Complexity to inform issues of methodology, taxation and politics in a context of crisis

This blog is not a review of my former colleague Keming’s book so much as a reflection on what it made me think about. However, I should at least notice it first, so here goes.  To quote the publisher’s blurb:

How can we analyse the intersectional effects of multiple factors on experiences of disenfranchisement? This book equips you with the methodological tools to uncover new insights.
First providing a critical examination of long-standing methodologies in intersectionality research, it then shines a spotlight on analytical techniques such as qualitative comparative analysis, multilevel models, mediation and moderation, and mixed methods designs.
With chapter objectives, real-world research examples, further reading and reflective questions, it will equip you with the methodological tools to understand intersectionality in specific social settings.
The book:

·       Bridges the gap between intersectionality as a theory and an empirical research practice.

·       Extends existing approaches to analysing intersectionality in a traditionally qualitative field.

·       Inspires creativity and celebrates a variety of effective methods for studying intersectionality.

Innovative and thought-provoking, this book is ideal for any student or researcher looking to harness the power of empirical evidence to explore inequality and injustice.

I agree with all of this.

Let me quote Keming’s view on what is the core notion of intersectionality as deployed in the book:

Intersectionality among two or more social attributes such as race, ethnicity, gender, sexuality, class, nationality etc. are associated with the observed social inequalities because existing social structure and processes interact with one another in distributing the variable resources across thesesign cross-classified groups unevenly. (page 28)

Note the use of the word interaction. He continues:

The theory of intersectionality assumes that all these power systems depend on, interact with, and mutually construct each other, and their interactions are more significant that each category or system alone. Put in statistical jargon, the effect of the interaction terms must outweigh the effect of each individual system alone. Most importantly, intersectionality as a critical social theory does not treat this as a statement to be verified with empirical evidence, it holds the statement as necessarily true…(pages 29-30).

I agree and really consider that only quantitative descriptive data can provide that empirical verification although all modes of qualitative inquiry can provide experiential accounts of its implications. Actually, on reflection I think that historical on documentary sources, with documents understood in the widest possible sense, can provide an empirical basis for asserting the reality of the processes which create intersectional inequality.  Look at the set of quotations in this link to see how they demonstrate the role of discrimination against catholics in Northern Ireland over most of the history of that statelet and that discrimination was intersectional in relation particularly to class: https://cain.ulster.ac.uk/issues/discrimination/quotes.htm

I will return to Northern Ireland subsequently.

There is a lot of good stuff in the book in terms of discussion of a range of primarily quantitative modes of empirical demonstration.  I will draw particular attention to Chapter Eight on Mediation, Moderation and Intersectionality. As Keming says a lot of models ‘… do not distinguish the attributes’ role in a logical orb temporal sense when affecting the outcome’ (page 117) His discussion of the value of multicollinearity is excellent. I will simply note that for me a good way of dealing with logical and sometimes even temporal ordering in causality is by the use of time ordered partial correlation coefficients. When there are longitudinal data sets both can be deployed in relation to cohort relative effects as was done in The Poverty of Education which I discussed in one of my life in methods blogs.  You need cohort effects to get quantitative data outcomes.

Keming has a very sensible and measured discussion of his own situation in relation to interaction effects. I will present my own with reference to my lived experience in relation to the implications of having been brought up in the North East of England rather than Northern Ireland where I lived (to be more fully discussed in a subsequent life in methods blog) in my early 30s during the later 1970s. My family background is Tyneside Irish in that my father’s family was entirely composed from Irish immigration in the late nineteenth century to Tyneside through his grandfather and father who arrived as immigrants and became coal miners. My mothers’ grandfather did the same although her father was a Welsh seaman who married the Irish miner’s daughter he met when taking his first mate’s ticket at the South Shields marine school. The NE of England has no meaningful history of religious sectarianism – class has always been the dominant source of  identity, in a complex interaction with gender. Actually part of my father’s family came from County Derry in Northern Ireland but before partition, My mother through sheer ability, although not allowed to sit the scholarship exam for entry to a secondary school in the 1930s because of the racing certainty she would pass – she was top of the class at ten in a class of eleven year olds – and her mother’s (a war widow) fear of the expense of uniforms and bus fares, got into the civils service in the revenue and rose through competitive examination through the clerical and executive grades to be on the cusp of promotion to Inspector of Taxes when her marriage and my birth knocked her out and she became a primary school teacher. My father as the youngest son in a mining family was able to go to Grammar School but after a brief and unsuccessful time as a miner, knocked about as a post office clerk, working in a chocolate factory and orderly in a psychiatric hospital until his war service as a field ambulance private in the front line in the RAMC in North Africa. He got into Medical School on a postwar scholarship and eventually became a GP although not until my parents had separated and subsequently divorce so I was in effect the son of a single parent school teacher (even when she was supporting her second husband through his medical school years she was the wage earner).

If I had grown up on in Northern Ireland my situation in class terms  would been much the same but my intersectional relationship with a catholic background would have been completely different.  I would as a University student have taken part in the Civil Rights movement in Northern Ireland (which I supported from Britain at the time), been beaten up by loyalist thugs at Burntollet Bridge on the march to Derry, and almost certainly have joined the IRA – first the Official and then the INLA after the split when the Officials gave up the armed struggle. My Catholic background had no significance in the NE of England. In Northern Ireland it would have determined most aspects of my life. Context always matters in relation to the causal powers of interaction.

I have a few of criticisms of this book. First, n the discussion of the value of Qualitative Comparative Analysis as a mode of exploration of the causal powers of intersectional / interaction effects Keming rightly says that when we use lots of attributes then the process becomes difficult to the point of unmanageability. There is a way round this. Instead of using single attributes measured at any level it is possible to cluster sets of related attributes using Two Step Cluster in SPSS. This is a useful mode of data reduction but to my mind it also reflects social reality in that we can cluster to generate more meaningful complex attributes.  This is a way of saying that interaction / intersectionality is embodied in real cases and can be used to relate aspects of the case to outcomes.

Second, Keming counterposes interpretive modes of qualitative reasoning to positivist approaches. Realist approaches, especially complex realist approaches, enable us to reconcile understanding based on both quantitative and qualitative research, provided  both are understood as exploratory in essential form. To be sure “the church of critical realism” is uncomfortable with quantitative materials but we must remember that innumeracy can never be an epistemological principle.

Finally, and this reflects the general tenor of discussion of interaction / intersectionality, there is not enough on class position as key attribute set. This is particularly important when the inane culture wars, inane on both sides of them, assert identities as what matters and wilfully disregard class positions. “White Privilege” is an extreme version of this nonsense. Just to show how all causal processes are embedded in changing historical context let me address the situation of white working class protestant young men in Northern Ireland. Historically this group had an effective monopoly on access to good industrial jobs but a combination of deindustrialization and effective legislative outlawing of discrimination (the loss of industry is much more important) have eliminated that intersectional advantage. The separate catholic school system across all secondary school types is much more effective than protestant state non selective schools at delivering a good education which in a labour market dominated by public sector jobs puts protestant working class boys at a significant disadvantage. An intersectional / interactional advantage in relation to employment outcomes has become a disadvantage.  More generally being a rich bourgeois of any sort eliminates all identity based disadvantages.

To conclude, this is an excellent book written by a very good teacher and is both a model of thinking and a  great text to use with students.