Guilty as framed

by Henry on December 4, 2007

John Sides at _The Monkey Cage_ asks whether whites are more likely to support the death penalty when they think black people are being executed, and finds that the answer is yes.

In a 2001 survey conducted by Mark Peffley and John Hurwitz, a random subset of whites was asked:

“Do you favor or oppose the death penalty for persons convicted of murder?”
Somewhat favor: 29%
Strongly favor: 36%

Another random subset of whites was asked:

“Some people say that the death penalty is unfair because most of the people who are executed are African-Americans. Do you favor or oppose the death penalty for persons convicted of murder?”
Somewhat favor: 25%
Strongly favor: 52%

That is a 12-point increase in overall support

See more “here”:http://www.themonkeycage.org/2007/12/are_whites_more_likely_to_supp.html (and original paper “here”:http://www.uky.edu/AS/PoliSci/Peffley/pdf/Peffley%20&%20Hurwitz%20Death%20Penalty%20ajps_293.pdf.

{ 52 comments }

1

jacob 12.04.07 at 2:05 am

Er, the first link is just back to Crooked Timber.

2

Dan S. 12.04.07 at 2:13 am

Er, the first link is just back to Crooked Timber.

Well, it does go “here”, and there is “more” here, so I guess it’s literally true . . . .

Anyway . . . well, it would be interesting to see responses to a question about how some people think it’s unfair because most of the people executed are poor, or something along those lines – y’know, is support increasing because of general antipathy to disadvantaged groups/ groups seen to have high crime rates, etc. –
but honestly, I wouldn’t be surprised if the effect would be smaller/nonexistent, and that there’s just a certain percentage of people who want the state to do their lynching for them (or whatever).
Bleh.

3

Kevin Drum 12.04.07 at 2:37 am

It’s worth noting that N is pretty small in this survey. The difference between white and black responses is statistically significant, but the difference between the two question wordings isn’t (p. 1002).

4

Kevin Drum 12.04.07 at 2:40 am

Actually, I semi-retract that. I’m not sure if the difference between question wording is significant or not. Apparently the difference between the “% favor” numbers isn’t, but the delta itself is. I don’t get that. Maybe somebody can educate me.

However, the N is still pretty small. I’d take this with a grain of salt.

5

ozma 12.04.07 at 2:46 am

I hate to say it, but this could explain quite a lot of things.

6

ozma 12.04.07 at 2:46 am

I guess what I meant to say was:

Do white people have something against black people? I hate to say it but that explains ALOT.

7

Adam 12.04.07 at 2:47 am

On top of the small sample size, the result seems hard to interpret. I can think of many reasons other than overt racism why being presented with an argument contrary to a position you hold would tend to strengthen your conviction. Basically, it’s called digging in your heels.

Which is not to say that this is a happy result, but I don’t think the bleakest reading is necessarily (or even likely) the correct one.

8

Tom 12.04.07 at 2:57 am

The first statement in the second question is one which many interviewees would find problematic.

I don’t agree that any law can be judged as unfair solely because it’s applied mainly to a particular group.

You could expect a statement like that to destabilise the respondent. Perhaps this might be an alternate explanation, or partial explanation, for the result.

9

yoyo 12.04.07 at 3:09 am

I don’t think thats necessarily the best way to interpret this. I think its often the case that if you provide a bogus reason as the ‘opposition’ people will favor things, as a way of being against the bogus reason. What it says is that white people don’t think disproportionate execution of blacks is an indictment of the death penalty. That’s different from favoring the death penalty because blacks are being thrown into it.

10

yoyo 12.04.07 at 3:11 am

If you had as the frame question “Some people say people on death row are actually just mentally ill people with a bad childhood who need a hug. Do you favor the death penalty?” the number in favor would be even higher.

11

kth 12.04.07 at 3:32 am

But that rationale–that hearing a squishy liberal talking point makes a fair number of people actually favor the death penalty; that for a definite percentage, support for the death penalty is a matter of cultural identity rather than moral or practical reasoning–seems disturbing enough in its own way.

Nevertheless, the insight seems utterly persuasive. The finding would probably not have surprised Lee Atwater or Karl Rove.

12

Brett Bellmore 12.04.07 at 3:34 am

[aeiou] Click on my name for a report on the racial demographics of murder and murderers. It seems the way the question is framed is indeed misleading; Blacks are not executed disproportionately, they’re committing murder disproportionately. This is fairly widely known, perhaps poll respondents are reacting adversely to being fed a misleading question.

13

Walt 12.04.07 at 3:45 am

Brett, do you have any hobbies other than discussing the limitations of black people?

14

Dan S. 12.04.07 at 3:45 am

The way the question is framed is indeed misleading; Blacks are not executed disproportionately

Is that how the question is framed?

15

Colin Danby 12.04.07 at 4:16 am

From page 1002: “∗The experiment also randomly manipulated the source of the argument as either “some people” or “FBI statistics show that,” which had no discernible influence on support for the death penalty.”

They also tried “Some people say that the death penalty is unfair because too many innocent people are being executed.” as the initial statement, and that had almost zero effect on white responses one way or the other, apart from shifting a few people from somewhat to strongly oppose.

They do some statistical work on factors that explain the shift. Generally I would suggest reading the paper before reaching for arguments.

16

Richard Campbell 12.04.07 at 4:21 am

One thing I have noticed, and that may correlate well with this survey’s results, is a white backlash to racial appeals.

That is, given a frame that suggests that whites in general may have racist attitudes, and that racist attitudes would cause action X, many whites in my experience tend to react in a way verbalized as “Fuck you for calling me a racist, I’m going to do X now.”

17

Henry 12.04.07 at 5:20 am

Brett – you know what the rules are. Any future comments from you in excess of your once a day allowance will be disemvowelled and trollmarked using our semi-automated system. Persistent efforts to continue to post comments in excess will result in a complete ban on participation in my threads.

18

Henry 12.04.07 at 5:21 am

yoyo – the paper itself doesn’t argue for a simplistic version of racism as the explanation for its findings.

19

Henry 12.04.07 at 5:32 am

Kevin – the _n_ for the various treatments is plenty high for a study of this sort as best as I understand these things (as suggested by the high degree of statistical significance for their results). I imagine that one might raise other objections to their findings or interpretations thereof – but low n is unlikely to be one of them.

20

yoyo 12.04.07 at 5:47 am

Yeah, i don’t see anything to disagree with in the paper. I just thought how you or John sides wrote the post was misleading/simplistic.

21

dsquared 12.04.07 at 8:03 am

It’s worth noting that N is pretty small in this survey

200-odd people isn’t small.

I think that the only plausible psychological mechanism that hasn’t been considered here (other than the obvious one of racism, which we can hardly rule out) is that support for the death penalty is going to be higher if you’re thinking of yourself as punishing someone else and lower if you’re thinking about yourself as potentially being punished. Reminding white people that application of the death penalty is basically cost-free for white people thus makes them more inclined to support it.

22

bad Jim 12.04.07 at 9:10 am

When Justice Blackmun wrote “I no longer shall tinker with the machinery of death”, he may have been as frustrated with its inability to cull white malefactors as with its efficiency in killing blacks.

23

novakant 12.04.07 at 9:32 am

I think it shows that it’s futile to argue against the death penalty using ancillary arguments like the racism involved or the number of innocents killed. I remember one activist who was pushing the latter angle saying, that while he had seen some success with this line, most people who could be won over by it were still fundamentally pro death penalty. His fear was, that if the process would be improved to the point were the killing of innocents was virtually ruled out, people would swing back to fervently supporting the death penalty and he’d be back to square one. The racism angle is structured differently but suffers from similar deficiencies. I think the only way ahead is to convincingly make the point that state sanctioned murder is fundamentally wrong.

