The second main group of analyses was designed to examine how exposure to the four different types of workplace risk changed between 2005 and 2010. In this analysis we included the 30 countries which were included in both the 2005 and 2010 rounds of the EWCS. We might expect substantial change as a result of the recession, though it is difficult to anticipate in advance the direction of change when it comes to workplace risks. On the one hand, an economic downturn might be associated with a disproportionate loss of jobs in more marginal employment characterised by unsafe working conditions. This might result in an improvement in the risk profile of the remaining jobs. On the other hand, economic pressures to cut costs may result in cuts to health and safety procedures and personnel, leading to an increased exposure to risk.
In this analysis we first asked whether there was a significant change in the level of exposure to each time of risk and whether this change differed by country. In the second part of the analysis, we asked whether there was a significant change over time in the association between risk and the characteristics of jobs and workers. In other words, was there any evidence that certain types of jobs were becoming more risky or less risky.
5.3.1 Overall Change in Exposure to Risk
In general across the 30 countries, we saw a small fall in exposure to physical risk, which was mainly accounted for by shifts in the composition of jobs and workers, but an increase in exposure in three countries: Ireland, Luxembourg and Turkey. There was no clear association between the changes in exposure and the extent to which the countries were hit by the recession. In spite of the increase in Ireland, the level of exposure to physical risk in 2010 remained relatively low.
There was also a general fall in exposure to chemical/biological risk, including in Ireland. About half the countries, including Ireland, experienced no change in the level of exposure to physically demanding work, while most of the remainder experienced a reduction in exposure.
There was a more substantial fall in the level of exposure to psycho-social risk, especially in countries like Finland and the Netherlands where levels had been highest in 2005. The fall in Ireland was not significantly different from the fall in most countries.
Note that these figures refer to changes over time in exposure to a range of different types of workplace risk rather than to the outcome for workers. An analysis of trends in workplace illness and injury in Ireland, based on an analysis of the Quarterly National Household Survey module on Health and Safety in the Workplace, found that the rate of workplace illness and injury fell between 2001 and 2012 (Russell et al., 2015). The authors find that the sharpest reduction in the injury rate occurred between 2008 and 2009 and was associated with the change in composition of jobs and workers with the recession, including the fall in employment in construction. The results here pointed to a mixed pattern in terms of exposure to workplace risks in Ireland, with a small increase in exposure to physical risk; little change in exposure to physically demanding work; a small fall in exposure to chemical/biological risk and a more substantial fall in exposure to psycho-social risk.
In terms of the overall change over time, then, there was a small fall in the exposure to physical and chemical/biological risk and to physically demanding work, while there was a more substantial fall in the level of exposure to psycho- social risk. Ireland differed from this general pattern with respect to physical risk (which increased slightly in Ireland) and physically demanding work (no change over time in Ireland).
5.3.2 Change in Correlates of Exposure to Risk
A second aspect of the research question regarding change over time concerned whether there was a general shift in the association between workplace risk and characteristics of jobs and workers. A small fall in exposure overall, for instance, might mask a more substantial fall in some industries and an increase in others. In general, most of the associations that were present in 2010 did not differ significantly from those found in 2005. So as well as examining whether the
change differed by country, as described above, we examined whether the change differed by characteristics of jobs and workers. This analysis was conducted across the 30 countries present in the 2005 and 2010 datasets.
We found some changes over time across the 30 countries in the association between exposure to physical risk and job or worker characteristics. The relative advantage experienced by some groups had narrowed. For instance, workers in the financial, real estate and business services sector and workers with third-level education were less distinctive than the reference categories (retail/whole sector and workers with higher second level education, respectively). On the other hand, the gap between managers and officials, on the one hand, and unskilled workers in elementary occupations had increased. By 2010, the unskilled workers had a level of exposure to physical risk that was 0.9 points higher than managers and officials on the ten point scale, compared to a gap of 0.6 points in 2005. Unskilled workers in elementary occupations also experienced an increase in exposure to chemical/biological risk compared to managers and officials, with the gap increasing from 0.6 points in 2005 to 0.8 points in 2010 on the ten point scale. Apart from this increase in the disadvantage faced by unskilled workers, the other significant changes over time represented a weakening of the association between job or worker characteristics and exposure to chemical/ biological risk. As with physical risk, the advantage experienced by those with degree-level education narrowed over time in terms of exposure to chemical/ biological risk. There was a weakening of the differences between economic sectors. The higher risk of exposure among those in agriculture, forestry and fishing in 2005 had disappeared by 2010 and the higher risk among those in the construction sector had become less pronounced. Where recently-hired workers in 2005 had a higher risk of exposure to chemical/biological risk than workers with more than ten years tenure, this was no longer the case in 2010.
We also saw a weakening of the association between certain characteristics of jobs or workers and the level of exposure to physically demanding work. The gap between officials/managers and certain other occupations narrowed (technicians and associate professionals; craft and trades workers); the difference between those working full-time and those working part-time (who had a higher level of exposure in 2005) narrowed; the difference between younger workers, who had a higher level of exposure in 2005, and older workers also narrowed.
In terms of exposure to psycho-social risk, apart from an increase in the risk of exposure among the self-employed,17 there was a narrowing of other differentials
with a larger fall in many public sector jobs that had a high level of exposure in 2005. Compared to the retail and wholesale sector, the higher levels of exposure experienced by workers in education and in public administration and defence had disappeared, while the higher level experienced by workers in the health and social work sector had been reduced. The advantage experienced by workers in the financial, real estate and business service sector had also disappeared. The association between long hours and exposure to psycho-social risk had become non-significant.
It is worth repeating that the general pattern of association between job and worker characteristics across the 30 countries remained unchanged over time. Of the more than 50 contrasts we examined for each of the four types of risk, only a handful were different in 2005 from the pattern in 2010. More often than not, the changes took the form of a narrowing of differentials between groups of workers. The exception, of course, was the increase in exposure to physical and chemical/biological risk among unskilled workers in elementary occupations.