Which law schools best prepare students for the bar exam?

Spoiler: I don’t know

Photo by Debby Hudson on Unsplash

Evaluating the effectiveness of law schools is challenging for two reasons. The first hurdle lies in defining effectiveness. What does it mean for one law school to provide a better legal education than other schools? Generally, this centers on producing effective lawyers, although such a definition simply punts the definition of effectiveness up a level. Regardless, you cannot be an effective lawyer without actually becoming a licensed attorney. And to do this, you have to pass the bar exam. Thus, let’s make ‘preparing students to pass the bar exam’ one measure of law school effectiveness.

Now for the second hurdle. How do we measure whether a school effectively prepares students for the bar exam? This question is causal in nature. For example, let’s take a student deciding on which law school to attend. She knows she is going to take the 2023 North Carolina bar exam. In evaluating law school effectiveness, we want to know her probability of bar passage in NC, based solely on her pre-law school characteristics, given various law schools. These differences in probabilities between law schools shed light on each law school’s effectiveness in preparing students for the bar exam.

This post is my attempt to answer this question. I lay out the problems with my approach at the end, but the bottom line is that I’m not convinced that it’s possible to answer this question with current data sources. But, I’ll let readers be the final judge.

Overview of Method

As already stated, we want to know how well law schools prepare students for the bar exam. We’ll do this by predicting each school’s bar passage rate based on school-level bar passage predictors and then measure how well each school performs against its predicted passage rate. To account for student quality, we will incorporate the school’s median, 25th percentile, and 75th percentile undergraduate GPA and LSAT score of incoming students.

We’ll also account for first-year attrition. The reason is that schools typically flunk-out poor performing students and these students have a low probability of bar passage. Thus, the quickest way for a school to raise its bar passage rate is to kick out more low-performing students.

Exploration of LSAT, undergrad GPA, and attrition as predictors

AccessLex Institute aggregates the data law school data. We’ll use this data to find each school’s yearly median, 25th percentile, and 75th percentile undergraduate GPA and LSAT score of incoming students, attrition rate, and bar passage rate by state and year. Years in this post represent the year students took the bar exam. For admissions factors such as median undergrad GPA and LSAT this means that the years represent the median values for students who entered law school school three years prior to the stated year.

Relationship between bar passage and both undergraduate GPA and LSAT scores

Prior to modeling, however, we’ll examine the individual relationships between the predictors and bar passage. This will provide us initial confirmation that we should at least test them in models. Figure 1 shows the relationship between both median undergrad GPA (top plot) and median LSAT score (bottom plot) and bar passage. Each reveal a positive relationship between the admissions factor and bar passage.

The top plot shows the relationship between a school's median undergrad GPA and bar passage rate. The bottom plot highlights the relationship between median LSAT and bar passage. Both median undergrad GPA and median LSAT correlate with bar passage.

Figure 1: The top plot shows the relationship between a school’s median undergrad GPA and bar passage rate. The bottom plot highlights the relationship between median LSAT and bar passage. Both median undergrad GPA and median LSAT correlate with bar passage.

Correlation between median LSAT and median undergrad GPA

We’ll next look at the correlation between median undergrad GPA and median LSAT. Our prior is that these two predictors are highly correlated, as more elite schools have higher median undergrad GPAs and LSAT scores. This could cause colinearity problems within our models, which might widen the uncertainty of predictions.

Figure 2 confirms our prior: median undergrad GPA and median LSAT are highly correlated. In the plot, both predictors are standardized to have a mean of 0 and standard deviation of 1, allowing us to compare their linear relationship - the blue line - with a prefect correlation represented by a line with a slope of 1 - the red dashed line. The plot shows an extremely strong relationship between median undergrad GPA and median LSAT. The linear relationship is only slightly flatter than the perfect relationship (red dashed line).

Correlation between each school's standardized median LSAT and standardized median undergrad GPA. Blue line is the fit line of the data, red line is a hypothetical perfect fit line - a slope of 1. LSAT and GPA are highly correlated, with the actual fit line mirroring the perfect fit line.

Figure 2: Correlation between each school’s standardized median LSAT and standardized median undergrad GPA. Blue line is the fit line of the data, red line is a hypothetical perfect fit line - a slope of 1. LSAT and GPA are highly correlated, with the actual fit line mirroring the perfect fit line.

