specifically, people with high levels on happiness are more relaxed, can regulate their emotions better and are more capable of facing and accepting problems (Yiengprugsawan, Somboonsook, Seubsman, & Sleigh, 2012). Happiness correlates negatively with anxiety, depression and rumination (Eldeleklioglu, 2015, Iani et al., 2014). Regarding stability over time, happiness is seen as a state construct with a higher temporal instability, whereas the construct satisfaction of life is considered to be more stable (Veenhoven, 1994). Kaczmarek, Bujacz and Eid (2015) found happiness not to be more occasion-specific compared to life-satisfaction. However, the items measuring happiness in their study were not solely affect-based, a cognitive component was included.

Life satisfaction is defined here as an overall “conscious cognitive judgement of one‘s life in which the criteria for judgment are up to the person” (Pavot & Diener, 1993, p. 164). Life satisfaction has an affective dimension that, however, is not identical with positive affect or happiness. Hence, life satisfaction should not be conceived as an overarching or superordinate concept of overall mental health but rather as one important aspect of positive mental health (Kjell, Daukantaite, Hefferon, & Sikstrom, 2016). Life satisfaction is positively associated with gratitude, social support, self-efficacy, continuous planning and consideration of future consequences (Azizli et al., 2015, Kong et al., 2015) and negatively associated with psychological distress, depression and anxiety (Beutel et al., 2016, Marum et al., 2014). It has been shown to be an indicator and predictor of functioning and presence of clinical symptoms and comorbidity in college students (Renshaw & Cohen, 2014). Regarding stability, between 30% and 50% of the between-person variation in life satisfaction is attributable to a stable trait factor (Lucas and Donnellan, 2007, Lucas and Donnellan, 2012, Schimmack et al., 2010, Schimmack and Lucas, 2010). Eid and Diener (2004) likewise found that 12–16% of the variance in satisfaction with life was due to occasion-specific influences, whereas 74–83% of the variance was determined by stable trait differences. However, life satisfaction is not as stable as personality factors. Anusic and Schimmack (2016) found that changing life factors had a greater influence on life satisfaction than on personality factors. This finding underlines the affective component of life satisfaction. However, in their study happiness was as stable as life satisfaction. Furthermore, happiness and life satisfaction are often linked to markedly different causes and consequences (Schimmack, Schupp, & Wagner, 2008). For example, life events such as bereavement or unemployment have a stronger effect on life satisfaction than on happiness, whereas childbirth has a stronger effect on happiness (Luhmann, Hofmann, Eid, & Lucas, 2012).

Positive mental health, as defined here, includes emotional and psychological components of well-being and indicates positive functioning (Lukat, Margraf, Lutz, van der Veld, & Becker, 2016). While happiness and life satisfaction are often subsumed under the term subjective well-being or hedonic well-being, the psychological component of positive mental health refers to eudaimonic concepts such as self-acceptance or meaning in life. Positive mental health is often used interchangeably with satisfaction of life or quality of life. However, it is a multidimensional construct and can be regarded as a component of quality of life including constructs of happiness and life satisfaction (Brown, Bowling, and Flynn, 2004). Positive mental health is often tested by the tripartite model proposed by Keyes (2007), which includes emotional, psychological and social aspects of well-being. However, the stability of the construct has been rarely investigated. Positive mental health and personality traits are closely related (Steel, Schmidt, & Shultz, 2008). Indeed, the stability of positive mental health in middle adulthood is high (rtt = 0.84; Kokko, Korkalainen, Lyyra, & Feldt, 2013). Similarly, it was high for a nine-month interval rtt = 0.65–0.68 (Lamers, Westerhof, Bohlmeijer, ten Klooster, & Keyes, 2011) and for a one-week and four-week interval (rtt = 0.81 and rtt = 0.74; Lukat et al., 2016).

Most studies investigate the constructs and levels of happiness, life satisfaction and positive mental health separately and cross-sectionally. Therefore, longitudinal studies focusing on the temporal stability, sequence and interplay of these constructs are needed. Furthermore, the relationships between these variables over time have not been investigated so far.

Due to high cross-sectional covariations, it is likely that positive factors covary strongly across time (Steger, Kashdan, & Oishi, 2008). Thus, increases or decreases in one positive construct may lead to subsequent increases or decreases in another positive construct. For example, higher levels of positive affect predict greater life satisfaction and positive mental health (Coffey et al., 2015, Datu and King, 2016). Lyubomirsky et al. (2005) found that happiness precedes numerous positive outcomes in work life, income, social relationships, physical well-being and coping. Life satisfaction is often used as an outcome measure (Chou and Chi, 1999, Coffey et al., 2015), only rarely as a predictor. It is assumed that happiness, when seen as positive affect, is associated with later life satisfaction. There is indeed evidence for such a relationship (Coffey et al., 2015), but the reversed relationship has not been extensively tested yet. Likewise, positive mental health is often used as an outcome measure. However, it has not been tested whether positive mental health can predict happiness and life satisfaction over a certain course of time. According to the broaden-and-build theory (Fredrickson, 2001), the constructs could mutually influence each other over time. A few cross-sectional studies have examined positive association between happiness and life satisfaction (Chui and Wong, 2016, Lin et al., 2010). However, there is no study including positive mental health as well as examining these associations in a prospective manner.

The Subjective Happiness Scale (SHS) measuring the construct of happiness is a widely and well-validated scale (Iani et al., 2013, Swami et al., 2009). However, its measurement invariance was only tested across cultures so far (Bieda, Hirschfeld, Schönfeld, Brailovskaia, Zhang, & Margraf, 2017); research into measurement invariance across time is lacking.

The LMI of life satisfaction, often measured by the satisfaction with life scale (SWLS), was tested in two small student samples over two and six months. In the general group of students, the SWLS was strong measurement invariant. In the athlete students’ sample, partial measurement invariance was found. Items 2, 3 and 4 were invariant over time (Wu, Chen, & Tsai, 2009).

Positive mental health is often measured by the Mental Health Continuum Short Form MHC-SF (Keyes, 2002). The Positive Mental Health Scale PMH-scale (Lukat et al., 2016), which we used in the present study, also assesses positive mental health and is a brief, unidimensional and person-centered measure including emotional and psychological well-being aspects. It exhibited strong measurement invariance over time (one week in a patient sample to 17 months in young women; Lukat et al., 2016). In sum, there is a great lack of evidence concerning constructs’ LMI over a period longer than one year in student samples.

In the present study, we first tested the assumed unidimensionality and LMI of the constructs happiness (SHS), life satisfaction (SWLS) and positive mental health (PMH-scale). If at least weak LMI could be established, the examination of the stability and interplay of the constructs of happiness, life satisfaction and/or positive mental health over time was possible. We examined the reciprocal relationships among the constructs in a student sample with a four-wave random intercept cross-lagged panel model. A better understanding of the time course and interplay of positive and probable protective factors is crucial for the development of interventions to promote positive mental health. We expected happiness, life satisfaction and positive mental health to be interrelated (Chui and Wong, 2016, Lin et al., 2010), but distinct constructs. Further, we expected the relationships among the positive constructs to be reciprocal: Happiness would predict life satisfaction and positive mental health at every preceding time point, but life satisfaction and positive mental health could also predict happiness.