CAPITULO 3 De las Ciencias sociales y humanas a las Ciencias Sociales del No –

Table 3.1 The distribution of overall job satisfaction (CERS)

Frequency Percent

Completely satisfied 230 10.79

Very satisfied 730 34.24

Satisfied 831 38.98

Neither satisfied nor dissatisfied 164 7.69

Dissatisfied 117 5.49

Very dissatisfied 26 1.22

Completely dissatisfied 24 1.13

Total 2,132

Note: 10 respondents did not state their level of job satisfaction.

The majority of the respondents to the survey are satisfied with their job (84% report some level of satisfaction; 8% are neutral; 8% report some level of dissatisfaction), the modal group being ‘satisfied’ (39%). The distribution of job satisfaction observed in Table 3.1 is consistent with other British studies and shows that the reported levels of satisfaction are very high (Blanchflower et al 1993, Millward et al 1999, Oswald and Gardner, 2001). These results may reflect a self­ selection effect insofar as workers sort themselves into the jobs that they like and quit those they dislike. However, this explanation over-states workers’ ability to find a suitable job match. There are very few observations in some categories and hence job satisfaction is collapsed into five discrete categories (see below).

A drawback of the WERS is that it does not contain an equivalent question on overall job satisfaction, but instead asks workers about their satisfaction with their pay: “How satisfied are you with the amount of pay you receive?”. An equivalent question is asked in the CERS on workers’ satisfaction with pay How satisfied or dissatisfied are you with your pay?”, and hence the distribution of workers satisfaction

report some level of satisfaction with pay, 14% are neutral, and 27% report some level o f dissatisfaction, with the modal group being satisfied with their pay (46%). In WERS, 35% report some level of satisfaction with pay, 23% are neutral, and 40% are dissatisfied, with the modal group being satisfied with their pay (32%).

Table 3.2A The distribution of satisfaction over pay (CERS)

Frequency Percent

Completely satisfied 59 2.77

Very satisfied 215 10.08

Satisfied 976 45.78

Neither satisfied not dissatisfied 301 14.12

Dissatisfied 403 18.90

Very dissatisfied 101 4.74

Completely dissatisfied 64 3.00

Total 2,119 99.39

Note: Out of the 2,132 respondents in the dataset, for 1 respondent, the question was not applicable, and 12 respondents did not state their level of satisfaction with pay.

Table 3.2B The distribution of satisfaction over pay (WERS)

Frequency Percent

Very satisfied 970 3.43

Satisfied 9,011 31.91

Neither satisfied nor dissatisfied 6,568 23.26

Dissatisfied 7,885 27.92

Very dissatisfied 3,480 12.32

Total 27, 914 98.84

Note: 1.16 percent of respondents (326) did not answer the question, or answered ‘I don’t know’.

Differences between the distribution of the two variables for satisfaction with pay may be due to the fact that CERS samples a different subset of employees, which is in size only 8% of the WERS sample. Additionally, differences may be due to data for the two surveys being collected in slightly different periods of time (WERS: October 1997 - July 1998, CERS: July 2000 - January 2001), and, consequently, due to

Britain experiencing slightly different influences upon its economy, politics and society.25

The CERS and the WERS datasets differ in that the former has a larger proportion o f workers from small firms (i.e. those with less than 10 employees), whereas the WERS excludes establishments with fewer than 10 employees. This means that WERS excludes in its sampling frame 73% of the total o f 1.3 million establishments in the UK (Cully et al, 1999). The critical advantage gained from using CERS becomes even clearer when one considers the considerable bias in the extant economic literature which favours large organisation studies due to easier availability o f data. However, by number of employees, WERS has taken into account for its sampling frame 82% of the total employee population. In conclusion, WERS is more representative of the employee population, whereas CERS is more representative of the establishment population since CERS includes in its sampling frame those workplaces / companies / organisations (these terms are used in CERS interchangeably) with under 10 employees. These two datasets can thus be seen as complements in this analysis. For a more detailed comparison between the datasets, please see Appendix B .l.

As shown in Figure 3.1, WERS includes more large workplaces than CERS, and one could assume that there is a higher scope for more generous pay schemes in large workplaces / firms. In fact, one would expect large firms to be associated with higher levels pay than small firms. This is a stylised fact of the labour market known in labour economics as the firm size wage effect. So, it may be puzzling

to note that workers sampled in WERS are less satisfied with their pay than workers sampled in CERS.

Nevertheless, the firm size wage effect does not imply that satisfaction

with pay is also higher in large firms. For instance, even though workers may receive a higher pay in large firms, relative-pay comparisons within large firms may leave workers feeling “dissatisfied” with a particular level o f pay, whilst that same level of pay may make workers who are employed in small firms rate themselves as “satisfied” with their pay. This is also because there would be a higher likelihood that in a large firm one’s level o f pay was ranked towards the lower side o f within-firm pay distribution. This likelihood would be higher if one were part o f a more numerous workforce, a workforce which is also known to have relatively higher levels o f pay than the rest o f the labour market, such as is the case o f large firms.

Figure 3.1 C om parative distribution of organisations (CERS) and workplaces (W ERS) by employee size (% in sample from CERS and WERS)

_n

m

" '■■111

HU

■

f c H_

i

Descriptive statistics for all variables constructed from both datasets are shown in Appendix B.2 while Appendix B.3 compares the means for all covariates for the CERS and the WERS. Variables for HRM and pay inequality variables are constructed from the questions listed in Appendix B.4. Appendix B.3 allows for comparison between the sample means for the variables extracted from the two datasets. It is clear that the CERS has fewer professional workers and more managerial/intermediate technical workers than the WERS.

To allow a comparison of the findings from these two datasets this chapter uses a set of covariates that are encoded to be as consistent as possible. In terms of HRM practices, the following sets of variables are identified: work organisation, supervision, employee involvement/voice, recruitment and selection, training and learning, and pay practices. In addition, variables for the workers’ perception o f pay

26