This is the second in a series of posts reflecting on how intergroup relations can be influenced by factors operating on different time-scales (including events from the distant past), with particular attention to the influence of social institutions. The first post is here.
Considering how long rhythms from the past might affect current day intergroup dynamics, we have recently begun to think about the impact of slavery in the United States – an institution that still haunts this country. Slavery was a set of policies and rules that allowed one group of human beings to own another group as property, and to do with that property more or less whatever they wanted, up to and including the desecration and destruction of their bodies. The institution of slavery allowed White people to exploit the lives and labor of Black people for their own massive economic gain, and also to torture and murder them largely at will.
Economists have in recent years started to study the effects of slavery on current day economic outcomes. US counties, for example, where more people were enslaved in the 1800s tend to underperform economically still today (e.g., Nunn, 2008). There is debate as to why this is the case, but one interesting theory proposes that it has to do with a lack of public goods in these areas. Counties with more slaves, and with more free Blacks after emancipation, invested less in public goods like libraries and schools, and that absence of investment in the potential of their people continues to affect outcomes today, generations later.
Our lab has similarly begun to investigate whether and how slavery in the United States might affect current day intergroup attitudes. We are not alone in this endeavor. A sophisticated recent working paper by Acharya, Blackwell and Sen (2014), for example, examines effects on explicit attitudes. But here I am going to focus on our (to date much more limited) work investigating relationships with implicit attitudes.
Here is a state-by-state map of implicit bias, based on data from about 1.8 million White participants who took an online Black/White IAT on the Project Implicit website between 2003 and 2013. In December of last year, we conducted an analysis for the Washington Post’s Wonkblog to identify what sorts of factors might predict this distribution of bias. There appears, after all, to be a spatial logic to it – with some the highest levels of bias clustering in the south. Indeed, statistical analysis reveals that levels of implicit bias are indeed greater in former slave-holding states. But bias is also higher in some perhaps less expected locations: New York, Illinois, Pennsylvania, Ohio…
What unites these less expected locations with the slaveholding south? These are states where many African Americans moved after emancipation, and particularly during the ‘great migration’ between the 1920s and 1970s. Black people moved in large numbers from the south to cities like New York, Chicago, Philadelphia and Cleveland. And the number of African Americans – or, more specifically, the number of African Americans relative to White people – turns out to be a potent predictor of implicit bias at the state level. Indeed, this ratio (shown below in reverse, so that higher numbers indicate more White relative to Black people) predicts more than 50% of the variance in IAT scores. Further, entering this ratio into a regression equation reduces the influence of former slaveholding status to statistical non-significance.*
Now, a common interpretation of this statistical pattern would be to say that the effect of the historical variable (slavery) is accounted for or mediated by a current day variable. Slavery is associated with current day implicit bias because it (and events following its abolition) brought Black people to certain locations, and where there are more Black people today relative to the White population, bias is heightened. But it would, I think, be a mistake to assume that all of the important psychological action in these relationships is taking place in the present.
The Washington Post piece that reported these findings generated a fair amount of online attention. It caused much tweeting – including, to my chagrin, tweets from the White supremacist community. Their interpretations of the pattern tended to emphasize current day dynamics. One tweet read, for example: “Leftist lie refuted in leftist publication: the more whites experience blacks, the more they hate them.” To which a follow up tweet read: “If contact leads to hostility, then segregation is the moral solution. As racists said from day one. #wearethegoodguys”
As social psychologists, we would probably put it differently. But we might actually make similar sorts of arguments (though without drawing the same egregious social conclusions): As the size of a racial outgroup increases, intergroup threats – whether symbolic or realistic – also increase, driving bias. In this interpretation, slavery set up the conditions, but it is current day population dynamics that drive the psychology that produces the bias.
This could be the case. But I don’t think we should write off more direct influences of the past that hastily…
Coming soon: Historical Threats to White Interests (and some county-level analyses)
*More details about these analyses can be found here. It is interesting to consider both the variables that predicted IAT scores (state-level income inequality, social capital (inversely)), as well those that did not (Black/White segregation, state median income, 2008 electoral college vote).