24

ben saunders 12.04.07 at 10:07 am

What would people say if #2 had been:

“Some people say that the death penalty is unfair because most of the people who are executed are men. Do you favor or oppose the death penalty for persons convicted of murder?”

25

notsneaky 12.04.07 at 10:09 am

This is one of the instances where I should have cites and sources to back up what I’m about to say but it’s stuff I read long time ago and it made sense at the time and I ain’t got it no more, so I’m just gonna pitch it out there;

Bunch of law studies; For the most part there is no statistically significant difference in how mostly X race juries deal with Y race defendants. In other words if you’re white you got no reason to worry if your jury is mostly black. And vice versa (I guess this is the amazing part). Mixed race juries will treat you just like … straithg up people would. They really do look at the facts.

However, the difference – and there is very strong difference which ends up being reflected in the incarcerations rates etc.- is what people of different races are charged with. And this is what all of this comes from. You’re a white dude busted with crime Z, you get charged with a misdemeanor. You know, your mom was a very nice lady from Sweden. You’re a black dude you’re looking at life.

In the end the juries are fair GIVEN THE CHOICE THEY ARE PRESENTED WITH. Spend sometime in America and you’ll be hard pressed to find real, blatant racism… it is purely institutional in the sense that everybody pretends that it doesn’t exits becauseit’s always the lawyer doing the racist work for you.

26

SG 12.04.07 at 10:38 am

notsneaky I don’t know if that is true or not but I remember an excellent example from recent history – that 17 year old black chap who got a consensual blow job from a 15 year old. The jury were presented with a crime he had clearly committed, but didn’t realise he was going to get some shocking punishment for it. Isn’t this the issue with the Jena 6 as well, that they are being charged with a ludicrously overserious crime? I will match your completely hazy recollection with those…

Also in Australia the high rate of Aboriginal death in prison is not actually due to higher death risk but due to extremely high rates of incarceration for minor misdemeanours (we even coined a phrase for the brace of charges, “the trifecta”).

In contradiction to your opinion though, I recall that the issue is a common example in introductory regression text books, and always ends up showing a huge bias once one adjusts for a confounder (either jury race or victim race, I can’t recall). So it could just be that you are wrong…

27

Katherine 12.04.07 at 11:29 am

I find it simply fascinating that a large proportion of the first round of comments were of the “no it simply can’t be racism (although I haven’t read the report)” variety. I genuinely mean no negative comment on those particular individuals, but it really is just fascinating how gut reactions will tend to (want to) deny racism in average white responses.

28

Brett Bellmore 12.04.07 at 11:38 am

Henry, you’re an ass, but I knew that already. You might have made it clear that you meant one comment per day site-wide, rather than per thread, before disemvolwing a comment which merely linked to what would appear relevant evidence, and said nothing particularly controversial.

The echo chamber here is getting smaller every day, people, with Henry at the controls.

29

Slocum 12.04.07 at 12:48 pm

I’d suggest that the second question form is being interpreted as “do you oppose the death penalty because most of the people executed are African American”? And so, I agree with #24 that if ‘African American’ had been replaced by ‘men’ or ‘the poor’ or ‘high school dropouts’ or ‘people with low IQs’ or ‘kids from single families’, you’d likely have seen a similar effect — with the increase being an indication that no, the respondents don’t think the death penalty is being unfairly applied to those groups (despite the numbers).

If, in fact, using ‘men’ or ‘the poor’ did not produce the same results as ‘African Americans’, then I’d say racism would be the most likely explanation. But I wouldn’t say that given just these data.

30

Barry 12.04.07 at 12:54 pm

And Brett lies again. Unless there are a bunch of other people being banned.

Katherine: “I find it simply fascinating that a large proportion of the first round of comments were of the “no it simply can’t be racism (although I haven’t read the report)” variety. I genuinely mean no negative comment on those particular individuals, but it really is just fascinating how gut reactions will tend to (want to) deny racism in average white responses.”

Katherine, it is amazing, isn’t it? And yet, when discussing racism as a major factor in differing behavior by race, the same people will deny it. Moreso, people who don’t have that reaction will observe that reaction in numerous others, and still not see just how racist they are.

31

Steve LaBonne 12.04.07 at 1:39 pm

I want to thank Henry for posting this story and thereby destroying my last tiny sliver of faith in humanity. I guess it’s better not to have any illusions left…

Seriously, that’s a hell of a depressing paper.

32

John Emerson 12.04.07 at 1:50 pm

There was an objective measure of racism in the 2000 Alabama election. The Alabama constitution had an anti-miscegenation article. It hasn’t been enforced for a long time, and there was a movement to repeal it. Probably the first evidence was the fact that the legislature didn’t dare do it themselves, but referred it to the people for a vote (though that may just be a peculiarity of Alabama law). In any case, the referendum passed something like 58-42.

The belief that interracial marriage should be illegal is a classically racist belief, so we can say that 42% of Alabamans are unmistakably racist. It shouldn’t be hard to to a comparison of distribution of Republican votes and votes on the measure in question. (The referendum was voted on in a by-election, not the main election, and it’s possible that if turnout had been greater the referendum would have lost.)

So anyway, there is a lot of pure old-fashioned racism out there.

33

dsquared 12.04.07 at 1:57 pm

32: That would be 42% of voters in that election – turnout tends to be lower among black voters than white, so probably less than 42% of all Alabamans.

34

Down and Out of Sài Gòn 12.04.07 at 1:57 pm

It’s pretty depressing that 66% of respondents think the death penalty is a good idea, period.

I’d be interested in response to the following question:

Some people say that the death penalty is unfair because some people are executed for crimes they didn’t commit. Do you favor or oppose the death penalty for persons convicted of murder?

35

Matt Weiner 12.04.07 at 1:58 pm

sg in 26: that 17 year old black chap who got a consensual blow job from a 15 year old — you’re thinking of Genarlow Wilson.

It’s my understanding that the real racial disparities in the death penalty come when you look at the race of the victim.

36

Matt Weiner 12.04.07 at 2:05 pm

Probably the first evidence was the fact that the legislature didn’t dare do it themselves, but referred it to the people for a vote (though that may just be a peculiarity of Alabama law).

According to Wikipedia constitutional amendments in Alabama must go to statewide vote unless they are approved unanimously by the legislature. Which I suppose means that there was a vote against.

37

John Emerson 12.04.07 at 2:25 pm

My guess is that the special election drew a higher proportion of the anti-miscegenation than pro-miscegenation voters, and had been set off in a special election (instead of the main fall election) for that purpose. IIRC the white non-bigot establishment supported the measure and there was a strong Vote Yes campaign — I don’t know if there was an active Vote campaign. Granted that the 41% (correction of 42%) only applied to this election, I don’t think that we can be sure which way the result should be corrected.

38

John Emerson 12.04.07 at 2:26 pm

“active Vote No campaign”

39

Sk 12.04.07 at 3:05 pm

“Brett – you know what the rules are. Any future comments from you in excess of your once a day allowance will be disemvowelled and trollmarked using our semi-automated system. Persistent efforts to continue to post comments in excess will result in a complete ban on participation in my threads.”

You really are intellectually dishonest.

Sk

40

John Emerson 12.04.07 at 3:12 pm

39: Who, Brett?

41

Benquo 12.04.07 at 3:47 pm

“They also tried “Some people say that the death penalty is unfair because too many innocent people are being executed.” as the initial statement, and that had almost zero effect on white responses one way or the other, apart from shifting a few people from somewhat to strongly oppose.”

I don’t think that’s a strong counterexample. After all, that may have been the implicit justification for the anti-death penalty respondents in the simplest formulation. Any framing of the question which brings up an argument that seems less irrelevant could make its case seem weaker.

Remember, most people (probably including you; certainly including me) are pretty bad at evaluating arguments fairly and ignoring irrelevancies.

42

Henry (not the famous one) 12.04.07 at 4:33 pm

The wonders of Google: there’s a paper entitled “Measuring the difference between white voting and polling on interracial marriage” to be found at http://64.233.167.104/search?q=cache:Cgqz9UYTuf4J:www.hmdc.harvard.edu/micah_altman/papers/MeasuringRacialDiffs.pdf+2000+alabama+%22election+results%22+miscegenation&hl=en&ct=clnk&cd=1&gl=us that offers some details about the background to the 2000 referendum in Alabama (and the 1998 vote in South Carolina). There was no significant opposition to the repeal in the Legislature (where it passed unanimously) or in the media; there was no organized opposition to it other than “the 200-member Alabama branch of the Southern Party.” Nonetheless, relying on precinct results as a basis for projecting the race of voters, the authors estimate that 49% [49-59%] of white voters, versus 8% [8-11%] of black voters voted against repeal. Similar results in South Carolina. This contrasted with 61% of potential white Alabama voters who said they would vote for repeal — what we call the “Tom Bradley effect” here in California. Which tells us something else about polls, regardless of the methodological problems that this study about the death penalty might have.

43

Henry (not the famous one) 12.04.07 at 4:43 pm

P.S. It was the November 2000 general election, not a by-election or special election.

Similar anecdotal results can be found in the 1990 gubernatorial election in Louisiana, in which David Duke, the photogenic Klansman, got the majority of white votes against Edwin Edwards. One incidental effect of that close call may have been the Republican Party softening its opposition to reform of civil rights laws, which gave us the Civil Rights Act of 1991. And the campaign also gave us a bumper sticker produced by Edwards supporters that has to be the gold standard for candor and incisiveness: “Vote For The Crook–It’s Important!”

44

Sebastian Holsclaw 12.04.07 at 4:59 pm

A lot of the responses are highly sensitive to what the answerer perceives as the validity of the framing. For example imagine a question framed: “Some people say that the death penalty is unfair because alien ghosts were exiled from another planet and have corrupted our communal energy. Do you favor or oppose the death penalty for persons convicted of murder?” Here the frame clearly cuts against the later test question. By pairing them together and offering the explanation as ‘related’, you make the second question look bad. People accept neither the factual claim (that alien ghosts were exiled from another planet and have corrupted our communal energy) nor the linkage (that the idea expressed about alien ghosts has anything to do with the death penalty). You get in the habit of saying “that’s ridiculous” and by the third time you are somewhat more predisposed to say that the thing as a whole is ridiculous—(you analyze it as a whole rather than as individual parts). This is why good interrogators ask lots of questions that the prisoner will agree with at the beginning.

Step to the innocent framing. Here I suspect that most people would accept the linkage, but they don’t accept the ‘some’ as being a statistically significant number which should change the argument. ‘Some’ people get struck by lightening, but we don’t worry about it much. This may be an incorrect view of the underlying factual issue, but I suspect it is the actual view.

Back to the original questions, I suspect that a few people don’t accept the facts, but that most people don’t accept the linkage. So while not as strongly as the alien ghosts example, they are still predisposed to reject the later proposition.

Note that the hardening of positions comes across at all levels. The number who strongly support the death penalty goes up, and the number who somewhat support it remains essentially steady. Unless people are jumping from “undecided” straight to “strongly support”, this is a hardening of position across levels.

Interestingly enough, this same shift is observed even among those who oppose the death penalty. The number who strongly oppose and who somewhat oppose drops noticeably in response to the racially charged question. This is an interesting result (though it also makes me wonder if the ‘other randomly selected group of white people’ was just straight up not as opposed to the death penalty). There are all sorts of potential explanations for that (racism, the framing I’ve discussed, drawing an odd group) but the result of the hardening in the anti-death penalty group is worth thinking about more.

Also we can look at the black respondents in the same analysis. I would suspect that they are much more likely to accept the validity of the fact presented and the linkage between the fact and the later question. Another very interesting result is that the black respondents react even more negatively to the innocent framing than the racial framing. I suspect that this result is because they strongly accept the linkage between the factoid and the later question in the innocent framing.

Anyway, no conclusions from me on the topic. I’m perfectly willing to accept, sadly, that many of those who seem to have a stronger position on the topic, really are racist. But on the other hand I hate when the interpretation intended by the interviewer is taken as gospel—the effect of people getting in a habit of disagreeing or agreeing in a particular interview really does exist.

45

eudoxis 12.04.07 at 6:05 pm

“most of the people who are executed are African-Americans”

Is this true?

46

Uncle Kvetch 12.04.07 at 6:12 pm

Is this true?

Apparently not. While African-Americans have been disproportionately represented among the executed, whites have been a numerical majority.

47

yabonn 12.04.07 at 7:07 pm

It would b interesting to compare/complete with something like :

“Some people say that being poor affects dispoportionately African-Americans. Do you favor or oppose policies that would help the poor?”

48

Watson Aname 12.04.07 at 11:24 pm

Uncle Kvetch: Do you know if it is still disproportional if normalized by income? Or by other socio-economic proxies?

49

Uncle Kvetch 12.05.07 at 1:32 pm

Uncle Kvetch: Do you know if it is still disproportional if normalized by income? Or by other socio-economic proxies?

Sorry, no idea. I was curious about eudoxis’ question so I dug up the linked site, but that’s all the research I’ve done on this.

50

Kathleen 12.05.07 at 8:38 pm

This elegant study manages to demonstrate that the Southern strategy is the American strategy — Republicans can push *anything* as long as they can suggest it punishes, or at least snatches any benefits from, blacks.

51

NullPointer 12.05.07 at 9:45 pm

I am very unimpressed by the report.

First the authors’s are very anti-death penalty and have an agenda (they even bring up WMD in Iraq, huh?!) and are not genuniely looking for confounding factors.

They even go so far as to offer sneaky coaching to death penalty opponents:
Thus, making more general arguments against the lack of fairness of the death penalty without making a direct reference to race may constitute a successful “stealth” strategy that appeals to
blacks but does not produce countermobilization among whites.

The clue to me that something is wrong is the table on p 7. First the “baseline” question only has N=117, not 200 has has been reported. Secondly the “Strongly Oppose” category actually goes down from 18% to 11.4% when they introduce the race information.

Based on the people you personally know who oppose the death penalty (strongly), do you think *any* of them would likely be swayed by the race of the offender? I am incredulous to the point where I think these too sample populations are different.

I also find it interesting (and troublesome) that they actual question is varied too. “The experiment also randomly manipulated the source of the argument as either “some people” or “FBI statistics show that,” which had
no discernible influence on support for the death penalty. p7” It seems to me that there is a big difference between saying that “some people say” and “law enforcement says that they are executing a lot of innocent people”. The authors tell us there is no significance, but it would have been much better if the numbers were broken out or the question stayed the same.

52

Brett Bellmore 12.05.07 at 10:53 pm

Again, and I don’t particularly care that it’s going to get me reamed for pointing this out, African Americans are not disproportionately executed. They’re disproportionately committing murder. If anything, African-Americans are under-represented on death row relative to the share of murders they commit. (Check out figure 3 on page 25.)

If there’s any disproportion, it’s in how often the death penalty is levied relative to the race of the victim. Which may very well be a result of racism, but may alternately be a consequence of African-Americans being hostile to the death penalty, and comprising most of the jury pool in areas where most of the black on black murders are taking place. It could be challenging to disentangle these effects, as it would require a detailed inquiry into all the circumstances surrounding each of the cases.

Comments on this entry are closed.