Adapting to the work-life interface: The influence of individual differences, work and family on well-being, mental health and work engagement By Prudence M R Millear B Sc Ag (Hons), Grad Dip Psych, B Psych (Hons) A thesis submitted in fulfilment of the requirements of the degree of Doctor of Philosophy School of Psychology and Counselling Faculty of Health Queensland University of Technology February 2010 i ii Keywords Bronfenbrenner, dispositional optimism, coping self-efficacy, affective commitment, skill discretion, job autonomy, life satisfaction, psychological well-being, mental health, work engagement, burnout, longitudinal modelling, gain spirals, loss spirals, Conservation of Resources, resource caravans, working adults iii iv Statement of Original Authorship The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution To the best of my knowledge and belief, this thesis contains no material previously published or written by another person except where due reference is made Signature …………………………………………………………………………… Date ………………………………………………………………………………… v vi Publications and presentations arising from the PhD research Journal articles Millear, P.M & Liossis, P.L., Gain spirals and resource caravans: An integrated longitudinal model of well-being, mental health and work engagement among Australian workers, under review, Journal of Occupational and Organizational Psychology Book Chapters Millear, P.M & Liossis, P.L (2010) Longitudinal modelling of individual differences and the workplace: well-being and work engagement Chapter 18 in Hicks, R.E (Ed.) Personality and Individual Differences: Current Directions Brisbane, Australia: Australian Academic Press Millear, P.M & Liossis, P.L Doing it for yourself: The choices and strategies of managing the work-life challenge, accepted for publication, Wayfinding through life‟s challenges: Coping and survival, Nova Science Publishers, NY, K Gow & M Celinski (Eds) Conference Presentations 2008 European Academy of Occupational Health Psychology conference at the University of Valencia, November 2008, presentations: Longitudinal modelling of well-being and mental health in Australian workers; Exploring burnout and work engagement in diverse occupations: A continuum or two separate factors? vii 2008 Australian Conference for Personality and Individual Differences (ACPID), Bond University, (November, 2008) “Longitudinal modelling of the influence of individual differences and the workplace on well-being and work engagement” 2009 8th Industrial and Organizational Psychology conference, Sydney (June 2009), Paper Presentation, “An integrated longitudinal model of well-being, mental health and work engagement among Australian workers” viii Acknowledgements I would like to acknowledge and sincerely thank my supervisors, Dr Poppy Liossis, my Principal Supervisor and Professor Ian Shochet, my Associate Supervisor for the guidance and support that they have provided throughout my candidature I would also like to thank Dr Cameron Hurst and Dr Trish Obst for their assistance with the Structural Equation Modelling that I undertook and thank Cameron particularly for deciphering the process of longitudinal modelling I would like to thank my postgraduate friends for their unstinting support, coffee and sympathy and thank my family and friends for helping where they could My greatest thanks are to my husband and children for bearing with me and understanding the work involved in completing my thesis, in the middle of family life, rugby and the house renovations We have done this together ix x e1 e2 e3 e4 e5 e1 IFwbmh1 PWFwbmh1 NSpwbmh1 OWBwbmh1 MIwbmh1 IFwbmh1 1 e6 1 e7 e8 IFwbmh2 PWFwbmh2 NSPwbmh2 IFwbmh3 e11 e10 OWBwbmh2 MIwbmh2 OWBwbmh3 NSPwbmh3 PWFwbmh3 e12 1 e13 e14 MIwbmh3 IFwbmh3 IFwbmh1 e6 PWFwbmh1 e7 NSpwbmh1 e8 1 IFwbmh2 PWFwbmh2 NSPwbmh2 IFwbmh3 e11 PWFwbmh3 e12 e4 e5 OWBwbmh1 MIwbmh1 1 e8 e9 e12 NSPwbmh3 e4 OWBwbmh3 e13 e14 OWBwbmh2 OWBwbmh3 1 e13 e14 Model D Reciprocal e6 MIwbmh3 e15 e11 NSpwbmh1 e7 OWBwbmh1 e5 MIwbmh1 e8 e9 e10 NSPwbmh2 OWBwbmh2 MIwbmh2 NSPwbmh3 PWFwbmh3 e12 e11 OWBwbmh3 1 e13 e14 Model C Reverse Causality e4 e3 PWFwbmh1 e7 NSpwbmh1 e8 PWFwbmh3 e12 NSPwbmh3 e5 1 OWBwbmh1 MIwbmh1 e10 e9 1 IFwbmh2 PWFwbmh2 NSPwbmh2 IFwbmh3 IFwbmh3 e15 e2 IFwbmh1 MIwbmh2 IFwbmh2PWFwbmh2 MIwbmh3 1 e10 e9 e6 MIwbmh2 e1 MIwbmh1 PWFwbmh1 1 OWBwbmh1 IFwbmh1 e4 1 e10 NSPwbmh2 OWBwbmh2 e5 e3 e2 e1 Model B Causality e3 e2 e11 e15 e3 NSpwbmh1 PWFwbmh3 NSPwbmh3 1 1 IFwbmh2 PWFwbmh2 Model A Stability e1 e7 1 PWFwbmh1 e6 e9 e2 1 OWBwbmh2 OWBwbmh3 1 e13 e14 MIwbmh2 MIwbmh3 e15 Model E Trimmed Figure J.3 Competing set of models for the Well-Being – Mental Health model, with the best fitting model, Model E 497 MIwbmh3 e15 e1 e2 e3 IFwa1 PWFwa1 NSPwa1 1 IFwa2 IFwa3 e9 PWFwa2 e2 e3 WEwa1 PWFwa1 NSPwa1 e8 NSPwa2 PWFwa3 e10 IFwa3 e9 e12 Model A Stability e2 e3 IFwa1 PWFwa1 NSPwa1 1 e6 e7 IFwa2 PWFwa2 NSPwa2 1 e9 e4 WEwa1 WEwa1 e11 PWFwa3 e10 WEwa1 WEwa3 NSPwa3 e11 Model D Reciprocal WEwa3 IFwa3 PWFwa2 PWFwa3 1 Model C Reverse Causality e3 e4 IFwa1 PWFwa1 NSPwa1 WEwa1 IFwa2 IFwa3 e7 PWFwa2 NSPwa2 PWFwa3 e10 e8 e6 e5 e9 NSPwa3 e11 e2 1 1 e10 e12 1 e12 IFwa2 1 WEwa2 e7 NSPwa2 1 NSPwa3 e11 e8 e6 e1 1 1 e9 e10 e4 e5 WEwa2 NSPwa3 e8 e5 e3 NSPwa1 Model B Causality e1 IFwa3 e2 PWFwa1 e8 NSPwa2 e1 IFwa1 1 PWFwa3 1 e11 PWFwa2 IFwa2 e4 e7 1 WEwa3 1 e6 e5 NSPwa3 1 WEwa2 1 e1 IFwa1 e7 e6 e5 e4 WEwa2 WEwa3 e12 Model E Trimmed Figure J.4 Models compared by the Work Engagement model, with Model E, the best fitting model 498 WEwa2 WEwa3 e12 e1 1 IFcm1 e2 IFcm2 e9 PWFcm2 e13 OWBcm2 MIcm2 OWBcm3 e15 e14 e11 e10 NSPcm3 e6 MIcm1 NSPcm2 PWFcm3 e5 OWBcm1 1 IFcm3 e8 e7 e4 e3 NSPcm1 PWFcm1 MIcm3 e1 WAcm1 IFcm1 IFcm1 IFcm2 WAcm2 WAcm3 IFcm3 1 e16 e17 e18 e13 IFcm2 IFcm3 e13 e4 e3 PWFcm1 e8 PWFcm2 PWFcm3 NSPcm1 e9 e14 e5 OWBcm1 e10 NSPcm2 NSPcm3 1 e8 PWFcm2 PWFcm3 NSPcm1 e9 e14 OWBcm1 e10 NSPcm2 NSPcm3 1 e5 e15 e6 e11 e15 e6 MIcm1 e11 OWBcm2 MIcm2 OWBcm3 e16 MIcm3 e17 Model D Reciprocal WEcm1 e12 WEcm2 WEcm2 WEcm3 e18 WEcm3 NSPcm1 PWFcm1 e8 IFcm2 PWFcm2 1 e16 e17 e18 e13 e10 PWFcm3 NSPcm2 NSPcm3 OWBcm3 e15 e14 IFcm1 e2 PWFcm1 e7 e8 IFcm2 PWFcm2 IFcm3 1 e13 PWFcm3 e14 e4 e3 NSPcm1 e5 OWBcm1 e9 e10 1 NSPcm2 NSPcm3 e15 e6 MIcm1 e11 OWBcm2 MIcm2 OWBcm3 MIcm3 WEcm3 WEcm3 1 e18 e18 e12 e17 MIcm3 e12 WEcm2 e17 WEcm2 e16 e16 1 e11 WEcm1 WEcm1 MIcm1 Model C Reverse Causality e1 e6 OWBcm2 MIcm2 1 e5 OWBcm1 e9 IFcm3 e4 e3 e7 1 e2 1 MIcm3 1 IFcm1 e12 OWBcm2 MIcm2 OWBcm3 e1 WEcm1 MIcm1 Model B Causality 1 e4 e3 1 1 e2 e7 e2 PWFcm1 e7 e12 Model A Stability e1 1 Model E Trimmed Figure J.5 The set of models compared by the Integrated model, with model E, the best fitting 499 Appendix J: Synchronous correlations in each model Table J.12 Synchronous correlations of the variables in the Well-Being model Time IFwb1 IFwb1 PWFwb1 PWFwb1 OWBwb1 WWBwb1 582*** 982*** 515*** 591*** 967*** OWB1 WWB1 495*** Time IFwb2 IFwb2 PWFwb2 OWBwb2 WWBwb2 453*** 964*** 369*** 426*** 943*** PWFwb2 OWBwb2 WWBwb2 296*** Time IFwb3 PWFwb3 IFwb3 PWFwb3 OWBwb3 WWBwb3 402*** 961*** 327*** 388*** 953*** OWBwb3 WWBwb3 283*** * p < 05, ** p < 01, *** p < 001 Note IF: Individual Factors, PWF: Positive Workplace Factors, OWB: overall Well-Being, WWB: Work WellBeing; „wb‟ composite variables in the Well-Being model; 1, 2, = times 1, 2, respectively 500 Table J.13 Synchronous correlations between variables at each time period of the Mental Distress model Time IFmi1 IFmi1 PWFmi1 PWFmi1 513*** NSPmi1 BURNmi1 MImi1 -.634*** -.711*** -.829*** -.663*** -.937*** -.451*** 857*** 816*** NSPmi1 BURNmi1 MImi1 717*** Time IFmi2 IFmi2 PWFmi2 PWFmi2 399*** NSPmi2 BURNmi2 MImi2 -.551*** -.624*** -.738*** -.491*** -.914*** -.229*** 770*** 715*** NSPmi2 BURNmi2 MImi2 572*** Time IFmi3 IFmi3 PWFmi3 PWFmi3 NSPmi3 BURNmi3 MImi3 317*** -.532*** -.571*** -.783*** -.567*** -.922*** -.271*** 809*** 727*** NSPmi3 BURNmi3 MImi3 1 591*** * p < 05, ** p < 01, *** p < 001 Note IF: Individual Factors, PWF: Positive Workplace Factors, NSP: negative Spillover, BURN: burnout, MI: Mental Illness; „mi‟ composite variables in the Mental Distress model; 1, 2, = times 1, 2, respectively 501 Table J.14 Synchronous correlations between variables in the Well-Being- Mental Health model Time IFwbmh1 IFwbmh1 PWFwbmh1 PWFwbmh1 NSPwbmh1 605*** NSPwbmh1 OWBwbmh1 MIwbmh1 -.572*** 978*** -.789*** -.550*** 667*** -.419*** -.537 *** 824*** OWBwbmh1 MIwbmh1 -.694*** Time IFwbmh2 IFwbmh2 PWFwbmh2 PWFwbmh2 NSPwbmh2 471*** OWBwbmh2 MIwbmh2 -.511*** 946*** -.729*** -.391*** 530*** -.260*** -.414*** 795*** -.537*** NSPwbmh2 OWBwbmh2 MIwbmh2 Time IFwbmh3 IFwbmh3 PWFwbmh3 PWFwbmh3 NSPwbmh3 403*** NSPwbmh3 OWBwbmh3 MIwbmh3 OWBwbmh3 MIwbmh3 -.535*** 936*** -.759*** -.455*** 483*** -.236*** -.451*** 789*** -.551*** * p < 05, ** p < 01, *** p < 001 Note IF: Individual Factors, PWF: Positive Workplace Factors, NSP: Negative Spillover, OWB: overall WellBeing, MI: Mental Illness; „wbmh‟ composite variables in the Well-Being-Mental Health model; 1, 2, = times 1, 2, respectively 502 Table J.15 Synchronous correlations between variables at each time period of the Work Engagement model Time IFwa1 IFwa1 PWFwa1 564*** PWFwa1 NSPwa1 NSPwa1 WEwa1 -.582*** 556*** -.484*** 967*** WEwa1 -.459 *** Time IFwa2 IFwa2 PWFwa2 425*** PWFwa2 NSPwa2 WEwa2 -.429*** 383*** -.302*** 939*** -.305*** NSPwa2 WEwa2 Time IFwa3 IFwa3 PWFwa3 NSPwa3 WEwa3 420*** -.419*** 415*** -.294*** 949*** -.254*** PWFwa3 NSPwa3 WEwa3 1 p < 05, ** p < 01, *** p < 001 Note IF: Individual Factors, PWF: Positive Workplace Factors, NSP: Negative Spillover, WE: Work Engagement; „wa‟ composite variables in the Work Engagement model; 1, 2, = times 1, 2, respectively 503 Table J.16 Synchronous correlations between variables in the Integrated model Time IFcm1 IFcm1 PWFcm1 NSPcm1 OWBcm1 MIcm1 WEcm1 558*** -.513*** 970*** -.778*** 509*** -.353*** 567*** -.388*** 974*** -.538*** 740*** -.398*** -.696*** 479*** -.417*** PWFcm1 NSPcm1 OWBcm1 MIcm1 WEcm1 Time IFcm2 IFcm2 PWFcm2 NSPcm2 OWBcm2 MIcm2 WEcm2 410*** -.425*** 938*** -.678*** 349*** -.208** 382*** -.166* 964*** -.428*** 627*** -.270*** -.529*** 257*** PWFcm2 NSPcm2 OWBcm2 MIcm2 WEcm2 -.203** Time IFcm3 IFcm3 PWFcm3 NSPcm3 PWFcm3 NSPcm3 OWBcm3 MIcm3 WEcm3 354*** -.457*** 902*** -.723*** 327*** -.375*** 363*** -.187** 969*** -.426*** 620*** -.425*** -.480*** 259*** 1 OWBcm3 MIcm3 WEcm3 † -.230** p < 10, * p < 05, ** p < 01, *** p < 001 Note IF: Individual Factors, PWF: Positive Workplace Factors, NSP: Negative Spillover, OWB: Overall WellBeing, MI: mental Illness WE: Work Engagement; „cm‟ composite variables in the Integrated model; 1, 2, = times 1, 2, respectively 504 Standardized regression weights of the auto-lagged and cross-lagged paths for the models Table J.17 Standardized regression weights for auto-lagged and cross-lagged paths in the Well-Being model „Input‟ variables a IFwb1 „Outcome‟ variables a IFwb2 IFwb3 685*** 304*** IFwb2 PWFwb2 860*** WWBwb2 WWBwb3 395*** 077** 219*** 808*** 184 OWBwb2 484*** 699*** 181*** 340*** 589*** 054* 611*** WWBwb2 † OWBwb3 143 PWFwb2 WWBwb1 OWBwb2 444*** PWFwb1 OWBwb1 PWFwb3 366*** -.340*** p < 10, * p < 05, ** p < 01, *** p < 001 Note Auto-lagged paths on the leading diagonal; Causality paths in upper triangular matrix; Reverse Causality paths in lower triangular matrix Note: a „Input‟ and „Outcome‟ variables refer to variables at the beginning and the end of the causal arrow, respectively, as shown by the explanatory diagram „Input‟ IFwb1 = 685*** „Outcome‟ IFwb2 „Input‟ IFwb2 = 444*** „Outcome‟ IFwb3 505 Table J.18 Standardized regression weights for the auto-lagged and cross-lagged paths of the Mental Distress model „Input‟ Variable a „Outcome‟ variable a IFmi2 IFmi3 IFmi1 897*** 270*** IFmi2 496*** PWFmi1 PWFmi2 PWFmi3 1.059*** NSPmi2 NSPmi3 MImi2 266*** MImi3 BURNmi2 -.162* -.048** -.142* -.331** BURNmi3 PWFmi2 352*** NSPmi1 631*** NSPmi2 MImi1 125* 065† 441*** 179** MImi2 BURNmi1 291*** -.106† -.119** -.158*** 067** 315† -.124* BURNmi2 310** 102** 244*** 472*** 208** 468*** -.609*** † 265*** 904*** p < 10, * p < 05, ** p < 01, *** p < 001 Note Auto-lagged paths on the leading diagonal; Causality paths in upper triangular matrix; Reverse Causality paths in lower triangular matrix a „Input‟ and „Outcome‟ variables refer to variables at the beginning and the end of the causal arrow, respectively, as shown by the explanatory diagram „Input‟ IFmi1 = 897*** „Outcome‟ „Input‟ IFmi2 IFmi2 = 496*** „Outcome‟ IFmi3 506 Table J.19 Standardized regression weights for the auto-lagged and cross-lagged paths of the Well-Being - Mental Health model „Input‟ Variables a IFwbmh „Outcome‟ variables IFwbmh2 IFwbmh3 PWFwbmh2 PWFwbmh3 NSPwbmh2 NSPwbmh3 OWBwbmh2 OWBwbmh3 MIwbmh2 MI3wbmh 647*** IFwbmh2 256*** -.213*** 651*** 229*** PWFwbmh1 840*** PWFwbmh2 361*** 509*** NSPwbmh1 803*** NSPwbmh2 254*** 171*** 882*** OWBwbmh2 -.141** 288*** 357** -.136† 050* MIwbmh2 † 134* 604*** OWBwbmh1 259*** MIwbmh1 -.175** 344*** -.179* 218*** 237*** p < 10, * p < 05, ** p < 01, *** p < 001 Note Auto-lagged paths on the leading diagonal; Causality paths in upper triangular matrix; Reverse Causality paths in lower triangular matrix a „Input‟ and „Outcome‟ variables refer to variables at the beginning and the end of the causal arrow, respectively, as shown by the explanatory diagram „Input‟ IFwbmh1 = 647*** „Outcome‟ IFwbmh2 „Input‟ IFwbmh2 = 651*** „Outcome‟ IFwbmh3 507 Table J.20 Standardized regression weights for the auto-lagged and cross-lagged paths of the Work Engagement model „Input‟ „Outcome‟ variables a Variable a IFwa2 IFwa3 IFwa1 867*** 240*** IFwa2 PWFwa2 PWFwa3 997*** 291*** NSPwa3 WEwa2 WEwa3 653*** PWFwa1 PWFwa2 422** 679*** NSPwa1 334* 749*** NSPwa2 271*** 561*** WEwa1 -.150 WEwa2 † NSPwa2 415** -.114 270*** 226 p < 10, * p < 05, ** p < 01, *** p < 001 Note Auto-lagged paths on the leading diagonal; Causality paths in upper triangular matrix; Reverse Causality paths in lower triangular matrix a „Input‟ and „Outcome‟ variables refer to variables at the beginning and the end of the causal arrow, respectively, as shown by the explanatory diagram „Input‟ IFwa1 = 867*** „Outcome‟ „Input‟ IFwa2 IFwa2 = 653*** „Outcome‟ IFwa3 508 Table J.21 Standardized regression weights for the auto-lagged and cross-lagged paths for the Integrated model „Input‟ „Outcome‟ variables a Variables a IFcm2 IFcm1 633*** 269*** IF2cm 616*** IFcm3 PWFcm2 PWFcm3 NSPcm2 NSPcm3 OWBcm2 OWBcm3 MIcm2 MIcm3 WEcm2 WEcm3 -.221** 176*** PWFcm1 845*** 329*** PWFcm2 736*** NSPcm1 084** 762*** 304*** 253*** -.115* -.015† 317*** 434*** 048*** 311*** 208*** -.113† 354*** 065* 651*** WEcm2 † -.022* 094* 828*** MIcm2 WEcm1 166*** 185*** 540*** OWBcm2 MIcm1 -.054 348* NSPcm2 OWBcm1 -.139* -.119 315*** 184 p < 10, * p < 05, ** p < 01, *** p < 001 Note Auto-lagged paths on the leading diagonal; Causality paths in upper triangular matrix; Reverse Causality paths in lower triangular matrix a „Input‟ and „Outcome‟ variables refer to variables at the beginning and the end of the causal arrow, respectively, as shown by the explanatory diagram „Input‟ IFcm1 = 633*** „Outcome‟ IFcm2 „Input‟ IFcm2 = 616*** „Outcome‟ IFcm3 509 Appendix K: Terms and glossary for Study 2, Longitudinal modelling Figure K.1 The set of non-nested longitudinal models that were compared in Study Table K.1 An explanation of the non-nested models used in the longitudinal models Model name Pathways in model Stability (A) Synchronous correlations between errors of variables at the same time and auto-lagged paths between same variables over time Causality (B) Stability + cross-lagged paths from „predictors‟ to „outcomes‟ over time Reverse Causality (C) Stability + cross-lagged paths from „outcomes‟ to „predictors‟ over time Reciprocal (D) Stability + Causality + Reverse causality models Trimmed (E) Reciprocal model with trivial paths, < 10 and p < 20 removed to show true non-zero pathways Designation of time in the models Time Time Time „tm1‟ or „1‟ „tm2‟ or „2‟ „tm3‟ or „3‟ Notes on the SEM figures: Double-headed arrows indicate correlations between the two variables Single headed arrows indicate the direction of causal influence, from „cause‟ to „effect‟ „e‟ indicates the measurement error for the variable 510 Table K.2 The assessment of good fit and parsimony of the CFAs and the longitudinal models Fit indices Range of good fit and parsimony X2/df 1.00 – 3.00 CFA 0.95 – 1.00 RMSEA (point estimate) Perfect fit = 00; Close fit ≤ 05; Reasonable fit between 05 and 08; Mediocre fit between 08 and 10; Poor fit ≥ 10 RMSEA (95% CI) Close fit if lower bound estimate < 05; Reasonable fit if upper bound estimate < 08 Poor fit if upper bound estimate > 10 AIC Lowest estimate is most parsimonious model ECVI Lowest estimate is model most likely to be replicated in similar samples Table K.3 Variables in Study 2, Longitudinal modelling Factor label Latent Variable Indicator variables IF Individual Factors Dispositional optimism Coping self-efficacy PWF Positive Workplace Factors Job autonomy Skill discretion Affective commitment NSP Negative spillover Negative work-to-family spillover Negative family-to-work spillover OWB Overall well-being Life satisfaction Psychological well-being MI Mental Illness Depression Anxiety Stress WWB Work Well-Being Work dedication Work absorption BURN Burnout Emotional exhaustion Cynicism Professional efficacy WE Work Engagement Work dedication Work absorption Professional efficacy Table K.4 Names of the models and the latent variables used in the CFAs Model Label of latent factors in model Well-Being IF, PWF, OWB, WWB Mental Distress IF, PWF, NSP, MI, BURN Well-Being-Mental Health IF, PWF, NSP, OWB, MI Work Engagement IF, PWF, NSP, WE Integrated IF, PWF, NSP, OWB, MI, WE Model Postscript wb mi wbmh wa cm 511 ... psychology to study the work- life interface of working adults Occupational and organizational psychology is focused on the demands and resources of work and family, without emphasising the individual. .. useful framework for the current research, showing the importance of the person as central to the individual s experience of the work- life interface By taking control of their own life, the individual. .. (measured as life satisfaction and psychological well- being) and high work engagement (as work vigour, work dedication and absorption in work) and as the absence of mental illness (as depression, anxiety