Skip to main content
Log in

Determinants of Work-life Balance: Shortcomings in the Contemporary Measurement of WLB in Large-scale Surveys

  • Published:
Social Indicators Research Aims and scope Submit manuscript


Research on work-life balance (WLB) has presented important insights into the problems of combining family aspirations with paid work in relation to policy relevant agendas. Using the ESS II (2004/2005), we examine work-related and household/family-related causes of WLB. We can corroborate other research findings that show that work-related aspects explain by far the largest part of the variation in WLB. However, we illustrate that the measurement of WLB is partly problematic. Because WLB scales conceptualize the work component more specifically than the life component, what ‘life’ means remains rather intangible apart from general references to the ‘home’, ‘housework’ and ‘family responsibilities’. This largely neglects different emic dimensions to WLB common to specific subgroups and renders the measurement rather abstract. Second, the wordings of WLB indicators already include their most probable explanations. There is the danger of a circular argument here and many explanations seem tautological. This makes it difficult to conclude on the effects of other than work-related aspects on WLB, which are, arguably, also important aspects of WLB. Finally, WLB scales hardly correlate with relevant external criteria, for instance subjective well-being. Following from these findings, we discuss what these WLB scales could really measure and propose to broaden quantitative empirical approaches to it.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others


  1. The search term was “work-life balance” used at on 19 November 2007. Blackle is the allegedly environmentally friendly version of Google.

  2. Some examples are the European Union (DG Employment and Social Affairs), the European Foundation for the Improvement of Living and Working Conditions, the OECD, for UK philanthropic institutions see, for instance, the Work Foundation.

  3. Hence, higher scores indicate imbalance rather than balance of work and life. We deliberately depart from the common practice that higher scores of an index/scale indicate positive outcomes to emphasize the nature of WLB. This highlights the important fact that researchers are usually interested in the causes of imbalance rather than balance. Needless to say, this coding has no effects on the result apart from the circumstance that positive associations with WLB indicate causes of imbalance.

  4. This is a reasonable limitation to the data as it cannot be assumed that a common measurement of WLB exhibits metric and/or scalar equivalence across countries. Our attempts to find such measurement models failed to the extent that only a measurement model of WLB including three indicators (‘worry’, ‘tired’ and ‘prevents time’) exhibited metric but not scalar equivalence (results are available form the author upon request). Because structural equation modelling (SEM) would conclude that measures cannot be compared across countries and thus terminate the analysis, we note to be cautious about substantial interpretations of country means using different methods. Differences in the mean can thus also be related to bias and error instead of merely substantial differences in WLB across countries.

  5. See specifically a lively discussion in the multilevel list at

  6. Factor analysis supports our theoretical assumption that all five items of WLB are related to one, and only one common dimension. The first Eigen value is equal to 2.55; the second Eigen value is smaller than 1 (0.80) and thus not extracted.

  7. Because variation in WLB is first explained by household-related aspects, the remaining indicators can only capture what is not explained away so far. We have alternated the order of inclusions of blocks of variables in the model and conclude that this order matters. Inserting work-related aspects prior to the household explanations almost renders household-related explanations redundant. In the presented model, we prioritize household-related explanations because our criticism later on will particularly focus on the low relevance of the household for WLB when considering the available measures.

  8. Less favourable models showed that household-related aspects explain <6% in total, with objective conditions explaining just 1%.

  9. For a short scale of WLB, including the indicators ‘worrying’, ‘too tired’ and ‘prevents time’, we have found metric invariance across all participating countries. Using multi-group confirmatory factor analysis (MGCFA) in LISREL, we find that the metric invariance model yields acceptable fit (χ2 = 277.44; df = 48, p-value < 0.001, RMSEA = 0.063 and NFI = 0.986). Though the χ2 value is significant, RMSEA and NFI are below the thresholds for rejection. This short WLB scale exhibits high loadings on all three indicators (0.55 on ‘worrying’, 0.79 on ‘too tired’ and 0.73 on ‘prevents time’). We have also run the analysis with this short version, but we decided against it because it was even ‘less favourable’ to household-related explanations of WLB. On the other hand, alternative statistical analyses have shown that ‘worrying’ does not fit other WLB indicators to the same extent.

  10. Wording of life satisfaction: ‘All things considered, how satisfied are you with your life as a whole nowadays? Please answer using this card, where 0 means extremely dissatisfied and 10 means extremely satisfied?’

  11. Wording of happiness: ‘Taking all things together, how happy would you say you are? Please use this card.’ 11-point scale of 00 ‘extremely unhappy’ to 10 ‘extremely happy’.

  12. Wording of subjective health: ‘How is your health in general? Would you say it is…very good (1), good (2), fair (3), bad (4) or, very bad (5)?’

  13. Wording of feelings: ‘Firstly, I am going to read out a list of statements about how you may have been feeling recently. For each statement, using this card, I would like you to say how often you have felt like this over the last two weeks. Please use this card: 1) I have felt cheerful and in good spirits. 2) I have felt calm and relaxed. 3) I have felt active and vigorous. 4) I have woken up feeling fresh and rested. 5) My daily life has been filled with things that interest me. 6-point answer scales from 1 ‘all of the time’, 2’most of the time’, 3 ‘more than half of the time’, 4 ‘less than half of the time’, 5 ‘some of the time’ to 6 ‘at no time’.

  14. The factor analysis of five feelings shows one common dimension of emotional well-being. All five indicators are (highly) correlated (Pearson correlations between 0.41 and 0.61) and factor analysis using the ML algorithm confirms adequacy of the one-dimensionality by a KMO of 0.84. Only one Eigen value is larger than 1 (3.10) whilst the second Eigen value is equal to 0.59. Emotional well-being explains 52.8% of the variation and loadings vary from 0.66 (‘my daily life has been filled with things that interest me’) to 0.80 (‘I have felt cheerful and in good spirits’). Cronbach’s α equals 0.84 and cannot be increased by deleting one of the indicators. Following from this, we constructed a scale of well-being including all five feelings. This scale ranges from 1 ‘high emotional well-being’ to 6 ‘low emotional well-being’. The scale has a means of 3.04 and a standard deviation of 1.09.


Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Florian Pichler.



Table A1 Indicators used in multilevel models

Rights and permissions

Reprints and permissions

About this article

Cite this article

Pichler, F. Determinants of Work-life Balance: Shortcomings in the Contemporary Measurement of WLB in Large-scale Surveys. Soc Indic Res 92, 449–469 (2009).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: