The architecture of psychological well-being: A network analysis study of the Ryff psychological well-being scale 

Abstract

The proliferation of mental health research is orienting its efforts towards the exploration of psychological well-being. One of the main burdens is the measurement challenges reported by the Psychological Well-being Scale (PWBS), which has often been criticized for inconsistencies between the theoretical and the empirical model. A potential alternative to understand the structure of psychological well-being is network models, which conceptualizes psychological phenomena as emerging systems of mutually connected indicators. We examined the network structure of the Spanish 29-item PWBS in a sample of 1,404 adults. We estimated a regularized partial correlation network using the graphical LASSO algorithm in the item and dimension level. We tested the stability of both networks and identified the most important variables of the network. The PWBS network model revealed four dimensions, with self-acceptance, life purpose and environmental mastery clustering together. Node strength centrality suggested that self-acceptance is the most central dimension in the psychological well-being structure as measured by the PWBS. Despite the network model of psychological well-being did not replicate the theoretical structure of Ryff’s model, it provides a novel conceptualization of psychological well-being and proposes target indicators for mental health interventions.

Introduction

The interest on the study of well-being has dramatically increased within the field of social sciences over the last decades. This advancement has been characterized by the proliferation of several measures that capture different aspects of well-being. One of the most well-known is the Psychological Well-Being Scale (PWBS; Ryff, 1989), which evaluates how individuals function psychologically in response to their life’s demands premised on the eudaimonic notion that well-being comes from realizing human potential (Ryan & Deci, 2001). The PWBS is conceived as a multidimensional instrument composed of six dimensions: self-acceptance (being aware of and accept one’s strength and weaknesses), positive relations (having deep, meaningful relationships with others), personal growth (experiencing progress as a result of developing one’s strengths), autonomy (being self-determined and taking independent decisions), environmental mastery (managing one’s circumstances to take advantage of opportunities), and purpose in life (establishing and being guide by goals). The conjunction of these dimensions resulted from converging different previous models that tried to describe mental health, like the theories of Jahoda, (1958), Maslow (1968), or Rogers (1961). The main goal was to offer a theory-driven basis to assess indicators of positive functioning (Ryff, 1989; Ryff & Keyes, 1995).

But despite its widespread application in research, psychological well-being has been surrounded by debates concerning the structure (e.g., validity and dimensionality) and therefore the applicability to the study of quality of life. Of note, Social Science Research held an academic discussion about the topic (see Ryff & Singer, 2006; Springer et al., 2006). The conclusions opened a still ongoing academic discussion about the adequacy of using the PWBS, since the Ryff’s theory model of six dimensions may not be supported by the PWBS measurement model. These measurement inconsistencies have been reported in other studies (e.g., Burns & Machin, 2009; Kafka et al., 2014; Triadó et al., 2007), so it is common to observe that Ryff’s six-factor model does not provide acceptable fit indices (Abbott et al., 2010; Díaz et al., 2006; Freire et al., 2017; Springer et al., 2006). A three-factor structure whereby self-acceptance, life purpose, personal growth and environmental mastery are combined to form a single dimension have been reported to explain the structure of psychological well-being (Abbott et al., 2010; Burns & Machin, 2009; Kafka et al., 2014), which resembles the self-determination theory of Ryan and Deci (2000). A four-factor model was found using the 29-item version of the PWBS (as in this study) composed of self-acceptance, environmental mastery, personal growth and life purpose, but the authors excluded the dimensions of positive relationships and autonomy (Freire et al., 2017). Other studies have supported the theoretical six-factor (Clarke et al., 2001; Díaz et al., 2006; Ryff & Keyes, 1995) or six-bifactor structure (Espinoza et al., 2018). Researchers generally agree that the PWBS appears as a good indicator of overall psychological well-being, but it fails to identify the six intended specific dimensions. The empirical findings point to a refinement in the measurement of this construct, but a potential limitation is that these findings are based on different analytic strategies (e.g., Exploratory Factor Analysis or Exploratory Structural Equation Modelling), with the majority of them rooting on factor analytic methods in which authors specify a measurement model based on the theoretical model (e.g., Confirmatory Factor Analysis). Factor models are premised on the idea that an underlying cause explains the covariation between indicators (commonly referred to as symptoms in psychopathology), and thus these indicators may not be causally connected. According to this perspective, the co-occurrence between having a sense of purpose in life and feeling confident about oneself is only a cause of (the latent construct of) psychological well-being. However, to better represent the complexity of psychological phenomena it seems a good idea to incorporate and combine contributions from factor models and network models (Fried & Cramer, 2017).

What Can the Network Approach Offer to The Study of Psychological Well-being?

A recent advance across several fields of psychology is the introduction of the network approach (see Fried et al., 2017 for a review). In network research, psychological phenomena are represented as networks, a complex organization of psychological characteristics that emerge from the mutual connections between the observable indicators that define them (e.g., items or subscales) (Borsboom & Cramer, 2013). Network models offer an alternative approach to assess the dimensionality of constructs without necessarily detecting a common latent variable. This approach has the advantage of unveiling existing categories by depicting in a graph the associations (edges; total or partial correlations) present between variables (nodes; Epskamp et al., 2018). Exploratory Graph Analysis (EGA; Golino & Epskamp, 2017) is a technique used to statistically determine the number of dimensions and the allocation of each item within the dimensions, providing a similar performance to exploratory techniques like parallel analysis or minimum average partial.

As a network model, psychological well-being is defined as a system that emerges from mutually interacting indicators, that together, form the construct. It is therefore assumed that indicators are not simply cause or by-product of psychological well-being, as represented by factor models, and the direct interactions between item responses conform the backbone of the network structure. More specifically, psychological well-being might not directly cause the emergence of self-acceptance, autonomy, life purpose, environmental mastery, positive relationships and personal growth. Rather, psychological well-being might emerge as a result of the mutual interactions between these indicators (see Fig. 1). The indicators of this structure can be built under the domain- or item-level, implying that we can examine the network structure of the PWBS as represented by the six dimensions or the 29 items.

The network approach can add valuable knowledge to evaluate which of these indicators (dimensions and items) are more important to psychological well-being. At dimension level, this would inform about how the six theoretical dimensions are built within the structural organization of the PWBS. At item level, it is also possible to analyze if the network clusters into different communities, which would inform about how items conform separate dimensions and the importance of each dimension within the network structure. Maybe some indicators are more important to define psychological well-being and some others are more peripherical. For instance, at the dimension level self-acceptance, environmental mastery, personal growth and life purpose tend to appear as a one single dimension and might be more central (Abbot et al., 2010; Burns & Machin, 2009; Kafka et al., 2014). At the item level, however, it becomes a convoluted task to predict which items would be more important to the network given the high heterogeneity of the PWBS measures used in prior research (e.g., 9-item, 39-item, 42-item, 52-item versions).

Despite the recent explosion of network models, to our knowledge only Kossakowski et al. (2016) advocated this perspective in the assessment of quality of life indicators, evidencing that the network approach has much to offer to the science of well-being. Given its novelty, our study can contribute to the ongoing investigation of psychological well-being by examining the network structure of one of the most expanded measures in the field. We seek to investigate the operationalization of psychological well-being by examining: first, the network structure of the PWBS at the item level; second, the network structure of the PWBS at the dimension level; and third, the importance of the empirically-driven dimensions in explaining psychological well-being by means of centrality test. We expect to find a multidimensional structure, with a potential dimension composed of self-acceptance, environmental mastery, personal growth and life purpose. Given the multiple disagreements in factor analytic studies, we anticipate that the theoretical model of six dimensions would not be identified in the network model.

Methodology

Participants and procedure

The data comprised 1,404 Spanish adults[1] (75.7% females; M = 21.65, SD = 7) who participated voluntarily in a project about personal development. A survey including measures of psychological and subjective well-being was completed as baseline. For the present study, only baseline results of psychological well-being were considered, as baseline results of subjective well-being were used in other studies (Blasco-Belled et al., 2019). We provided participants with access to an online platform to complete a survey including all the measures. They completed the questions in around 15 minutes and obtained an individualized report with their results at the end. We provided an explanation of the study’s aims in the survey, including the terms of confidentiality and anonymity. The participants were extensively informed about the procedure and the purpose of the project, and they signed an informed consent form prior to enrollment in the program. Only participants who completed the whole protocol were accepted into the study and their data were collected.

Instruments

Psychological Well-being Scale (Ryff, 1989; Spanish adaptation of Díaz et al., 2006). This 29-item scale evaluates psychological well-being through 6 dimensions: autonomy, environmental mastery, personal growth, positive relationships, purpose in life, and self-acceptance. It uses a 7-point Likert-type scale (1 = strongly disagree to 7 = strongly agree). A sample item is "I have a sense of direction and purpose in life".

Analytic plan

We used the network analysis approach to assess the structure of psychological well-being. The analyses were carried out in Rstudio (Team R Core, 2020) and the analytic plan consisted of the following steps; first, we estimated two networks and the centrality indices. Second, we assessed the accuracy of these inferences. Third, we scrutinized the dimensionality of the network, and fourth, we tested the stability of the structure. The data and R code are available at Open Science Framework.

Network estimation. First, we estimated a Gaussian Graphical Model (GGM; Lauritzen, 1996), which assumes multivariate normal variables, to represent the structure of psychological well-being in the item and dimension level. In GGM networks, the connection (edges) between variables (nodes) are based on the partial correlation of two nodes after controlling for all other variables of the network. The width of edges represents the strength of the connection between two nodes and thus represent the unique associations between two variables. The edges are undirected (with no arrows), so the associations should not be interpreted as causal (Epskamp, Borsboom, et al., 2018). To avoid overfitting due to the estimation of a large number of parameters, we used the graphical least absolute shrinkage and selection operator (glasso; Epskamp, Waldorp, et al., 2018), a regularization technique that shrinks all connections and sets small coefficients to zero (Tibshirani, 1996). This leads to a sparse network in which only a few edges are included to explain the associations. By this means, glasso preserves false positive edges and prevents erroneous conclusions about the presence of an edge in the model. To estimate the networks, we applied EGA (Golino & Epskamp, 2017) using the EGA function of EGAnet R-package version 0.9.5 (Golino & Christensen, 2020). EGA estimates the correlation matrix of the observed variables before applying a graphical lasso estimation to obtain a sparse inverse covariance matrix in the form of a network (Friedman et al., 2008). To identify which nodes were more important to the network, we used the centrality metric of node strength (i.e., how strongly a node directly connects to other nodes, based on the absolute sum of edge weights connected to a node) using the centralityPlot function of qgraph R-package version 1.6.9 (Epskamp et al., 2012). Recent work has suggested to rely on node strength to interpret the relevance of a node in the network (Isvoranu & Epskamp, 2021).

Dimensionality. Highly connected nodes often cluster and form communities in the network, which are equivalent to latent factors (Golino & Epskamp, 2017). Here we analyzed the internal structure of psychological well-being by examining the number of communities (i.e., dimensions) within the network. Because we were interested in the clustering of items forming communities, dimensionality was only calculated in the item-level network. After identifying the allocation of the items in each dimension, we inspected the network loadings using the net.loads function of the EGAnet R-package 0.9.5, which uses the walktrap algorithm to detect communities. Since the algorithm is deterministic, the resulting communities are not theory-driven, allowing researchers to discover the clustering of items based on the empirical data. Network loadings are equivalent to factor loadings and describe the extent to which a node contributes to a dimension in the network (Christensen & Golino, 2021). As network loadings represent partial correlations loadings, they are typically lower than factor scores. In the present study, we reported the standardized node strength for each node in each dimension in the network.

Network accuracy and structural consistency. After network estimation, it is recommended to test the accuracy in the inference of the network structure. First, we assessed the stability of the dimensions in the item-level network using the bootEGA function of the EGAnet R-package version 0.9.5. We calculated dimension stability based on 1,000 bootstraps applying parametric bootstrapping. This procedure generates data from a multivariate normal distribution with the same number of cases as the original sample and repeatedly evaluates the model. Second, we assessed item stability to inspect the proportion of times that each item is replicated in each dimension. Finally, to ascertain the stability of centrality metrics, we assessed whether the order of strength centrality changed after re-estimating the network with less cases. If the correlation between the original centrality indices and the bootstrapped indices does not change, the interpretation of centralities is plausible (Epskamp et al., 2018). To that end, the correlation stability coefficient (CS-coefficient) tests the maximum proportion of cases that can be dropped. By default, the CS-coefficient computes a correlation of .70 with 95% CI. According to the authors, CS-coefficients > .25 (and preferably >.50) are recommended to interpret centrality differences.

[1] Network analysis are based on (partial) correlations, and despite power analysis techniques are not applied easily, samples > 1000 are considered large (see Isvoranu & Epskamp, 2021)

Results

Table 2 shows that dimension 4 had a perfect structural consistency, and dimensions 2 and 3 almost a perfect structural consistency. However, dimension 1 was less stable than the others. To inspect these results, we analyzed the stability of items in each dimension.

Table 2.

Discussion

Substantial questions about the multidimensional nature of psychological well-being are often examined following empirical measurement models that, theoretically, derive from the conceptualization of psychological well-being. But the existing evidences relying on factor models, which assume that an underlying latent variable is causing psychological well-being, indicate that the resolution of these questions is far from clear. Network models conceptualize psychological well-being as emerging from mutual interactions between observable indicators. In psychopathology and personality, network models have provided a different framework to understand the conceptualization, comorbidity and treatment alternatives of psychological phenomena (see Robinaugh et al., 2020 for a review). Research on well-being seems a potential “client” to benefit from the network approach when it comes to clarifying inconsistencies about the composition of well-being measures.

We estimated the network model of psychological well-being as determined by the Spanish 29-item PWBS to examine its structure in the item and dimension level. Under this scope, the observable indicators measured by the PWBS items covary to generate the construct of psychological well-being. The mutual interaction between PWBS items yielded four dimensions; while dimensions 2, 3 and 4 successfully replicated the theoretical dimensions of autonomy, personal growth and positive relationships, dimension 1 comprised three of the theoretical dimensions (self-acceptance, life purpose and environmental mastery). From the network perspective, this suggests that self-acceptance, life purpose and environmental mastery might be highly enough interconnected and share similar features to form an individual dimension.

A study that used the same Spanish version of the scale also found a four-factor structure (Freire et al., 2017) but the content of dimensions differed from our results. The authors removed the dimensions of positive relationships and autonomy, claiming that this final structure represented the core of eudaimonic well-being found in the literature. However, statistical decisions of items or scales removal without theoretical inferences are unadvisable because they can have substantial implications, specially if we consider psychological constructs as causal systems (Fried & Cramer, 2017). Our results showed that the item-level network does not replicate the six-factor model of Ryff, in accordance with previous studies (Abbott et al., 2010; Burns & Machin, 2009; Espinoza et al., 2018; Kafka et al., 2014; Triadó et al., 2007; Springer et al., 2006). Besides the dimensional analysis, another fundamental finding concerns the identification of important variables (i.e., indicators) of the network structure as potential intervention targets.

If psychological well-being arises from the causal interaction between indicators, either at item or dimension level, it seems useful to determine the importance of each indicator. In network models we can identify important nodes by measuring centrality metrics. The item-level network revealed that feeling proud, confident and positive about oneself were the most central variables, which corresponded to self-acceptance. A sense of a lack of progress in life and feeling responsible for life situations were the least influential variables in the network, which corresponded to personal growth and environmental mastery. Of note, the item referring to progress in life reported the lowest structural consistency in the network. In the dimension-level network, dimension 1 was the most important, yet the least consistent, plausibly because it comprised three dimensions from the original Ryff model. Altogether, self-acceptance seems to exert a substantial influence over the rest of dimensions in the network of psychological well-being.

Activation of central nodes can produce and spread a flow of activation with other nodes in the whole network (Borsboom & Cramer, 2013). Echoing findings from psychopathology (McNally, 2016), by “turning on” central nodes, it may be feasible to influence other nodes that help trigger, develop and maintain states of well-being. Therefore, positive interventions focused on enhancing mental health might be interested in targeting specific indicators of psychological well-being. Aspects relating to self-acceptance seems, according to the present results, potential targets to promote downstream benefits.

Limitations

Our study is not without limitations. Most importantly, we investigated the network structure of a single measure of psychological well-being, adapted to a single language, and applied to a single cross-sectional nonclinical sample. These preclude generalization of findings. Of importance, complex psychological phenomena are not to be seen as a dualism between factor and network models (Fried & Cramer, 2017). For this reason, our study provides the first empirical attempt to understand psychological well-being as a system of interconnected variables. Further research using different samples, contexts and methodological approaches is needed to substantiate our findings. As stated previously, the PWBS has reported psychometric problems in terms of model fit (Springer et al., 2006). To ascertain the structure of psychological well-being, future research employing different measures may be needed.

Conclusions

In sum, our study was the first attempt to subject the PWBS to network analysis. The network architecture of the PWBS provided four dimensions composing the structure of psychological well-being, with self-acceptance, life purpose and environmental mastery clustering together, which is partially in line with previous studies. The theoretical model of Ryff is not empirically supported from the network approach. Using a different perspective from previous research, our results tested the emergence of psychological well-being as resulting from mutually connected indicators rather than from a latent, underlying variable. These findings bring a novel perspective to evaluate the complex multicomponent nature of psychological well-being, and open new paths to inquiry into the promotion of psychological well-being interventions by targeting specific indicators that activate the network. Unfortunately, our findings support previous assertions regarding the discrepancies between the theoretical claims and the empirical distinctions of the Ryff’s model, which poses an academic challenge to the reconsideration of measurement scales of psychological well-being.

Declarations

Competing interests: The authors declare that they have no conflict of interest.

References

Abbott, R. A., Ploubidis, G. B., Huppert, F. A., Kuh, D., & Croudace, T. J. (2010). An evaluation of the precision of measurement of Ryff’s Psychological Well-being scales in a population sample. Social Indicators Research, 97(3), 357–373. https://doi.org/10.1007/s11205-009-9506-x

Blasco-Belled, A., Rogoza, R., Torrelles-Nadal, C., & Alsinet, C. (2019). Emotional Intelligence Structure and Its Relationship with Life Satisfaction and Happiness: New Findings from the Bifactor Model. Journal of Happiness Studies, First Onli. https://doi.org/10.1007/s10902-019-00167-x

Borsboom, D., & Cramer, A. O. J. (2013). Network analysis: An integrative approach to the structure of psychopathology. In Annual Review of Clinical Psychology (Vol. 9, pp. 91–121).  Annual Reviews . https://doi.org/10.1146/annurev-clinpsy-050212-185608

Burns, R. A., & Machin, M. A. (2009). Investigating the structural validity of Ryff’s psychological well-being scales across two samples. Social Indicators Research, 93(2), 359–375. https://doi.org/10.1007/s11205-008-9329-1

Christensen, A. P., & Golino, H. (2021). On the equivalency of factor and network loadings. Behavior Research Methods, 1–18. https://doi.org/10.3758/s13428-020-01500-6

Clarke, P. J., Marshall, V. W., Ryff, C. D., & Wheaton, B. (2001). Measuring psychological well-being in the Canadian Study of Health and Aging. International Psychogeriatrics, 13(SUPPL. 1), 79–90. https://doi.org/10.1017/S1041610202008013

Costantini, G., Epskamp, S., Borsboom, D., Perugini, M., Mõttus, R., Waldorp, L. J., & Cramer, A. O. J. (2015). State of the aRt personality research: A tutorial on network analysis of personality data in R. Journal of Research in Personality, 54, 13–29. https://doi.org/10.1016/j.jrp.2014.07.003

Díaz, D., Rodríguez-Carvajal, R., Blanco, A., Moreno-Jimenez, B., Gallardo, I., Valle, C., & van Dierendonck, D. (2006). Adaptación española de las escalas de bienestar psicológico de Ryff. Psicothema, 18(3), 572–577.

Epskamp, S., Borsboom, D., & Fried, E. I. (2018). Estimating psychological networks and their accuracy: A tutorial paper. Behavior Research Methods, 50(1), 195–212. https://doi.org/10.3758/s13428-017-0862-1

Epskamp, S., Cramer, A. O. J., Waldorp, L. J., Schmittmann, V. D., & Borsboom, D. (2012). qgraph: Network visualizations of relationships in psychometric data. Journal of Statistical Software, 48(1), 1–18. https://doi.org/10.18637/jss.v048.i04

Epskamp, S., Waldorp, L. J., Mõttus, R., & Borsboom, D. (2018). The Gaussian Graphical Model in Cross-Sectional and Time-Series Data. Multivariate Behavioral Research, 53(4), 453–480. https://doi.org/10.1080/00273171.2018.1454823

Espinoza, J. A., Meyer, J. P., Anderson, B. K., Vaters, C., & Politis, C. (2018). Evidence for a Bifactor Structure of the Scales of Psychological Well-Being Using Exploratory Structural Equation Modeling. Journal of Well-Being Assessment, 2(1), 21–40. https://doi.org/10.1007/s41543-018-0008-y

Freire, C., Ferradás, M. del M., Núñez, J. C., & Valle, A. (2017). Estructura factorial de las Escalas de Bienestar Psicológico de Ryff en estudiantes universitarios. European Journal of Education and Psychology, 10(1), 1–8. https://doi.org/10.1016/j.ejeps.2016.10.001

Fried, E. I., & Cramer, A. O. J. (2017). Moving Forward: Challenges and Directions for Psychopathological Network Theory and Methodology. Perspectives on Psychological Science, 12(6), 999–1020. https://doi.org/10.1177/1745691617705892

Fried, E. I., van Borkulo, C. D., Cramer, A. O. J., Boschloo, L., Schoevers, R. A., & Borsboom, D. (2017). Mental disorders as networks of problems: a review of recent insights. Social Psychiatry and Psychiatric Epidemiology, 52(1), 1–10. https://doi.org/10.1007/s00127-016-1319-z

Friedman, J., Hastie, T., & Tibshirani, R. (2008). Sparse inverse covariance estimation with the graphical lasso. Biostatistics, 9(3), 432–441. https://doi.org/10.1093/biostatistics/kxm045

Golino, H. F., & Christensen, A. P. (2020). EGAnet: Exploratory Graph Analysis – A framework for estimating the number of dimensions in multivariate data using network psychometrics. R package version 0.9.5. https://cran.r-project.org/web/packages/EGAnet/citation.html

Golino, H. F., & Epskamp, S. (2017). Exploratory graph analysis: A new approach for estimating the number of dimensions in psychological research. PLOS ONE, 12(6), e0174035. https://doi.org/10.1371/journal.pone.0174035

Isvoranu, A.-M., & Epskamp, S. (2021). Continuous and Ordered Categorical Data in Network Psychometrics: Which Estimation Method to Choose? Deriving Guidelines for Applied Researchers. https://doi.org/10.31234/OSF.IO/MBYCN

Jahoda, M. (1958). Current concepts of positive mental health. Basic Books. https://doi.org/10.1037/11258-000

Kafka, G. J., Kozma, A., & Kafka, J. (2014a). The Construct Validity of Ryff ’ s Scales of Psychological Well-Being ( SPWB ) and Their Relationship to Measures of Subjective Well-Being THE CONSTRUCT VALIDITY OF RYFF ’ S SCALES OF WELL-BEING ( SPWB ) AND PSYCHOLOGICAL THEIR RELATIONSHIP WELL-BEING. 57(2), 171–190.

Kafka, G. J., Kozma, A., & Kafka, J. (2014b). The Construct Validity of Ryff ’ s Scales of Psychological Well-Being (SPWB) and Their Relationship to Measures of Subjective Well-Being. Social Indicators Research, 57(2), 171–190.

Lauritzen, S. L. (1996). Graphical Models. Clarendon Press. https://books.google.es/books?hl=es&lr=&id=mGQWkx4guhAC&oi=fnd&pg=PA1&dq=Graphical+models+lauritzen&ots=2Ldjm8DpWj&sig=N8dX-7jBztOMJZ017SAoOgiMrSo#v=onepage&q&f=false

Maslow, A. (1968). Some Educational Implications of the Humanistic Psychologies. Harvard Educational Review, 38(4), 685–696. https://doi.org/10.17763/haer.38.4.j07288786v86w660

McNally, R. J. (2016). Can network analysis transform psychopathology? In Behaviour Research and Therapy (Vol. 86, pp. 95–104). Elsevier Ltd. https://doi.org/10.1016/j.brat.2016.06.006

Robinaugh, D., Hoekstra, R., Toner, E., & Borsboom, D. (2020). The Network Approach to Psychopathology: A Review of the Literature 2008–2018 and an Agenda for Future Research. Psychological Medicine, 50(3), 139-353–366. https://doi.org/10.1017/S0033291719003404

Rogers, C. R. (1961). On Becoming a Person. Houghton Mifflin.

Ryan, R. M., & Deci, E. L. (2000). Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being Self-Determination Theory. 55(1), 68–78.

Ryan, R. M., & Deci, E. L. (2001). On Happiness and Human Potentials: A Review of Research on Hedonic and Eudaimonic Well-Being. Annual Review of Psychology, 52(1), 141–166. https://doi.org/10.1146/annurev.psych.52.1.141

Ryff, C. (1989). Happiness Is Everything, or Is It? Explorations on the Meaning of Psychological Well-Being. Journal of Personality and Social Psychology, 57(6), 1069–1081.

Ryff, C. D. (1989). Happiness Is Everything, or Is It? Explorations on the Meaning of Psychological Well-Being. In Journal of Personality and Social Psychology (Vol. 57, Issue 6).

Ryff, C. D., & Keyes, C. (1995). The Structure of Psychological Well-Being Revisited. Journal of Personality and Social Psychology, 69(4), 719–727.

Ryff, C. D., & Singer, B. H. (2006). Best news yet on the six-factor model of well-being. Social Science Research, 35(4), 1103–1119. https://doi.org/10.1016/j.ssresearch.2006.01.002

Springer, K. W., Hauser, R. M., & Freese, J. (2006). Bad news indeed for Ryff’s six-factor model of well-being. Social Science Research, 35(4), 1120–1131. https://doi.org/10.1016/j.ssresearch.2006.01.003

Tibshirani, R. (1996). Regression Shrinkage and Selection Via the Lasso. Journal of the Royal Statistical Society: Series B (Methodological), 58(1), 267–288. https://doi.org/10.1111/j.2517-6161.1996.tb02080.x

Triadó, C., Villar, F., Solé, C., & Celdrán, M. (2007). Construct Validity of Ryff’s Scale of Psychological Well-Being in Spanish Older Adults. Psychological Reports, 100(3_suppl), 1151–1164. https://doi.org/10.2466/pr0.100.4.1151-1164