Tuesday, April 22, 2008

The Distortion of Statistics for Political Goals: Considerations Concerning the Case of Mumia Abu-Jamal

The Distortion of Statistics for Political Goals

Some Elementary Considerations Concerning the Recent Mumia Abu-Jamal Court Decision re Batson

By Michael Schiffmann

Much has been said in recent days, and rightly so, about how the 3rd Circuit Court of Appeals in its March 27 decision to deny Mumia Abu-Jamal a new trial or at least a hearing on the so-called Batson issue – prosecutorial racism in jury selection – once more created a new “Mumia law” in demanding that the defense objects to such racism already at the time of the trial.

In the following remarks, which I hope to expand into a fuller study in the weeks to come, I want to concentrate on the second part of the court’s ruling on Batson – the one where the court claims to deal with the issue “on the merits,” i.e., not in a formal but in a substantive fashion.

In a nutshell, what the court majority claims in its 77-page part of the whole 118-page decision of the 3rd Circuit is that the defense lacks the data to lay the prima facie case for its claim that prosecutor Joseph McGill used his peremptory challenges in a systematic fashion to strike blacks.

The majority concedes that the defense did supply data on the so-called “strike rate,” which is “computed by comparing the number of peremptory strikes the prosecutor used to remove black potential jurors with the prosecutor’s total number of peremptory strikes exercised.”

There is no question here that the prosecutor’s strike rate–he undisputedly used at least 10 out of the 15 peremptories he used altogether to strike blacks, which yields a strike rate of 66.7 percent – is in stark contradiction to what one would expect from the racial composition of the city at the time, whose population according to the official 1980 census was 37.8 percent black.

But according to the court majority, in order to properly evaluate this strike rate in the Abu-Jamal case, data have to be supplied on another statistical rate, the so-called “exclusion rate,” “which is calculated by comparing the percentage of exercised challenges used against black potential jurors with the percentage of black potential jurors known to be in the venire [jury pool].”

That means that the defense is supposed to supply data on the race of all the jurors that were questioned during the so-called voir dire, i.e., the process of the jury selection, which in the case of Abu-Jamal lasted seven days and according to the Philadelphia Inquirer, June 17, 1982, involved the examination of 157 jurors (in my own files, the transcripts for the next to last day of the jury selection are missing, but that number squares well with the rest of the data that I do have).

In demanding these data, the court majority cites alleged legal precedent which will certainly be ably dealt with by Abu-Jamal’s defense, but which time and space doesn’t allow to go into here. The substantive question, however, is: Why should these data even matter?

The large majority of the jurors in question, 107, were struck, not peremptorily, i.e., without giving a reason, but “for cause,” and therefore their race shouldn’t matter whatsoever: The final arbiter of who gets struck for cause and who is left for the parties to either accept or strike peremptorily is the judge, not the prosecutor, and the whole Batson issue is not about the judge, but the prosecutor, and not about strikes for cause, but about peremptory strikes.

The Two Elephants in the Room

Of course we know that court rules and precedent often defy logic, rationality and mere common sense. But in an utterly astounding move, the March 27 court ruling goes even further and proceeds to distort the record, in ignoring or even expressly claiming the absence of data that the defense did supply both in its October 15, 1999 habeas corpus petition and in filings preceding the May 2007 Abu-Jamal hearing in Philadelphia.

As noted above, 107 of the 157 jurors questioned during the seven-day pre-trial empanelling of Abu-Jamal’s jury were struck for cause, the overwhelming majority for either (1) personal hardship involved in serving two to three weeks on a sequestered jury, (2) doubts whether they could be fair (in my data on six days of jury selection, not one juror said he had a fixed opinion that Abu-Jamal was innocent, but many had already concluded he was guilty), or (3) opposition against the death penalty, which was by no means only constrained to blacks.

5 persons were either seated as one of the 4 alternate jurors or – in the case of 1 person – peremptorily struck as alternate juror, a matter which I will come back to in a moment.

This leaves us with 45 persons who were either accepted into the jury or struck peremptorily by either the defense or the prosecution. 12 of these were seated as jurors, 19 were peremptorily struck by the defense, and 15 were peremptorily struck by the prosecution. As the defense states in a July 19, 2006 filing, 6 of these “were struck by the defense before the prosecutor had an opportunity to either strike or accept them.”

These 45 alone already constitute a sizable 28.7 percent of the whole pool of potential jurors questioned during the voir dire process – the so-called “venire.” This is certainly a statistically significant number whose racial composition, if known, should allow for reasonable conclusions about the probable racial composition of the whole venire of 157 persons.

Subtracting from these 45 the 6 persons struck by the defense before the prosecutor could strike or accept them, we are left with 39 persons (24.8 percent of the whole venire) where the prosecutor had an opportunity to display either racial neutrality or racial bias via the number of black versus white persons he struck peremptorily.

The racial composition of this set of 39 persons, and it alone, should logically be the basis to put the prosecutor’s “strike rate”
of 66.7 percent against black persons into perspective, since they, and they alone, were the persons against whom he could use peremptory strikes.

What about these 39 persons who together formed almost a quarter of the whole venire? In connection with them I could barely trust my eyes when I read the March 27 court decision. Buried in footnote 18 on page 47, it says:

Abu-Jamal contends the prosecutor had the opportunity to strike thirty-nine venirepersons, of which fourteen were allegedly black, but he does not cite any record support for these numbers. We see no record support for these numbers.

This is stunning. Both the 39 persons who the prosecutor had an opportunity to accept or peremptorily strike and the additional 6 persons struck first by the defense are given by name, race and voir dire day and transcript page numbers on p. 18-20 of the July 19, 2006 defense filing quoted above. Even if the court were to insist to cast doubt on one or another step in the data collection in this defense filing, which it doesn’t even try – the contention that the defense “does not cite any record support for these numbers” is simply absurd.

In fact, the defense presents solid data showing that indeed of these 39 persons, 14 were black – and that the rest, 25, were white, that is, the composition was 35.9 percent black versus 65.1 percent white. This percentage of African Americans is already slightly smaller than their 1980 share in the racial composition of the city, but the prosecutor still used the vast majority, 66.7 percent, of his peremptories to remove even more of them.

Actually, if one wants to talk about “exclusion rates” in the sense defined by the court, the only thing that rationally makes sense is a comparison between these two numbers – 35, 9 percent blacks among the set of person where the prosecutor could strike peremptorily and 66.7 percent blacks among the set of persons where he did strike peremptorily.

Pushing this a little further and factoring in the 6 persons, all white, struck by the defense before the prosecutor could accept or strike them, we arrive at still 14 black but now 31 white persons, and the black/white relation is now 31.1 versus 68,9. These 45, all given by name and race in the July 19, 2006 defense filing, were the persons considered for service in the jury itself.

As mentioned above, there were also 5 persons who were considered as alternate jurors, one of whom was peremptorily struck (by the defense). Abu-Jamal’s 1999 habeas corpus petition[1] identifies all of them as white, which is not in doubt or even contested in the case of the 4 that were seated, and easily verified in the case of the juror peremptorily struck, who identified himself as “Italian” when he was questioned.

This raises the number of jurors whose race is either given in the July 19, 2006 defense filing (45) or identified in the 1999 habeas petition and easily checkable from the record (another 5) to 50, or 31.8 percent of the entire venire, certainly a not insubstantial percentage. Looking at the racial composition of these 50 persons, we find 28 percent blacks and 72 percent whites.

None of these data are mentioned anywhere in the March 27, 2008 ruling, not even in Judge Ambro’s 41-page dissent on the Batson question, even though, to his credit, it must be said that he argues for a new hearing for Abu-Jamal even without considering these data.

Also, very strikingly, the whole 118-page court decision fails to even mention any of the statistical data supplied by the defense on a systematic pattern of discrimination by the Philadelphia District Attorney’s Office in general or by Abu-Jamal prosecutor Joseph McGill in particular, data that went far beyond and supplied background to McGill’s 66.7 percent strike rate of blacks in Abu-Jamal’s June 1982 trial.

But that doesn’t mean that these data are not there and were not supplied by the defense. They were just ignored by the court, apparently being to inconvenient.

In its centrally important July 19, 2006 filing the defense clearly argues, from the known number and from the record that considering the 39 really relevant venirepersons the prosecutor struck 71% (10/14 [10 of 14]) of the blacks he had an opportunity to strike, but struck just 20% (5/25 [5 of 25]) of the whites he had an opportunity to strike – i.e., he struck blacks at 3.6 times the rate than he struck whites. The odds of being struck if you were black were 2.5-to-1 (10/4 [10 to 4]), but the odds of being struck if you were white were just 0.25-to-1 (5/20 [5 to 20]) – i.e., a black person’s odds of being struck were 10 times higher than someone who is white. [Emphasis in original.]

Can anyone regard this as a statistical “warp” or accident? To pose the question is to answer it. If we factor in the 4 white alternate jurors that the prosecutor could have struck peremptorily but did not, the picture gets even starker.

These two facts – that the defense has supplied statistically significant hard data on the race of approximately one third (50 out of 157) potential jurors, and that if one compares the rates with which the prosecutor struck blacks when he could with the rates with which he struck whites when he could, the result one finds an almost grotesque disparity where a black person was at least ten times as likely to be struck as a white one – these two facts are the two big, big elephants in the courtroom in this case which won’t go away and are there for everyone to see but which none of the judges of the 3rd Circuit wanted to talk about.

You Do Not Need to Be a Rocket Scientist…

As we have just seen, even if one were to insist, against logic and common sense, that the composition of the whole venire, as opposed to the set of those jurors whom the prosecutor had an opportunity to strike or accept, is of tremendous importance, the defense has already supplied significant data that very much indicate that composition.

In fact, the numbers supplied above, taken from defense filings which in turn took them – contra page 47, footnote 18 of the recent court decision – right from the record, are hardly surprising.

Actually at the May 17, 2007 Abu-Jamal court hearing two questions were brought up for the first time: (1) Should the composition of the whole venire be regarded as a decisively important question, and (2) Could it not be that there was such a heavy black overrepresentation of blacks in Abu-Jamal’s venire that could possibly justify the prosecutor’s 66.7 percent anti-black strike rate?

As for the first question, I have already argued above that a positive answer simply makes no sense. If the anti-black peremptory “strike rate” of a prosecutor is to be evaluated by context, that context should be the racial composition of the set of persons he had an opportunity to peremptorily strike at all, not to the composition of the entire venire, the vast majority he could not strike peremptorily in the first place.

Court precedent on Batson clearly – and rightly – says that statistical data to evaluate a claim of discrimination should not be applied “mechanically,” but rather, in a meaningful way. So it should be here.

As for the second question, journalist and author Dave Lindorff pointed out right after the May 17, 2007 court hearing that the argument is not only highly speculative but, given the concrete conditions in the case at hand, also bordering on the absurd.

Since in 1982 prospective jury pools were (theoretically randomly) drawn from voter lists, the likelihood of a heavy black overrepresentation – say, 50 percent, or 67 percent (the approximate “break-even” point), or 80 percent in a city like Philadelphia with a black population of around 38 percent in 1980 – is very small, since black people nationwide, and in Philadelphia in particular, tended, if anything, to be underrepresented in the voter registration lists.

One does not need to be a rocket scientist to understand Lindorff’s point, and, thinking of the language of Batson, one could even call it a prima facie case.

Some Additional Data

If one tries to go beyond the obvious and combs through the data, as I have done in at least a preliminary fashion with the transcripts of six of the seven Abu-Jamal jury selection days including data on 134 of the altogether 157 venirepersons, the picture is very much the same.

70 (or 52%) of the questioned jurors indicated where they lived by larger section, such as South Philadelphia, Germantown etc.; sometimes they also indicated the neighborhood, such as Roxborough, Nicetown etc., but I have not included these data here. 28 additional jurors (or 21%) only indicated the neighborhood. That is, there are residential data on 73% of the 85% (134 of 157) of the venirepersons for which I have the voir dire transcripts.

Philadelphia being the racially divided city it was at the time (and still is), the data once again very strongly suggest what one would have assumed from the start. 25 (or 36%) of the 70 persons who identified themselves by larger area came from Northeast Philadelphia, which in 1980 was almost lilywhite. 15 (or 21%) came from heavily black North Philadelphia. Another 12 (or 17%) came from racially mixed South Philadelphia – but most of these were Italian, i.e., white.

It is similar with the finer grained data for the neighborhoods – they reflect exactly the same picture of a strong statistical likelihood of a racial composition of the venire similar to the racial composition of the city as a whole in 1980, with a tendency, if any, of black underrepresentation – exactly what one would expect from the “prima facie” case made by Lindorff already in May 2007.

This is what the record whose alleged absence in the defense filings the 3rd Circuit judges deplore reflects when subjected to an even closer scrutiny than the one the defense has presented over the years. I will subject this conclusion to further, more exact research in the near future, but one can say already now that the result will hardly be very different.

Even at this preliminary stage, there is overwhelming evidence that the court’s speculation that prosecutor McGill’s 66.7 percent “strike rate” against blacks and the fact that a black juror was at least ten times as likely to not be accepted by him as a white one might be explained by some purported massive black overrepresentation is not only logically, but also factually wrong.

The whole argument claiming that racial data about all 157 venirepersons in the Abu-Jamal case are even relevant to evaluate his Batson claim about racial discrimination in jury selection is transparently illogical and absurd, but everything indicates that once all indisputable facts about the venire composition are in, even this last ditch argument to deny Abu-Jamal “on the merits” relief in the Batson issue will lie in shambles. It will then have turned out as np more than a pseudo-statistical sham to justify a ruling that the court wanted to reach.

Readers in Philadelphia in particular, stay posted. Your help might be needed to find out more and conclusively establish these facts. You can contact Abu-Jamal News, or myself: mikschiff@t-online.de.

Michael Schiffmann, April 17, 2008

[1] The habeas petition identifies 15 instead of 14 black persons as contenders for the jury proper, but I will ignore this fact here since the race of the additional person continues to be unknown and since it alters nothing substantial.

No comments: