An Examination of the Evidence on Sectarianism in Scotland: 2015 Update

Understanding Difference: The Relative Impact Of Religion Compared With Other Key Variables In Relation To Economic Outcomes

As was previously highlighted, we cannot conclude simply from the data about different outcomes that Catholics or indeed Protestants have been systematically discriminated against at either the personal or structural level. This is because of the difficulty in extricating the effect of 'religion' from demographic factors which are likely to have a key role in influencing outcomes. Indeed, many of the findings set out above are likely to be inter-related, with the younger age profile of the Catholic population and higher proportions living in more urban and more deprived locations likely to provide a large part of the explanation for differences in imprisonment, unemployment and victimisation rates. The link between health and deprivation is also well understood, although the direction of the relationship may be reciprocal. In order to investigate the relative impact of these differences on outcomes (as recommended in the 2013 review), we therefore conducted some exploratory analysis using logistic regression looking specifically at the area of economic outcomes (see appendix for detailed models).

Economic activity was selected as a key area for exploration as Catholics have a slightly lower employment rate than those of the Church of Scotland (69% compared with 72% according to the 2011 census) and a slightly higher unemployment rate (8% compared with 6%), with a gap which has persisted over time (evident in the 2001 and 2011 censuses). However, it is worth reiterating that the unemployment rate for Catholics was the same as for those of 'no religion' and close to the 'all people' average figure at 7%. It was also considerably lower than Muslims (who had the highest unemployment rate of all at 12%).

Logistic regression analysis was employed to look at how far religion or more specifically Catholicism is an 'explanatory' factor in terms of predicting the outcome variable^[98] 'in education or employment'. This was classified as being a full-time student or currently employed^[99] (i.e. those individuals who were not unemployed or economically inactive: looking after family, permanently sick, disabled or 'other') over the other key demographic and socio-economic variables. The model was designed to be purposefully simple, in line with the regression literature that explores predictors of employment outcomes (Isengard^[100], 2002; Manley and van Ham 2010^[101]). It therefore allowed key demographic and socio-economic or 'individual' variables (i.e. dichotomised variables such as gender, lone parent, ethnicity and categorical such as health and qualifications), as well as regional factors (dichotomised variable for 15% most deprived area and council area) that affect individual employment prospects, to be controlled for before investigating the impact of including Catholicism as a 'predictor' of being in employment or education. A final model included a dichotomous 'Catholic' and 'non-Catholic' variable^[102], and in order to explore whether there were any generational differences, analysis was applied separately to the following four age groups: 16-24, 25-34, 35-49 and 50-64.

It found that when other key factors were controlled for, the impact of 'Catholicism' was marginal but positive for the youngest age group (16-24), but that overall 'individual' effects have the biggest impact on economic outcomes. These changed only slightly in the regional model but did not change after taking account of Catholicism^[103]. Similar results were found for the 25-34 age group but with a less pronounced positive effect of 'Catholicism'. For the 35-49 age group the Catholic variable was not a useful predictor of economic outcomes (with no effect for the Catholicism variable) and so it might be concluded that there was no significant difference in employment outcomes between Catholics and non-Catholics. In this model, the results again show that individual effects seem to be the most important with a slight, statistically significant effect from including regional effects.

However, in terms of the oldest group (50-64) 'Catholicism' had a negative effect. Amongst those in the oldest group Catholics were more likely to be economically inactive or unemployed compared with non-Catholics. However, as with the youngest age group, the effect of Catholicism was marginal, with 'individual' factors (i.e. gender, lone parent, ethnicity, health and qualifications) having a greater predictive value.

Overall then key 'individual' factors have a large effect on economic outcomes, which do not change greatly when Catholicism is controlled for, and only change slightly after regional effects are accounted for. However, there does appear to be a slight generational difference, with Catholicism found to have a positive, but marginal effect on economic outcomes for young people aged 16-24, but a negative (again marginal) effect for older people aged 50-64.

A number of explanations for these effects are plausible. Firstly, while there may have been some disadvantage (possibly based on discrimination), related to their religion for the older Catholic population, this does not appear to be the case for the younger age groups. In fact, Catholicism has a slight positive effect for the youngest age group once other variables are controlled for. Alternatively, it may be that other religious or vulnerable groups face more difficult socio-economic challenges and/or (worse) discrimination and so it appears as if Catholicism has a positive effect on economic outcomes. However, there are some other interesting observations that can be made from the regression results that may also provide an explanation for this difference, i.e. for younger age groups, residence in Glasgow (which has a large Catholic population) appears to have a positive effect on economic outcomes, whereas for older age groups the opposite is true.

It should be noted that this model does not claim that the younger Catholic population is 'more successful' or indeed 'less successful', than the non-Catholic population. It simply shows that 'individual' effects, and not 'Catholicism', have a stronger association with economic outcomes. Likewise one cannot rule out discrimination more generally. The Catholic population may be disproportionately represented by certain variables. For example, we know that Catholics had a slightly higher proportion of people with no qualifications (30%) than the Scottish average (27%) although Church of Scotland were the group most likely to have no qualifications (35%). It may be that forms of discrimination or other processes lead to a situation whereby the Catholic (or other) population is disadvantaged in certain variables, which, in turn lead to related negative outcomes. Further analysis would allow greater understanding of other socio-economic factors that may play a significant role.

Further breakdowns by religion may also lead to a different picture, as it may be that being of any faith has a positive impact for employment outcomes for the younger age groups. Alternatively, it may be that other groups face more difficult socio-economic challenges and/or (worse) discrimination and so it appears as if Catholicism has a positive effect on economic outcomes. Further analysis of the Scottish-born population may also remove any lingering effects of economic migration on the Catholic coefficient.

Overall then, while it is recognised that there are a number of limitations to the analysis (including the dichotomous categories of Catholic and non-Catholic and the inability to distinguish between high and low paid employment), based on the models as specified, it would seem that individual factors have a greater effect than religion in shaping economic outcomes, with no evidence found to suggest persistent anti-Catholic discrimination.