Belonging Community

The Learning Opportunities Index (LOI) is one of the Toronto District School Board’s key tools for directing resources to the neediest students in the system. Therefore, it’s vital that the index measure deprivation accurately and reliably.

The newly modified LOI dropped less predictive measures of student performance, such as average income, housing type, and immigration status and now includes variables which are better able to measure poverty. Of the new variables, the most powerful are “families on social assistance” and families in the bottom income quartile (as measured by the LIM).

Trustees bite the bullet

So, even though some schools shifted down the ranking and would now potentially lose resources, Trustees (or most of them) bit the bullet and voted to adopt the new instrument.

Still there were some misgivings.

For instance, in terms of external challenges, critical race scholars in the U.S.A. have shown race and poverty have separate effects on student achievement. That, even when income and other demographic characteristics are controlled for, students of different racial identities perform differently within the American school system. This finding has been used, reasonably, as the basis for the creation of Africentric and other race-based schools.

When the new LOI removed the variable of immigration status — often conflated with race in the Canadian context —, the TDSB faced the problem that race, in any form, had been excised. The LOI faced the critique it had been homogenized, to the detriment of its mission of accurately measuring external challenges, and to the detriment, especially, of black students.

So the Board asked the LOI review committee (of which I am a member) to also examine how and whether race should be included in the LOI.

A question for policy wonks or for research geeks

Given the range of views on the question, perhaps the task is really better suited for politicians and policy wonks than for statisticians and research geeks.

However, the review committee has begun its review. We will look at the broader literature, and we will test the utility and strength of any new race-based variable within the Toronto context.

A problem of definition

The first problem has been trying to figure out how to approach the problem.

For instance, producing an accurate description of the term”race” is tricky because race is a social, rather than a biological construction. Its definition and boundaries are blurry and ever–changing. Statistics Canada doesn’t even use the term, but instead says “visible minority” — a bare truth in Toronto — for anyone who has a heritage other than white.

Yet, within the Toronto context, when we compare the performance of “visible minority” students against that of their white peers, there are only subtle differences, sometimes in favour of students of colour. “Visible minority” status alone is not correlated to students’ academic performance. And, that’s a relief. In fact, it’s as it should be.

However, others remind us, we know there are differences between some racial groups.

So we have to explore the term further. Some advocates have been quite clear, we need to stop skirting the issue and name the problem of academic underachievement as one of Black and Aboriginal students, and a few other historically–disadvantaged groups. If we are prepared to do that, academic interventions can be better targetted.

Reliable school–level data

So, if this is the next step, to look at particular racial groups, can we get reliable school-level data? (School–level data is needed to calculate the LOI so that each school can be accurately assessed and ranked in comparison to the others.)

The school board census is the obvious answer. Among its many questions, the TDSB’s student/parent census asked respondents to identify their racial background. However, this won’t work for the LOI.

While useful at a system– or even ward–level, the census data won’t allow reliable comparisons at the school level. For example, some schools had a high non-response rate (students wrote in “Martian” as their answer to the question of their racial background, and various classes never even did the census). The census also happened long enough ago that it no longer supplies a current picture of the Board’s students.

Ranking and weighting races

Ethnic origin might be another usable category from Statistics Canada data, and one which may give more subtlety to the analysis.

Board research has shown that groups of students born in various parts of the world perform differently. Should we parse, weight and rank the value of my children’s English⁄Celtic heritage against their Chinese heritage? (As the discussion unfolds, one can’t help but feel like the evolutionary psychologist University of Western Ontario professor, Philippe Rushton wading into the world of measuring head size to explain intelligence.)

What are we trying to measure? And where does ethnicity blend into culture or language?

And, in the end, does the Board have the stomach to rank one ethnic group against another in the allocation of scarce resources?

Fixed identities

This exercise is different from research which shows different outcomes for students who have already gone through the education system. In this exercise, we are saying that because a student comes from a specific racial background, a priori,  we will award additional resources. We are pre-judging their performance.

The awkwardness of this is that a student’s racial background is different from all the other measures currently used within the LOI because race is fixed. All the other measures, such as parental marital status, education level, and income, can be changed, even re-mediated through social policy and individual effort.

Measuring racism rather than race

Perhaps then, more accurately, this quest to measure the impact of race should be more fittingly seen as a quest to measure racism. We should be measuring the disadvantage which led to the poor environment which created the external challenge some students face. Those who argue for reparations would argue for such.

So, then, the questions becomes, how to measure this.

Use a geographic lens

There is no general “measure of racism” which we can easily access to measure how Toronto students are doing in school. So this is where geography can help. We may well be looking for a measure of concentrated disadvantage or a measure of a neighbourhood peer effect.

Racism creates the inequitable conditions whereby students of colour are more likely live in poor neighbourhoods with low levels of education, fractured families, and little access to good jobs — all variables now included in the LOI and which make it a strong measure of external challenge.

Neighbourhoods may well be the key driver in a student’s performance. And it’s a premise which has some credence.

In 2005, Robert Sampson at Harvard (one of my favourite researchers), investigated the connection between race and violence; he found that the main differences between different racial groups’ levels of violence were explained by demographics and neighbourhood conditions. He recommended that interventions which “improved neighbourhood conditions and support families” would be the most effective way to reduce violence.

Sampson also found that neighbourhood distress was inversely related to the number of workers in professional occupations and the proportion of married parents. Higher levels of recent immigrants also had a dampening effect on violence. Tom Carter, at the University of Winnipeg, has cited research supporting similar conclusions in his studies on the inner city.

In effect, what looked like racial differences were actually problems rooted in poverty and deprivation.

Furthermore, an American study found that while racial segregation has been declining, educational segregation has increased. So neighbourhoods are more divided along, arguably, class lines than racial ones. (I don’t know of a similar study in a Canadian urban centre.)

More to thresh out

In the end, what seemed like an easy question may have a complex answer.