• Keine Ergebnisse gefunden

Estimation results

Im Dokument RESEARCH REPORT 221 (Seite 89-98)

WORKPLACE INJURIES AND ETHNIC MINORITIES: AN ANALYSIS OF THE LABOUR FORCE SURVEY

5.4 THE PROBABILITY OF REPORTING WORKPLACE ACCIDENTS

5.4.2 Estimation results

Figures 5.1 to 5.3 present results from logistic models that incorporate variables to distinguish individuals by ethnicity and length of residence in the UK. These characteristics are shown as sets of categories, with one category excluded in each set in order to act as the reference. The first set of variables distinguishes ethnicity, with ‘whites’ acting as the reference category. The second set of variables distinguishes migrant status, with ‘born in UK’ acting as the reference category. Where the bar drops below the horizontal axis, this indicates that individuals within that category exhibit a lower probability of reporting a workplace accident compared to the reference category. Where it lies above, this indicates a higher probability of reporting a workplace injury compared to the reference category. Shaded bars indicate whether the relationship was estimated to be statistically significant at the 5% level.

In terms of ethnicity, it is estimated that Indians and Pakistanis are less likely than Whites to report that they had experienced a workplace accident during the previous 12 months. This result is estimated to be statistically significant among both Indians and Pakistanis for all accidents (respectively 31% and 43% less likely), excluding road accidents (33% and 53% less likely) and reportable accidents (26% and 48% less likely). No statistically significant differences are estimated for Black Caribbeans, Black Others (incorporating Black African, Black Other, Black Mix) or Chinese workers, although there is weak evidence to suggest that workers within the Black Others category have an increased probability of reporting a workplace accident. In terms of length of residency in the UK, it is estimated that individuals not born in the UK are less likely to report the occurrence of a workplace injury. It appears, however, that as these individuals spend longer within the UK, their propensity to report the occurrence of a workplace injury during the 12 months prior to the survey approaches that of individuals who were born in this country. This is highlighted by those workers who have resided in the UK for longer than 10 years having very similar reporting behaviour compared to those who were born within the UK.

The reduced likelihood of Asians and those born outside of the UK to report the occurrence of a workplace accident does raise concerns about the quality of information collected from such respondents. A potential source of error in the LFS is that information related to approximately one third of individuals is collected through a proxy respondent. While this information is generally regarded as being of acceptable quality, it is possible that proxy respondents may under-report the occurrence of certain events (see Dawe and Knight, 1997). This could be of particular importance for the reporting of workplace injuries that requires respondents to recall events over a 12-month period. Figures 5.4 to 5.6 therefore consider whether there is a tendency for workplace injuries to be under-reported by proxy respondents.

All Accidents - by ethnicty and length of residence

Figure 5.1 Probability of reporting an accident by ethnicity and length of residence

Excluding Road Accidents - by ethnicty and length of residence

i

black other ndian pakistan chinese other

Reatsk (%

Figure 5.2 Probability of reporting an accident (excluding road accidents) by ethnicity and length of residence

Reportable Accidents - by ethnicty and length of residence 40.00

ack other ndian pakistan chinese other

Relative Risk (%) 0.00

Figure 5.3 Probability of reporting a ‘reportable’ accident by ethnicity and length of residence

All Accidents - Influence of Proxy Response

proxy-abroad wh te-direct wh te-proxy ack-direct black-proxy asian-direct asian-proxy

Reatsk (%

Figure 5.4 Probability of reporting an accident: effect of proxy response

Excluding Road Accidents - Influence of Proxy Response

di i i bl

proxy-abroad wh te-direct wh te-proxy ack-direct black-proxy asian-direct asian-proxy

Reatsk (%

Figure 5.5 Probability of reporting an accident (excluding road accidents): effect of proxy response

Reportable Accidents - Influence of Proxy Response

i i asi

white-d rect white-proxy black-d rect black-proxy asian-direct an-proxy

Relative Risk (%

Figure 5.6 Probability of reporting a ‘reportable’ accident: effect of proxy response

Considering proxy response by ethnicity for all accidents, it is estimated that proxy respondents reporting on behalf of Whites are 30% less likely to report the occurrence of a workplace injury compared to Whites who respond to the LFS directly. In comparison, it is estimated that proxy respondents reporting on behalf of Black (incorporating Black Caribbeans, Black Africans and Black Others) are approximately 49% less likely to report the occurrence of a workplace injury compared to Black who respond to the LFS directly. Therefore, there is possible cause for concern regarding the quality of information collected via proxy respondents relating to workplace injuries among Black. This is particularly important given that Black Caribbeans and Black Others were estimated to have an increased probability of reporting a workplace accident compared to whites, although not significantly so (see Figure 5.1 to 5.3). Under-reporting among proxy respondents may be disguising the actual extent of any increased risk exhibited by this group.

