Friday, December 19, 2014

What Predicts the Spatial Distribution of IAT Scores in the United States?

On December 15, an article by Chris Mooney on the Washington Post’s Wonkblog presented an interesting spatial distribution of state-level IAT scores across the United States. Based on a sample of more than 1.5 million participants from Project Implicit, IAT scores appear to be systematically higher in the southeastern and eastern US. (For an interactive version of the map, click here.)


The map above raises (at least!) two questions. First, what might predict this distribution of IAT scores? And second, given that the differences between scores (though significant with this large sample) are quite small (range = .341 - .456), are they meaningful?

To address these questions, we examined a number of potential predictors of state-level IAT scores. We selected predictors either because they were theoretically plausible (e.g., state level indices of racial segregation, median income, income inequality, the ratio of White:Black state residents, political preferences) or because they are known to share a similar spatial distribution to the observed IAT scores (e.g., ‘social capital’, status as a former slave-holding state). In addition, we examined whether state-level IAT scores could be used to predict actual behaviors – specifically, numbers of racist tweets recorded after the 2012 Presidential election.

What Predicts this Distribution of IAT Scores? 

Correlational analyses revealed several significant correlates of state-level IAT scores.  Note that where possible we examined variables around the year 2009 because this was the modal year of IAT data collection.

·         We used the ‘Civic Life Index’as a measure of social capital.  This index captures a composite of volunteering behavior, neighborhood engagement, voter participation and civic infrastructure.  There was a strong negative correlation between this measure of social capital and IAT scores (r (49) = -.576, p < .001)
·         State levels of income inequality (GINI) were positively correlated with IAT scores (r (49) = .409, p < .01).
·         The ratio of White to Black state residents (log transformed because it is skewed) was a strong negative predictor of IAT scores (r (49) = -.733, p < .001) – the more White relative to Black state residents, the lower the IAT score.
·         We indexed political preferences in terms each state’s Electoral College vote in the 2008 Presidential election.  States that voted for McCain (M = .405, SD = .030) vs. Obama (M = .398, SD = .027) did not have significantly different IAT scores (F (1, 49) = 0.79, p > .30).
·         In contrast, status as a former slaveholding state (in 1860) was a strong predictor of current day IAT scores.  Former slave states (M = .426, SD = .022) had significantly higher IAT scores than non-slaveholding states (M = .390, SD = .023; F (1, 49) = 27.45, p <. 001).
·         An index of Black/White segregation and state-level median income did not significantly correlate with IAT scores (rs < |.17|, ps > .20)

Of course, many of these variables were correlated with each other, so we ran a series of regression analyses that allowed us to identify the unique effects of each potential predictor.  Entering all of the variables EXCEPT population composition (ratio of White to Black residents) into a regression equation showed that two predictors remained significant – the Civic Life Index negatively predicted IAT scores (β = -.449, p < .01), whereas status as a former slaveholding state positively predicted IAT scores (β = .439, p < .01). 

When, however, the ratio of White to Black state residents (log transformed) was entered into a regression equation with these two variables, the contributions of the Civic Life Index (β = -.173, p > .10) and slaveholding status (β = .214, p > .10) dropped to non-significance.  By far the strongest single predictor of state-level IAT scores was the ratio of White:Black state residents (β = -.470, p < .01), accounting for a rather astonishing 53.8% of the variance.  States where Whites outnumber Blacks substantially in the population have lower average IAT scores.  In contrast, states where Blacks make up proportionally more of the population have higher average IAT scores.

Figure: Relationship between White:Black Population Ratio (log) and IAT Scores



Interpretation.  The strong association between population composition and IAT scores is consistent with the idea that bias increases when and where there is greater perceived competition between groups.  In the political science and sociology literatures this sort of population composition effect is often termed the “racial threat hypothesis” (Key, 1949).  In states where the African American outgroup comprises a larger proportion of the population, Whites may perceive greater competition for political, cultural and economic resources (e.g., Tolbert & Grummel, 2003).  Further, due to negative stereotypes about African Americans, they may also perceive greater risk for cross-race crime (e.g., Eitle, D'Alessio & Stolzenberg, 2002).

In contrast, these findings are inconsistent with any simple hypothesis that contact between members of different racial groups will lead to reduced bias.  (Of course, social psychologists know this anyway!)

Critically, however, this pattern should be interpreted within its historical context.  It is not a coincidence that bias is greater in former slaveholding states.  In these and other states with large African American communities, for example, emancipation and extension of full voting rights during the civil rights era posed a significant threat to White’s political power and historical dominance. Understandings of racial identity and the nature of intergroup relationships forged in the past persist today and likely continue to exert influence (see Acharya, Blackwell & Sen, under review; Jackman, 1994). In a different historical context, however, the relative size of racial groups might not have the same relationship to bias. 

Do State-Level IAT Scores Predict Anything?

After the 2012 election, geographers at the University of Kentucky plotted the distribution of racist Tweets geolocated by state.  They created a racist tweet location quotient “that indicates each state's share of election hate speech tweet relative to its total number of tweets”. We found that the distribution of tweets they observed can be predicted by state-level IAT scores, such that states with higher average IAT scores tended to emit more racist tweets (r (49) = .326, p <. 05). 

We further examined whether this relationship varied as a function of which candidate (Romney or Obama) won the Electoral College vote in each state.  Perhaps unsurprisingly, numbers of racist tweets were greater in states that voted for Romney (who lost the election; β = .409, p < .01; R2 = .25).  However, there was also a significant interaction between state vote and IAT scores (β = 6.78, p < .01; ΔR2 = .14).  The positive relationship between IAT scores and racist tweets was significant among states that voted for Romney (β = .48, p < .05), but was non-significant among states that voted for Obama (β = -.25, p > .20).

 Figure: Relationships between IAT Scores and Racist Tweet Location Quotient for States that Voted for Romney Vs. Obama


  
Interpretation.  This pattern is consistent with contemporary theories of racial bias (e.g., aversive racism). Given that social norms generally prohibit overtly racist behavior, biased attitudes often only predict behavior when people feel they have an excuse to behave in a biased fashion or perhaps when something (like frustration or anger about a preferred candidate losing an election) reduces control over biased responses.  Alternately, the normative climate in states that voted for Romney might be have felt sufficiently accepting of prejudice to some individuals following the election to sanction the translation of biased attitudes into behavior.

Conclusion

We started with two initial questions. 

What predicts the spatial distribution of IAT scores in the United States?  A cluster of variables are correlated with state-level IAT scores, including indices of social capital, income inequality and historical status as a slaveholding state.  Notably, however, the strongest correlate was the ratio of White to Black state residents, consistent with the idea that intergroup competition and identity threat contributes to the type of  bias indexed by the IAT.

Given that the differences between states’ IAT scores are quite small, are they meaningful?  The answer to this second question appears to be yes.  First, the fact that IAT scores are related to variables like social capital and population composition suggests that they vary in a systematic fashion.  Second, we found that state-level IAT scores predicted a form of biased behavior – namely the number of racist tweets produced in each state following the 2012 election of President Barack Obama. Thus, although the observed differences between states are indeed small, the very large sample (1.5 million+) from which these estimates of bias are derived means that they are stable, predictable and predictive.

Finally, it is important to note that the current analyses sample in a limited way from a much larger universe of possible predictors.  We believe it was an informed sampling, but there are certain to be other important predictors out there.  Further, a more sophisticated approach would examine effects at the county-level.  We are currently engaged in some county-level analyses, and hope to report findings in the near future. 

The Data

If you'd like to take a look yourself or have ideas about additional predictors (urban/rural differences being an obvious contender), the data for these analyses can be downloaded here.  And many thanks to Brian Nosek and Project Implicit for providing open access to the IAT data!

Thursday, November 27, 2014

Echoes of the Past in Ferguson

“History is about remembering the past, but it is also about what we choose to forget.”
                                                                                                                 - Margaret MacMillan

The past is present in Ferguson, Missouri.  But whereas it echoes with booming resonance for the African American community, the past vibrates at a frequency to which many Whites are deaf.

For obvious reasons, White people in America like to focus on the progress that has occurred in race relations, and when we remember the past at all, it is in highlight reel (not real) form.  Slave owners bad, Lincoln good.  Segregation immoral, MLK non-violent hero.

From the perspective of progress, Michael Brown’s shooting and the subsequent failure to hold anyone accountable seems like a setback.  Michael’s death is a tragedy and the grand jury decision not to bring the case against Officer Darren Wilson to a proper trial is perhaps a shame.  But the evidence just wasn’t strong enough, and that’s all there is to it.  An injustice, yes; but hardly warranting all this anguish, all this rage.  Surely?

But this is not just a specific tragedy or a single instance of possible injustice.  Neither is it a new type of injustice in a long history of injustices.  To those who remember the past, it can seem to be exactly the same injustice as there always was and always has been – the destruction and desecration of a Black body, and a gaping hole where equal protection under the law is supposed to be.

It is 2014 and we have a twice-elected Black President, and yet in many ways we could be back, to pick a non-arbitrary date, in 1944.

In 1944, Gunnar Myrdal published the second volume of his voluminous survey of US race relations, An American Dilemma (subtitled, The Negro Problem and Modern Democracy).  A Swedish economist and sociologist, Myrdal was invited by the Carnegie Corporation to study racial discrimination and disparity.  They specifically wanted his perspective as an outsider, and as WWII raged in Europe, he travelled extensively throughout the United States and particularly the South, observing, interviewing and writing.

In an effort to remember the past and by hearing its echoes to better understand the present, I have returned to Myrdal in the past few days.  His observations generally need little interpretation, and their parallels in the present are strikingly, startlingly obvious. 

On Grand Juries and Failures to Indict

“When the offender is a white man and the victim a Negro, a grand jury will often refuse to indict… “

“It is notorious that practically never have white lynching mobs been brought to court in the South, even when the killers are known to all the community and are mentioned by name in the local press.  [But] when the offender is a Negro, indictment is easily obtained and no such difficulty at the start will meet the prosecution of the case.”

On Witness Testimony

“Greater reliance is ordinarily given a white man’s testimony than a Negro’s.  This follows an old tradition in the South, from slavery times, when a Negro’s testimony against a white man was disregarded; and the white judge may justify his partisanship by what he feels to be his experience that Negroes are often actually unreliable.”

On Policing and Police Testimony

“…the police often assume the duty not only of arresting, but also of sentencing and punishing the culprit.”

“The Negro’s most important public contact is with the policeman. He is the personification of white authority in the Negro community.
‘There he is “the law” with badge and revolver; his word is final; he is the state’s witness in court, and as defined by the police system and the white community, his word must be accepted.’ (citing Raper, 1940)”

On Public Officials, including Prosecutors

“The American tradition of electing, rather than appointing, minor public officials has its most serious features in regard to the judiciary branch of the government…”

“…the fact that administration of justice is dependent on the local voters is likely to imply discrimination against an unpopular minority group, particularly when this group is as disfranchised as Negroes in the South.  The elected judge [or other officer of the law] knows that sooner or later he must come back to the polls, and that a decision running counter to local opinion may cost him his position.  He may be conscious of it or not, but this control of his future career must tend to increase his difficulties in keeping aloof from local prejudices and emotions.”

The relevance of these quotations to current events needs little elaboration.  I will comment here only on the last observation.  Myrdal noted that elected officials – including public prosecutors – might have competing loyalties: to the law they are sworn to uphold, but also to the voters that elected them.  With regard to the questionable decisions of St. Louis County prosecutor Robert McColloch in this case, therefore, one sincerely hopes it is not relevant that 70.3% of the voters in this county are White and that an estimated 71% of those White county residents thought (pre-Grand Jury decision) that the officer involved should not be charged.


It is impossible to read these observations from the early 1940s – before Brown vs. Board of Education, before Selma, before the March on Washington, before the Civil Rights Acts, before President Obama - without hearing the past reverberating in the present. To understand why the Black community is so upset by Michael Brown’s death and this grand jury decision, we must understand how similar these events seem to the terrible rhythms of the past, and how those rhythms have never subsided.  What feels like the same injustice triggers the same anguish, the same fear, the same frustration and anger. For real (not reel) racial progress to be made, it is time for us – and for White people in particular – to choose to stop forgetting.

Monday, August 25, 2014

What is our purpose? What are we working for?

As we start a new academic year, it's hard to imagine a better mission statement than these concluding words from Gunnar Myrdal's first volume of An American Dilemma...

"Social study is concerned with explaining why all these potentially and intentionally good people so often make life a hell for themselves and each other when they live together, whether in a family, a community, a nation or a world...

The rationalism and moralism which is the driving force behind social study, whether we admit it or not, is the faith that institutions can be improved and strengthened and that people are good enough to live a happier life.

With all we know today there should be the possibility to build a nation and a world where people's propensities for sympathy and cooperation would not be so thwarted."

- Gunnar Myrdal, 1942, An American Dilemma

Thursday, August 14, 2014

Assume a dangerous crowd and you get one: How unsophisticated policing can make things much much worse


Events in Ferguson – the suburb outside St. Louis, where last week 19-year-old Michael Brown was fatally shot by police – are changing fast, and currently for the better.  As I write this, police tactics have changed.  The riot gear, assault weapons, and faux tanks are gone.  Officers are out, mingling with the protestors, connecting, talking.  The protesters, in turn, are calm, expressing relief at the positive turn of events.  Observers around the country are also relieved, for not long ago, as tear gas canisters rained down and cameras were ordered off, one couldn’t help feeling that we were back in the worst, repressive days of the 1960s.

It was like the 1960s, except that the weapons and gadgetry sported by the police are very much 2014.  But what we have observed over the last few days is that none of this modern, sophisticated equipment makes a police force sophisticated.   What makes a police force sophisticated in protest situations is how well it understands the psychology of crowds.

Unsophisticated policing assumes that protestors are the same as a mob – unthinking, irrational and inherently dangerous. Unsophisticated policing assumes that all members of a crowd have the same goals, and treats them accordingly.  Unsophisticated policing assumes that the only check on anarchy is the police force itself – and that the best means of control is through fear and submission.

 All of these assumptions are wrong.  Further, they are counter-productive, exposing everyone, including police officers themselves, to unnecessary danger.  Outstanding research by social psychologists – including Steve Reicher, Clifford Stott and John Drury – has found that the assumptions made about crowds tend to be self-fulfilling.  Assume a unified crowd and you end up with a unified crowd.  Assume a dangerous crowd and treat it as dangerous – weapons drawn, orders shouted by megaphone, heads knocked – and you get a dangerous crowd. 


“If one believes that all crowd members are potentially if not actively dangerous, then one will (1) treat all crowd members alike and hence create unity amongst them, (2) react to the violence of some crowd members by imposing restraint on all, thus increasing the likelihood of violating ingroup conceptions of legitimacy and uniting the crowd in hostility and opposition to the police, and (3) increase the influence of those advocating conflict in the crowd and undermine self-policing amongst crowd members.”

So what are crowds actually like?  They rarely start out unified – but are instead made up of disparate subgroups with disparate goals.  They rarely contain many members intent on violence and destruction.  Certainly there may be some such individuals – who would like to escalate aggression, vandalize or loot – but most of the time, most protestors are there to peaceably express their frustration, anger or disappointment at something they consider unjust.  And in a democratic society, this is a most important right.

What sophisticated crowd management understands is that crowds will seek a unity – a common identity and sense of direction– and that it could go either way.  The group may solidify around peaceable goals and “crowd out” members advocating aggression and hooliganism.  These sorts of crowds are self-policing, setting and enforcing their own pro-social norms for behavior.  When this happens there is more than just a thin blue line between protest and anarchy.  Alternately, however, crowds may coalesce around more hostile and aggressive norms, pushing out advocates for moderation.  Which of these identities wins out has a great deal to do with the way the crowd is treated.

It is not, of course, inevitable that aggressive policing leads to an aggressive crowd.  Non-violent protest movements around the world consciously choose to respond to repression with non-aggression – and they gain moral authority and are often successful because of it.  However, when this happens it is usually due to strong moral leadership from within the group, fostering the strength to resist an oppositional identity. 

And in this sense, as much as events as Ferguson have felt reminiscent of the worst of the 1960s, they have also been all the more worrying because we are without the best of that era.  Without the non-violent ethos of leaders like the Rev. Martin Luther King, there is the very real danger that aggressive police tactics will lead to retaliation and violent escalation.

Fortunately – for now – it appears that things are getting a bit more sophisticated out there.  It’s about time.  In 2014, we should (and do) know better.

Monday, June 16, 2014

How Do World Cup and Economic Performance Relate? Historically Little Gain for Winners, But Bad News for Runners Up. Can't Claim Much About the Future.


A couple of weeks ago, Goldman Sachs published a World Cup and Economics Report.  They found, among other things, that winners have tended to experience a short ‘honeymoon bounce’ in equity market performance – outperforming the market by about +3.5% in the following month.  In contrast, runners up tend to experience a rough patch - underperforming by about -5.6% over the following 3 months. 

We were interested in whether World Cup winners might also experience broader economic gains.  Winning the World Cup is clearly a Good Thing to many people.  Might the morale boost, the surge in national identity and pride following a victory be associated with greater productivity, consumer spending, etc., etc., resulting in a boost to, say, Gross Domestic Product (GDP)?

To find out, we did a quick analysis and examined percent change in GDP per capita from the year prior to the World Cup to the year after for the 1st, 2nd, 3rd and 4th placing nations in every contest from 1930 through 2010*.  And the answer appears to be: No!  Average increases in GDP per capita for the winning nation have been no greater than increases for the 3rd and 4th placing countries.

So, no good news for the winner.  There has, however, typically been bad news for the runner up.  As shown in the Figure below, nations placing 2nd in the World Cup have tended to perform worse economically (in the following year) than countries placing 1st, 3rd and 4th – exhibiting gains in GDP per capita that have been, on average, less than half that of their rivals.


The next Figure illustrates that this pattern is fairly robust.  The 1930s were a bad decade for everyone, and results for contests since WWII (i.e., excluding 1930, 1934 and 1938) show the same configuration.  Further, if we break the contests down by time period – the 30s, 50-60s, 70-80s and 90-00s – we find the same pattern in each case.  In every period, the 2nd placing nation tended to exhibit the weakest economic gains, although this trend seems to have lessened in the recent past.


Taking a finer-grained look, the next Figure shows percent changes in GDP per capita following every World Cup.  The trend lines indicate rolling 5-contest averages, and it is notable that the line for 2nd placing nations falls below all of the others at all time points except two.


Descriptively, this seems to be a historically robust pattern.  Treating the Cups to date as the full population of World Cup contests, it is accurate to say that in the year following the  World Cup, second placing countries have exhibited only ~44% of the gains in GDP per capita shown by winning nations (and still less compared to the 3rd and 4th placing nations).   

Inferentially, however, we are on much less solid ground.  Treating the Cups to date as a sample of all possible World Cup contests (and the data as providing estimates for the future population), it is not valid to assume that this pattern will hold going forward.  The first Figure above shows the 95% confidence intervals for each estimate.  There is clearly a great deal of overlap in these intervals, and with the limited sample they do not differ statistically from one another (although the 2nd place estimate is the only one to overlap with zero growth).  For this reason, coupled with the fact that these data are correlational, we cannot claim that placing second in the World Cup hurts an economy.  We can simply say that 2nd placing nations have tended – to this point – to have smaller economic gains in the following year.

This pattern could well be an historical fluke.  But is there any plausible mechanism for why placing second might hurt an economy? Psychologically, this pattern is reminiscent of findings that individual athletes who place second (e.g., in Olympic contests) tend to be less happy than athletes who place third (Medvec, Madey & Gilovich, 1995).  The standard explanation for this has to do with the counterfactuals – the alternate histories – that these athletes can imagine.  The runner up can almost taste her victory, and compares her #2 spot to the glory of the victor.  The bronze medalist can imagine not having made it to the podium and is, as a result, happier with her lot.  Perhaps something similar could be going on at a national level – a large-scale deflation in morale with economically damaging consequences – although this would be remarkable.

Finally, we should note that the underperformance exhibited by 2nd placing nations has historically been reasonably short-term.  The final Figure plots percent change in GDP per capita from the year prior to the World Cup to three years after – and shows that by this point, second placing nations had, on average, caught up to their rivals.


*Economic data is from WorldEconomics.com (1990 GK$), except for Turkey in the 2002 WC and all nations in the 2010 WC, which is from IndexMundi.com (2011 GK$).  Thanks to Liana Mitteldorf for checking the figures.

And here it is - % Change in GDP from the year prior to World Cup to the year following…


Year
1st Place
GDP %  Change
2nd Place
GDP %  Change
3rd Place
GDP %  Change
4th Place
GDP %  Change
1930
Uruguay
-9.020
Argentina
-14.999
USA
-17.510
Yugoslavia
-7.551
1934
Italy
8.328
Czechoslovakia
-5.564
Germany
15.861
Austria
2.612
1938
Italy
6.086
Hungary
11.600
Brazil
1.040
Sweden
10.454
1950
Uruguay
10.013
Brazil
2.592
Sweden
6.225
Spain
10.719
1954
W. Germany
18.186
Hungary
10.194
Austria
22.142
Uruguay
4.145
1958
Brazil
11.384
Sweden
5.812
France
3.209
W. Germany
10.551
1962
Brazil
1.067
Czechoslovakia
-1.767
Chile
6.253
Yugoslavia
10.270
1966
UK
3.046
W. Germany
2.297
Portugal
12.249
USSR
7.100
1970
Brazil
14.615
Italy
2.854
W. Germany
6.102
Uruguay
2.785
1974
W. Germany
0.627
Netherlands
2.240
Poland
8.764
Brazil
7.912
1978
Argentina
-0.927
Netherlands
3.315
Brazil
7.119
Italy
9.016
1982
Italy
1.447
W. Germany
1.272
Poland
2.098
France
2.729
1986
Argentina
6.774
W. Germany
3.705
France
4.044
Belgium
3.766
1990
W. Germany
0.556
Argentina
7.009
Italy
3.569
UK
-1.566
1994
Brazil
7.272
Italy
4.740
Sweden
5.995
Bulgaria
7.117
1998
France
4.935
Brazil
-1.969
Croatia
2.421
Netherlands
9.383
2002
Brazil
0.199
Germany
-0.360
Turkey
1.419
South Korea
8.229
2006
Italy
3.553
France
3.487
Germany
5.820
Portugal
2.578
2010
Spain
0.266
Netherlands
2.654
Germany
4.791
Uruguay
7.669
AVERAGE
4.653
2.059
5.341
5.680