Association between 1L attrition and both median LSAT and median undergrad GPA

Let’s also look at the association between first year (1L) attrition and both median undergrad GPA and median LSAT. Our expectation, based on prior knowledge, is that schools with lower median undergrad GPA and LSAT scores will also have higher attrition. Figure 3 confirms this prior. But, the relationship is not too strong.

The top plot shows the relationship between a school's median undergrad GPA and 1L attrition. The bottom plot shows the relationship between median LSAT and 1L attrition. Both median undergrad GPA and median LSAT correlate slightly with 1L attrition.

Figure 3: The top plot shows the relationship between a school’s median undergrad GPA and 1L attrition. The bottom plot shows the relationship between median LSAT and 1L attrition. Both median undergrad GPA and median LSAT correlate slightly with 1L attrition.

Association between 1L attrition and bar passage

Finally, figure 4 highlights the association between attrition and bar passage. Not surprisingly, it’s negative: schools with higher attrition rates have lower bar passage rates. This finding is not unexpected because we saw from figure 1 that higher median undergrad GPA and LSAT scores correlate with higher bar passage rates and figure 3 showed us that lower median undergrad GPA and LSAT scores are associated with higher attrition rates. Putting both findings together, we would assume that higher attrition rates correlate with lower bar passage rates.

Correlation between bar passage rates and 1L attrition. There is a slight negative relationship, but there is also a lot of variance in the relationship.

Figure 4: Correlation between bar passage rates and 1L attrition. There is a slight negative relationship, but there is also a lot of variance in the relationship.

Difference between predicted and actual bar passage

Models with admissions factors as predictors

We now turn to the heart of this post. We’ll predict each school’s bar passage rate and see how actual rates stack up against predictions. But, we need a good model to do this. We know our potential predictors: median, 25th percentile, and 75th percentile undergraduate GPA and LSAT score, and first year attrition. But, we don’t know the functional form of the model that provides the best predictions. Do we need all predictors or only a couple of them?

We will use two steps to search for the best model. First, we will conduct model comparison and validation on four different models that only include the undergrad GPA and LSAT predictors, along with year and state. Then, we will take this model and compare it against models with attrition as an additional predictor. Models will be compared by testing their predictions out of sample using leave-one-out cross validation (LOO). Posterior prediction checks will also be used to ensure the fit is reasonable.

All models are Bayesian hierarchical logistic regressions. The response is a school’s bar passage rate for a given state and year, with the model weighted by the number of takers. All models have the same group-level predictors: state and year. They only differ in their school-level predictors. The four initial models are below:

  1. Linear: median undergraduate GPA and median LSAT score for the given graduating class year.

  2. PCA: Figure 1 showed that median undergraduate GPA and median LSAT score are highly correlated, raising fears of colinearity. The 25th, median, and 75th percentiles of each metric are also highly correlated with themselves (the 25th percentile undergrad GPA is correlated with the median undergrad GPA). We’ll use Principal Component Analysis (PCA) to reduce the six highly correlated percentile variables into a smaller number of uncorrelated ones, while maintaining as much of the information in the variables as possible. The table below shows that the first two principal components contain 98% of the variance in the six LSAT and GPA variables. By only using these two principal components as predictors, we get 98% of the variance with only two uncorrelated variables.

PC 1 PC 2 PC 3 PC 4 PC 5 PC 6
Standard deviation 2.24 0.42 0.22 0.14 0.11 0.07
Proportion of Variance 0.95 0.03 0.01 0.00 0.00 0.00
Cumulative Proportion 0.95 0.98 0.99 1.00 1.00 1.00
  1. PCA spline: To account for possible non-linear relationships between the two PCA variables and bar passage, we’ll examine a model with splines added to the two PCA variables.

  2. Spline: This model contains median undergraduate GPA and median LSAT score like model 1, but adds splines to them.

Compare admissions factors models using LOO

The table below compares each model by its expected log posterior density (ELPD). The Difference in ELPD column shows the difference in ELPD between the best fitting model - the model on the first row - and the model in the given row. The PCA spline model performs best. Plus, the standard error of the difference in ELPD between the best and second best model is small enough that we can have confidence that the PCA spline model did not perform best due to some randomness in the sampling draws. Assuming the model’s posterior predictive check pans out, we’ll start with it when adding attrition.

Model Difference in ELPD Std. