[The following post was originally published on Law360.com on February 19, 2014.]

With the publication of “The Behavior of Federal Judges,” by Lee Epstein, William M. Landes and Judge Richard A. Posner, rigorous statistical analysis of the appellate courts is beginning to move from academic publications to mainstream bar journals. Although academic analysts have focused largely on the federal appellate courts — the United States Supreme Court in particular — my focus for the past several years has been on the civil docket of the Illinois Supreme Court.

For 2013, the court decided 34 civil cases (not including attorney discipline and juvenile matters). More than 80 percent of the civil docket consists of appeals taken from final judgments and orders. The court decided four civil cases, each where the primary issue was civil procedure, domestic relations and constitutional law, as well as three cases each in insurance, wills and estates, and workers compensation. In addition, the court decided two cases each in the areas of taxation, labor law, tort and public pensions.

Not surprisingly, a dissent at the Appellate Court helps in getting the court’s attention — 29.4 percent of the civil cases involved dissents below, right in line with the court’s trend in recent years. The court rarely allows petitions for leave to appeal from unpublished decisions (known in Illinois as Rule 23 orders) — only 8.8 percent of the civil docket in 2013.

The court decided 58.8 percent of its civil cases unanimously. This is similar to the court’s experience in 2012, when the unanimity rate was 52.6 percent, but significantly below the court’s unanimity rate for most of the past decade. From 2003 to 2005 and 2007 to 2011, the court’s unanimity rate in civil decisions fell along a narrow range, from a low of 69.8 percent in 2003 to a high of 82.1 percent in 2009. The only exception was in 2006, when the cCourt dipped to 56.5 percent.

As always, the court produced decisions much more quickly in 2013 when there was no dissent. Unanimous decisions came down an average of 103.7 days after oral argument, while cases with dissenters took much longer — 185.8 days after argument. The court’s average lag time on nonunanimous decisions has been relatively static since 2011, but the average lag time on unanimous decisions has been cut by more than three weeks in that time.

The court reversed in 55.9 percent of its civil decisions in 2013. With the exception of 2012 (78.4 percent) and 2009 (75.7 percent), the court’s reversal rate has narrowly fluctuated around the 50 percent mark since 2003. Since 2003, the court has reversed in 57.99 percent of its civil cases.

Every year at the end of the United States Supreme Court’s term, the legal press reports on the rise and fall of reversal rates for the federal circuit courts. The problem with overemphasizing this statistic is that in any single year, an intermediate court’s reversal rate is based on a small number of cases. This is particularly true for my work on the state Supreme Court’s civil docket, so rather than focusing on year-to-year ups and downs, I look for sustained deviations from the norm over time.

The single biggest part of the court’s civil docket comes from Chicago’s First District, which comprises between 30 and 40 percent of the case load each year. Reversal rates in four of the six divisions of the First District (the Second, Third, Fifth and Sixth) were down in 2013 from 2012. Since 2003, four of the six divisions’ reversal rates are clustered between 50 and 60 percent. The Third Division is a bit higher (61.8 percent), and the Fourth a little lower (44.4 percent).

Both the Second and Third Districts saw lower reversal rates in 2013 — 60 percent for the Second, 50 percent for the Third, but in both cases, the courts were reverting to form. Since 2003, 61 percent of the Second District’s civil decisions reviewed by the court have been reversed, while 52.5 percent of the Third District’s decisions have been.

Last year, I noted that Springfield’s Fourth District had shown the lowest reversal rate in the state — only 25 percent. This was part of a three-year swing in the numbers, with only 30 percent of the court’s decisions reversed between 2010 and the end of 2012. But in 2013, the Fourth reverted to its long-term pattern as the Supreme Court reversed in six of 10 civil cases. Since 2003, the reversal rate for the Fourth District is 53.1 percent.

Many observers consider the Fifth Appellate District to be the most pro-plaintiff appellate court in the state. The Supreme Court’s response has been relatively consistent: In seven of the 11 years since 2003, the Fifth’s reversal rate has been 67 percent or more, and 2013 was no exception. The Fifth District leads the state for the entire period with a 75.9 percent reversal rate.

Justices Anne B. Burke and Lloyd A. Karmeier wrote for the court’s majority most often this past year, with seven majority opinions apiece in civil cases. Justice Robert R. Thomas added six, Chief Justice Rita B. Garman wrote five and Justice Mary Jane Theis four.

Collectively, written dissents were down 20 percent in 2013 over 2012. Justices Thomas L. Kilbride and Charles E. Freeman, who wrote the fewest majority opinions in civil cases, wrote the most dissents — five and three, respectively. Justices Burke and Thomas filed two dissents apiece, with the other justices dissenting only once each in civil cases.

In order to study individual justices’ voting patterns, I next considered how often each justice votes when the court is divided. Chief Justice Garman and Justice Mary Jane Theis each voted with the majority in 92.9 percent of the court’s nonunanimous civil decisions last year. Justice Thomas voted with the majority in 84.6 percent of such cases in 2013, almost identical to his percentage in 2012 (83.3 percent). Justice Karmeier voted with the majority in 78.6 percent of the court’s nonunanimous civil decisions, only slightly down from 2012 number. Only Justice Freeman’s percentage was slightly increased, voting with the majority in divided cases 78.6 percent of the time in 2013, up from 63 percent a year earlier. Justice Kilbride voted with the majority in 46.2 percent of nonunanimous cases.

The court’s center is even more sharply defined when we limit the database to two and three-dissenter decisions. In such decisions, the chief justice was in the majority every time. Justice Theis joined the majority decision in 85.7 percent of such cases, with Justices Thomas and Karmeier voting with the majority 71.4 percent of the time. Most often in the minority were Justice Freeman, voting with the majority in 57.1 percent of closely divided cases; Justice Kilbride, with the majority half the time; and finally, Justice Burke, who voted with the majority in 42.9 percent of closely divided cases.

I turned next to agreement rates between pairs of justices. For 2013, Chief Justice Garman and Justice Theis voted together in 85.7 percent of nonunanimous civil cases. The chief voted with Justice Thomas in 84.6 percent of such cases, and with Justice Karmeier 71.4 percent of the time. Similarly, Justice Theis voted with Justice Thomas in 76.9 percent of such cases and with Justice Karmeier 71.4 percent of the time. Justices Thomas and Karmeier voted together in 76.9 percent of nonunanimous civil cases.

On the other hand, Justice Burke voted with Chief Justice Garman and Justices Karmeier and Theis in 57.1 percent of such cases. Justice Burke voted with Justice Thomas 46.2 percent of the time. Similarly, Justice Kilbride voted with the chief justice in 38.5 percent of nonunanimous civil cases, with Justice Thomas in 58.3 percent, and with Justice Theis in 53.8 percent of such cases. Because Justice Freeman voted with the majority in each of the court’s seven one-dissenter cases, his agreement rates are a bit higher — 85.7 percent with Justice Burke, 71.4 percent with Chief Justice Garman, 61.5 percent with Justice Thomas, 57.1 percent with Justice Karmeier and 71.4 percent with Justice Theis.

When appellate specialists get together, we frequently debate whether or not an experienced appellate attorney should be able to predict the outcome of a case or even the vote at the conclusion of an oral argument. To study whether this question can be approached objectively, I added data on questioning patterns to my study.

The court asked 848 questions during arguments of civil cases decided during 2013: 417 to appellants during their opening remarks, 316 to appellees and 115 to appellants during rebuttal. Justice Thomas asked 222 questions. Justice Theis was second with 171. Justice Burke was third with 126 questions. After Justice Burke came Chief Justice Garman with 109 questions, Justice Karmeier with 101, Justice Freeman with 61 and Justice Kilbride with 58 questions. Appellants were asked an average of 15.4 questions per argument, appellees 9.3.

There are a lot of theories among appellate lawyers about questions from the court. Some lawyers insist the justices sometimes ask questions to play devil’s advocate, or to attempt to persuade another justice. There is no evidence to support either of these theories in the court’s 2013 civil arguments. Rather, the court’s questions tend to indicate that the inquiring justice may be having difficulty with that side’s argument. Losing appellants average 17.7 questions per argument to 13.6 for winners. Similarly, losing appellees average 10.6 questions per argument, while winners average 7.9 per argument.

In nonunanimous affirmances, appellants averaged 20.0 questions to 15.6 for appellants in unanimous cases. Appellees received an average of 8.9 questions in nonunanimous decisions to 7.0 in unanimous decisions. But in nonunanimous reversals, the difference was far smaller: Appellants received 14.4 questions in nonunanimous cases, compared to 13.2 in unanimous decisions. Appellees received an average of 9.3 questions in nonunanimous cases, but more in unanimous decisions: an average of 11.1 questions.

In order to account for the effect of complex cases on the data, I next asked whether the difference between total questions asked each side in a particular case might suggest a probable winner.

The answer is yes, at least in 2013. Appellees received more questions than their opponents in eight civil cases; they lost seven of eight. Appellants received more questions in 24 cases, winning 11 (the sides received an equal number of questions in two cases). The more an appellant’s total questions exceeded the appellee’s, the more likely the court would ultimately affirm. Losing appellants averaged 9.8 questions more than their opponents, while winning appellants averaged only 3.06 questions more than their adversaries.

Each of the seven justices averages more questions to appellants than appellees. For some justices, such as Chief Justice Garman (1.8/1.4), Justice Kilbride (1.0/0.8) and Justice Karmeier (1.7/1.4), the difference was small, but Justices Burke (2.5/1.5), Freeman (1.4/0.5), Thomas (4.2/2.7) and Theis (3.5/1.3) tended to ask appellants significantly more questions on average.

Dividing the data into unanimous and nonunanimous decisions does not make a consistent difference in the justices’ patterns. Justices Burke (1.8/3.1), Kilbride (0.7/1.5), Thomas (3.7/4.8) and Karmeier (1.1/2.4) averaged more questions to appellants when the court wound up divided, but Justices Freeman (1.6/1.1) and Theis (4.3/3.1) averaged fewer. Chief Justice Garman (0.3/4.8), Freeman (0.4/0.6) and Karmeier (1.0/1.8) averaged more questions to appellees in cases decided unanimously, but Justices Burke (2.1/0.9), Kilbride (0.9/0.5) and Thomas (3.0/2.5) averaged fewer.

However, several Justices’ questioning patterns might be suggestive of how they will ultimately vote. Five of the seven justices — Burke (2.4/2.2), Kilbride (1.3/0.8), Thomas (6.4/2.4), Karmeier (2.4/1.0) and Theis (4.4/3.0) — asked more questions of appellants they ultimately voted against than of appellants they voted for. Three justices — Chief Justice Garman and Justices Kilbride and Thomas — asked more questions of appellees they voted against than of appellees they voted for.

This past year suggests three lessons for counsel appearing before the Illinois Supreme Court: (1) for most issues, the court has a centrist voting bloc of Chief Justice Garman and Justices Thomas, Karmeier and Theis; (2) the court does not grant review predominantly to reverse, like some appellate courts with discretionary dockets; and (3) answer every question carefully — there’s a good chance those are the justices you must persuade to prevail.

Image courtesy of Flickr by anneh632 (no changes).