INTRODUCTION
PREVIOUS WORK ON THE IMPACT OF ATTORNEY SKILL
For scholars in all fields, two fundamental problems have stood in the way of investigating the effect of lawyering skill. Here and throughout the paper, “lawyer skill effect” simply refers to the effect of lawyer skill on trial outcomes. designing an experiment to test the effect of prosecutorial and defense skill on trial outcomes).
Concerns about the legal skill effect predate even Justice Sutherland, however, and studies of legal skill began as early as . Due to a lack of data, twentieth-century researchers had little success determining the effect of lawyering skills on trial outcome. Abrams and Yoon could say nothing about the effect of the prosecuting attorney's skill—and whether it was more or less important than the defense attorney's skill.
Moreover, because Abrams and Yoon's analysis only controlled for attorney-matching effects on the defendant's side, their estimate of the effect of defense attorney skill could still have been biased. An oversampling of low-stakes cases only matters if it somehow biases the estimate of the lawyer's skill effect.
GOING BEYOND ABRAMS AND YOON's ANALYSIS
However, perhaps the lawyer's skill makes a difference only in these low-stakes cases; if so, then Abrams and Yoon's ratings would be biased upwards. Although Abrams and Yoon's ratings are more solid than those of their predecessors, they remain problematic. BEYOND THE ABRAMS AND YOONA ANALYSIS: NEW DATA SETAbrams and Yoon's landmark was revolutionary in the. due to the random assignment of public defenders.5 9 However, since this data set actually contains the power of case variables, it allows us to test their claims.
For each case in this data set, the court clerk distributed a two-part survey to each juror, attorney, and judge. The first part was administered at the conclusion of the trial, but before the jury deliberations. the second part was administered after the verdict. Thus, the cases in this data set contain the reaction of each lawyer, juror, and judge to the decision, and their assessment of the lawyer's skills, the complexity of the case, and the evidence.
Again, this data set may go beyond Abrams and Yoon's study because while they used proxies for attorney skill such as education and years of experience, their data did not have a direct measure of attorney skill. About half of the cases resulted in convictions, and most were theft and drug-related crimes. However, prosecutors had, on average, fewer years of experience and prior criminal convictions than defense attorneys.
Notes: Evaluation of hung juries in Bronx County, New York, Los Angeles County, California Maricopa County, Arizona, and Washington, DC data comes from the Inter-Universit3 Consortium for Political and Social Research. Notes: Evaluation of hung juries in Bronx County, New York, Los Angeles County, California, Maricopa County, Arizona, and Washington, DC data comes from the Inter-University Consortium for Political and Social Research. At first glance, it might appear that the researchers somehow oversampled hung jury cases—especially since the title of the data set is Evaluation of Hung Juries.
Yet hung juries are much more common in criminal trials than most people think, as all twelve jurors must reach the same verdict. These detailed and unique observations were previously unobtainable, making broad research into the effect of attorneys' skills in the courtroom impossible. Accordingly, this new data set can provide valuable insight into the influence of attorneys' skills on the trial outcome of the common suspect.
MODELING THE ATTORNEY SKILL EFFECT
In their 2007 study, Abrams and Yoon found that defense counsel skill had a statistically significant effect on trial outcome.65 If Abrams and Yoon's results—not to mention anecdotal. The skill of the prosecuting attorney has a statistically significant effect on trial outcomes, but the skill of the defense attorney does not. In each table, specifications one, two, and three use the judge's, opposing counsel's, and jury's ratings, respectively, as a measure of the lawyer's skill.
Specification four adds the juror's case complexity rating to the juror attorney's skill rating specification. The jury case complexity variable can help control for juror confusion, which can have effects on jurors' perceptions of attorneys and trial outcome. Notes: Hung's evaluation of jurors in Bronx County, New York, Los Angeles County, California, Maricopa County, Arizona, and Washington, DC data comes from the Intercollegiate Consortium for Political and Social Research.
Controls for the race and gender of both the victim and defendant as well as crime fixed effects are included in each specification; for complete results, see Appendix Table I. Controls for the race and gender of both the victim and defendant as well as crime fixed effects are included in each specification; for complete results, see Appendix Table II. Notes: Evaluation of Hung Juries in Bronx County, New York, Los Angeles County, California, Maricopa County, Arizona, and Washington, DC data comes from the Inter-.
Controls for the race and gender of both the victim and the defendant as well as crime fixed effects are included in each specification; for complete results, see Appendix Table III. Notes: Evaluation of Hung Juries in Bronx County, New York, Los Angeles County, California, Maricopa County, Arizona, and Washington, DC data comes from the Inter-University Consortium for Political and Social Research. Controls for the race and gender of both the victim and the defendant as well as crime fixed effects are included in each specification; for complete results, see Appendix Table IV.
Notes: Rating data for Judge Hung in Bronx County, New York, Los Angeles County, California, Maricopa, Arizona, and Washington, DC are obtained from the Intercollegiate Consortium for Political and Social Research. Controls for the race and gender of the victim and the defendant and the fixed effects of the crimes are included in the eacl. In other words, the judge's and opposing counsel's perceptions of the legal skills do not affect the outcome of the trial.
According to specification four, the probability of conviction increases by 30.4 percent for every one-point increase in a juror's rating of the prosecution. Yet a skilled defense attorney has no such effect - jurors' assessment of the defense attorney's skill is never significant.
DISCUSSION AND IMPLICATIONS
The importance of prosecutorial skill also indicates the need for a policy change in the hiring and pay practices of district attorneys. Their high-achieving classmates grabbed all the private sector jobs, and the public sector was the only option. However, the importance of prosecuting attorney qualification levels to trial outcomes suggests that it would be beneficial for the government to increase pay.
At the end of the trial, the prosecutor's office could ask third parties—perhaps victims, victims' families, or even judges—to fill out a survey evaluating the prosecutor's performance.8 0 The office could then require prosecutors with consistently poor ratings to work under prosecutors with good ratings. In fact, the main motivation for the economist Radha Iyengar to study the defense system of the poor was the fact that "[t]he right to an equal and fair trial regardless of wealth is a hallmark of American jurisprudence."84 But contrary to her results and regarding According to the ACLU, the results of this study show that poor defendants receive an equal and fair trial regardless of their wealth. They were so preoccupied with the importance of defense counsel that they forgot about the prosecution.
These results certainly demonstrated the importance of a good prosecutor, and the data set allowed this note to surpass all previous studies of the effect of attorney skill. Since there are only 320 observations in the sample, and 80 percent of the attorneys appointed are indigent defendants, there may not be particularly skilled defense attorneys in the sample. In the event that the court-appointed attorney variable was highly correlated with the measure of defense counsel ability, all regression specifications were reestimated without the court-appointed attorney dummy variable.
However, this possibility is worth mentioning here because the results of the current study do not rule out possible unequal treatment before trial. This note also fails to demonstrate that the effects of counsel's skill last into the sentencing phase. Perhaps when judges or juries sentence convicted defendants, they look only to the strength of the case.
However, the general lack of significance in the sentencing regressions suggests that the effects of attorney skills may be a victim of small sample sizes; only 142 of the defendants in this sample were sentenced to prison. Still, it offers a unique opportunity to observe the effect of raw lawyer skills on trial outcomes without all the problems of Abrams and Yoon's research. The media, the ACLU, and most scientists are disproportionately concerned with the effect of the lawyer's skill on the outcome of the trial.
Instead, this research suggests that we should be much more concerned about the prosecutor's skills. Nevertheless, the results of this study suggest that prosecutor selection can determine who goes to prison and who goes home.