However, this is in contrast to Asians. Considering the probability of reporting all workplace accidents, proxy respondents reporting on behalf of an Asian are only approximately 15% less likely to report the occurrence of a workplace injury compared to Asians who respond to the LFS directly. Under reporting by proxy respondents therefore does not appear to explain the reduced risk of workplace injury among Asians. Similarly, there is no evidence to suggest that injury data collected through proxy respondents on behalf individuals born outside of the UK is contributing towards the lower risk of workplace injuries estimated for this group. The relative quality of data from proxy respondents does not appear to be the cause of the lower probability of workplace injuries among Asians, or among those not originally born in the UK. However, it is important to note that the lower estimated risks of workplace injury could be the result of a general under-reporting of workplace injuries among such groups, among both proxy respondents and those who respond directly to the LFS.

Another source of potential bias in terms of the under-reporting of accidents among ethnic minority or immigrant workers may be related to establishment size. Available empirical evidence tends to support the view that larger establishments exhibit lower rates of workplace injury (e.g. Currington, 1986, Lanoie, 1992). It is suggested that this pattern is indicative of the economies of scale achieved by larger establishments in terms of expenditure on health and safety. The average cost of health and safety measures per employee will be lower in larger firms, increasing the incentive of firms to invest in these areas. Previous analysis of the LFS however indicates that the risk of injury is lower within smaller establishments with fewer than 25 employees compared to larger establishments (Elias et al, 2001). The inability to distinguish between medium and larger sized establishments within the LFS means that it has not been possible to identify effects that may be attributable to economies of scale at larger establishments (see Nichols, Dennis and Guy, 1995).

The lower probability of workers in small establishments reporting a workplace injury could be attributable to safer behaviour at small establishments, where the operation of the business would be more adversely affected by the absence of a member of staff. Alternatively, these findings may indicate that such businesses are less able to cope with unplanned periods of absence, encouraging a faster return to work following a workplace injury. Finally, small firms may be characterised by a lack of awareness among both employees and employers regarding health and safety responsibilities. Whilst the lack of regulatory control at such workplaces may have adverse level upon actual workplace safety, workers in such establishments may have a lower propensity to report workplace accidents. The relationship between establishment size and workplace injuries may differ further when focussing upon ethnic minorities or migrant workers as such workers may be more likely to work in family run businesses. This may have an affect both upon workplace behaviour and upon how such workers may respond to questions regarding workplace injuries.

Figures 5.7 to 5.9 therefore consider the relationship between establishment size and the probability of reporting a workplace accident among different ethnic groups and by migrant status.

Relative Risk (%)

Figure 5.7 Probability of reporting an accident: effect of establishment size

Excluding Road Accidents - Establishment Size

Figure 5.8 Probability of reporting an accident (excluding road accidents): effect of establishment size

Figure 5.9 Probability of reporting a ‘reportable’ accident: effect of establishment size

A common theme emerges across these figures, with workers at smaller establishments having a lower probability of reporting a workplace injury. Considering all workplace accidents (see Figure 5.7), it can be seen that among those born in the UK, workers in establishments with 11­

24 employees are 19% more likely to report a workplace accident compared to those in establishments with 1-10 employees. UK born workers in establishments with 25 or more employees are 26% more likely to report a workplace accident compared to the reference category. Considering those workers born outside the UK, workers in establishments with 11­

24 employees are 18% more likely to report a workplace accident compared to those born outside the UK and working in establishments with 1-10 employees. Immigrant workers in establishments with 25 or more employees were also 26% more likely to report a workplace accident compared to those born outside the UK and working in establishments with 1-10 employees. A similar pattern emerges when excluding road accidents and for reportable accidents. The lower probability of workers who were born outside of the UK reporting a workplace accident is observed across all establishment sizes and is not due to a disproportionately lower probability of reporting within establishments of a particular size.

Considering different ethnic groups, it is observed across all accident types that White and Asian workers at smaller establishments have a lower probability of reporting a workplace injuries. The lower probability of Asian workers reporting a workplace accident is observed across all establishment sizes and is not due to a disproportionately lower probability of reporting within establishments of a particular size. The positive relationship between establishment size and the probability of an individual reporting a workplace accident is however not observed among Black survey respondents. Although not statistically significant, Black workers in establishments with 1 to 10 employees are estimated to have a 20% higher probability of reporting a workplace accident compared to White workers at establishments of similar size. The lower probability of reporting an accident within that is observed among white workers and Asian workers in smaller establishments does not emerge among black workers.

The lower likelihood of ethnic minorities reporting a workplace injury does not appear to be attributable to a disproportionately lower probability of reporting a workplace injury among ethnic minorities and migrant workers employed within small establishments.

We next consider the issue of under-reporting by making the distinction between those people employed in the public and private sector. Utilising data from the LFS and having controlled for a variety of personal, establishment and job related characteristics, Elias et al (2001) estimate that workers in the public sector have approximately a 30% lower probability of experiencing a reportable workplace injury. This finding could be due to otherwise unobserved differences in the characteristics of those employed in the public and private sector that makes these workers differ in their susceptibility to experiencing a workplace injury. This finding could also be due to public sector workers taking more time off work as the result of a workplace injury. Alternatively, public sector workers may be more likely to report accidents because of the presence organisational regulations that are clear in identifying what constitutes a workplace injury and which encourage the reporting of these accidents. While information collected about workplace injuries in the LFS is not dependant upon whether an injury was actually reported under RIDDOR, the effect of higher reporting levels in the public sector may spill over into responses to the LFS. Figures 5.10 to 5.12 therefore consider the probability of reporting a workplace accident among different ethnic groups and by migrant status according to whether these workers are employed in the public or private sector.

All Accidents - Public/Private Sector 40.00

30.00

20.00

10.00

0.00

home- home-public abroad- abroad- white-private white-public black-private black-public asian- asian-public

private private public private

Relative Risk (%)

-10.00

-20.00

-30.00

-40.00

-50.00

Figure 5.10 Probability of reporting an accident: distinguishing the public/private sector

Excluding Road Accidents - Public/Private Sector 30.00

20.00

10.00

home- home-public abroad- abroad- white-private white-public black-private black-public asian- asian-public

private private public private

Relative Risk (%) 0.00 -10.00

-20.00

-30.00

-40.00

-50.00

Figure 5.11 Probability of reporting an accident (excluding road accidents):

distinguishing the public/private sector

Reportable Accidents - Public/Private Sector 40.00

30.00

20.00

home- home-public abroad- abroad- white-private white-public black-private black-public asian- asian-public

private private public private

Relative Risk (%) 10.00 0.00

-10.00

-20.00

-30.00

-40.00

Figure 5.12 Probability of reporting a ‘reportable’ accident: distinguishing the public/private sector

Considering the results for all accidents presented in Figure 5.10, it can be seen that among those born in the UK, workers in private sector establishments are 23% less likely to report a workplace accident compared to those in the public sector. Considering those workers born outside the UK, workers in private sector establishments are 31% less likely to report a workplace accident compared to those born outside the UK and working in the public sector. This wider difference among migrant workers however does not emerge when excluding road accidents and for reportable accidents. The results do not appear to be indicative differences in reporting levels among migrant workers in public and private sector establishments.

The patterns are less clear when we consider the relationship between ethnicity and the probability of reporting a workplace injury by sector. Considering all workplace injuries, white public sector workers are estimated to be 23% more likely to report a workplace injury than private sector workers. Black public sector workers are estimated to be 34% more likely to report a workplace injury compared to Black private sector workers. Similarly, Asian public sector workers are estimated to be 32% more likely to report a workplace injury compared to Asian private sector employees. However, these wider difference among ethnic minorities do not emerge when excluding road accidents and for reportable accidents only. When excluding road accidents, the difference in the probability of reporting a workplace injury between public and private sector employees is almost identical among white, Black and Asian workers. The increased probability of white public sector workers incurring a reportable workplace injury (32% more likely) is not observed among Black and Asian workers.

CHAPTER 6

Im Dokument RESEARCH REPORT 221 (Seite 89-98)