This paper explores the link between employee perceptions of working conditions and the desire for worker representation in Britain and the US. We find that the distribution of employee perceptions of poor working conditions is similar in Britain and the US; similar factors affect the number of perceived poor working conditions; and the perception of poor working conditions is strongly associated with the desire for union representation. The nature of workplaces, as opposed to employees’ characteristics, is the predominant factor determining employee perceptions of poor working conditions.
This is a preview of subscription content, log in to check access.
Buy single article
Instant access to the full article PDF.
Price includes VAT for USA
Subscribe to journal
Immediate online access to all issues from 2019. Subscription will auto renew annually.
This is the net price. Taxes to be calculated in checkout.
Farber and Krueger (1992) argued that falling interest in unions contributed to the 1980s decline in density. If desire for unionisation is constant over time, declines in union density should raise the proportion of non-organized workers who want unionism. Since the percentage of non-union workers saying they would vote for a union remained roughly constant over the period they examined they attributed some of the decline to loss of interest in unions. The increase in unfilled demand for unionisation in the 1990s through to 2005 runs against this story.
The probability of worker selection is the product of the probability of the workplace being selected and the probability of an employee being selected from within that workplace. The survey includes weights to allow the analyst to run population estimates. This weighting scheme compensates for sample non-response bias in the employee survey (Airey et al. 1999: 91–92). We test the sensitivity of our results to weighting the data.
Full details of the factor analyses conducted on all three poor working conditions scales are available from the authors on request.
An alternative approach is to use the full information in the distribution of responses on each question so that the additive scale weights responses according to how much of a problem it was. Using a simple scale with 1 to reflect the lowest possible response to poor conditions, 2 to reflect the next level, and so on, we formed a summated rating of these responses and obtained results that parallel those in the paper.
We regressed ln (the % reporting number of poor conditions) on the number of poor conditions to fit an exponential distribution and regressed ln (the % reporting number of poor conditions) on ln (the number of poor conditions) to fit a power law. The exponential fits the BWRPS and WRPS better than a power law, whereas the power law fits better for the WERS.
The mean values and standard deviation for each independent variable from the three surveys used in the analyses are presented in the data appendix.
Union membership is not exogenous with respect to perceptions of poor working conditions, of course. We are not making any claims about the nature of causal linkages between membership status and perceptions of working conditions.
The maximum workplace mean of poor perceived conditions in the data is 10.66 from a possible 13. The mean value for the workplace mean is 3.10.
Specialists are designated according to their job title.
The model for non-members is perhaps of most interest given unions’ objective of reaching out to new members. The model accounts for roughly one-quarter of the variance in the desire for union representation among non-members. Desire for union representation rises with age initially but then tails off; it is greater among non-white ethnic minorities, those with higher qualifications, and those in lower occupations. It falls with tenure and wages.
Full models for Table 4 and the BWRPS analysis are available on request.
In these models, available from the authors on request, a simple 9-item count variable containing poor working conditions other than the four dissatisfaction variables, was positive and statistically significant having conditioned on the job dissatisfaction count variable and all the other variables in Table 4. The coefficients and t-stats were 0.025, t = 5.86 for all employees, 0.017, t = 2.41 for members, and 0.29, t = 5.48 for non-members.
As in the British case, the number of poor working conditions was significantly positively correlated with voting union conditioning on job dissatisfaction.
Airey C, Hales J, Hamilton R, Korvovessis C, McKernan A, Purdon S (1999) The Workplace Employee Relations Survey (WERS) 1997–8: technical report. National Centre for Social Research, London
Brown A, Charlwood A, Forde C and Spencer D (2006) “Changing job quality in Great Britain, 1998-2004”, Department for Business, Skills and Innovation, Employment Relations Series No. 70, London
Bryson A, Cappellari L, Lucifora C (2004) Does union membership really reduce Job satisfaction? British Journal of Industrial Relations 42(3):439–459
Bryson A, Cappellari L, Lucifora C (2010) Why So unhappy? The effects of unionization on Job satisfaction. Oxford Bulletin of Economics and Statistics 72(3):357–380
Budd J (2004) Employment with a human face: balancing efficiency, equity and voice. Cornell University Press, Ithaca
Charlwood A and Green F (2012) “Job quality and employee wellbeing”, http://cep.lse.ac.uk/seminarpapers/WB-11-05-11GREEN.pdf
Cully M, Woodland S, O’Reilly A, Dix G (1999) Britain at work: as depicted by the 1998 workplace employee relations survey. Routledge, London
Diamond W, Freeman RB (2002) What workers want from workplace organisations. Trade Unions Congress Publications, London
Edwards, R. E. (1979) Contested terrain: the transformation of the workplace in the 20th Century, Basic Books
European Foundation for the Improvement of Living and Working Conditions (2010) Changes over time – First findings from the Fifth European Working Conditions Survey http://www.eurofound.europa.eu/pubdocs/2010/74/en/3/EF1074EN.pdf
Farber HS and Krueger AB (1992) Union membership in the United States: the decline continues, NBER Working Paper No. 4216
Fiorito J (2001) Human resource management practices and worker desires for union representation. Journal of Labor Research 21(2):335–54
Freeman RB (1978) Job satisfaction as an economic variable. American Economic Review 68:135–141
Freeman RB and Medoff J (1984) What Do Unions Do, Basic Books
Freeman RB, Rogers J (1999) What workers want. Cornell University Press, Ithaca
Freeman RB, Boxall P, Haynes P (2007) What workers Say: employee voice in the Anglo-American workplace. Cornell University Press, Ithaca
Goldberg D, Williams PA (1998) User guide to the general health questionnaire. NFER-Nelson, Windsor
Green F (2006) Demanding work. The paradox of job quality in the affluent society. Princeton University Press, Princeton
Hackman JR (1987) The design of work teams. In: Lorsch J (ed) Handbook of organizational behavior. Prentice-Hall, Englewood Cliffs
Hammer TH, Avgar A (2005) The impact of unions on Job satisfaction, organisational commitment and labor turnover. Journal of Labor Research, XXVI 2:241–266
Kaufman B (2010) The theoretical foundation of industrial relations and its implications for labor economics and human resource management. Industrial & Labor Relations Review 64(1):74–108
Kelly J (1998) Rethinking industrial relations: mobilisation, collectivism and long waves. Routledge
Klandermans B (1986) Psychology and trade union participation: joining, acting, quitting. Journal of Occupational Psychology 59(1):189–204
Marx K (1867) Capital: critique of political economy. Verlag von Otto Meisner
Maslow AH (1943) A theory of human motivation. Psychological Review 50:370–96
Towers B (1997) The representation gap: change and reform in the British and American workplace. Oxford University Press
Townsend P (1979) Poverty in the United Kingdom: a survey of household resources and standards of living. University of California Press
Data Appendix: Employee Perceptions of Poor Working Conditions
Data Appendix: Employee Perceptions of Poor Working Conditions
There are a variety of ways in which one might wish to construct an index of the poor working conditions employees face at the workplace. Traditionally, analysts have contented themselves with a single measure, such as an overall job satisfaction measure or perception of the climate of employment relations. However, single measures suffer from a number of drawbacks. First, any single item will only capture a part of an underlying multi-dimensional concept such as poor conditions at work. Measuring such a concept across various dimensions helps reduce the degree to which a proxy for poor conditions suffers from this type of measurement error. Second, there is no reason, a priori, why one should give precedence to one facet of poor conditions. Third, particular problems fall or rise in salience depending on external factors, such as the point in the business cycle. For instance, workers may be less likely to cite problems with pay satisfaction when the labor market is tight and employers are having to meet fairly large wage demands to attract and retain workers. A multi-item index can ‘smooth’ these idiosyncracies so giving a potentially more accurate measure of poor conditions. Another big advantage of a multi-item scale scored as a fraction of the total possible number of poor conditions recorded is that it offers a possibility of making comparisons across surveys where the specific survey questions are not identical.
Having chosen a multi-item scale, one needs to consider what enters the scale and how items should be added together. We were constrained in the items available to us since the surveys had already been undertaken. Fortunately those for both the USA and Britain contained items relating to key domains, notably the climate at the workplace, ‘gaps’ in influence between what workers had and what they wanted, satisfaction with various aspects of their jobs, ratings of management, and so on. We identified the cut off for “poor” conditions in the way described below. Results were not sensitive to whether we used the full distribution of answers to a particular item, or simply entered it as a dummy (0,1) variable where 1 = a poor condition. One might consider giving greater weight to some items than others in an additive scale, but there were no a priori reasons for doing so. Inter-item correlations were generally positive suggesting adding items together was not an unreasonable strategy. The Cronbach alpha for all items in the BWRPS was .80, while the alpha for those in WRPS was .64.
There are difficulties running principal components analyses on the items entering the scales because some questions were randomly assigned to sub-samples. Consequently, whereas one can readily add up scores to similar questions to form a scale, principal components analyses would have to be run on the sub-samples asked the same set of questions.
The BWRPS poor working conditions scale runs from 0 to 23. All items in Appendix Table 1 are 0–1 dummy variables apart from the last one, which is an “influence gap” running from 0–3. The individual is asked “How much direct involvement and influence do you have in....deciding how to do your job and organise work; setting working hours including breaks, overtime and time off; deciding how much of a pay rise the people in your work group or department should get; the pace at which you work; deciding how to work with new equipment or software; deciding what kinds of perks and bonuses are offered to employees”. Answers are coded on a four point scale from “a lot” to “none”. For the same items respondents are then asked “tell me how important it would be to you to have a lot of influence over…” with responses coded on a four point scale from “very important” to “not at all important”. When a person views having “a lot of influence” as “very important” but does not have “a lot of influence” this is identified as a poor working condition. Because respondents are randomly split into two groups, both of whom answer only three of the six items, respondents can score between 0–3 on the influence gap scale.
The WRPS is similar to BWRPS because the latter was partly based on the WRPS. The influence gap is constructed in a similar fashion, too, with 1 point added to the scale every time a person views having a lot of influence as ‘very important’ but does not have a lot of influence up to a maximum of 4. Some were randomly routed to an alternative question about satisfaction with influence on these items and so these people scored 1 point every time they expressed dissatisfaction with their influence.
About this article
Cite this article
Bryson, A., Freeman, R.B. Employee Perceptions of Working Conditions and the Desire for Worker Representation in Britain and the US. J Labor Res 34, 1–29 (2013). https://doi.org/10.1007/s12122-012-9152-y
- Working conditions
- Worker representation
- Trade unions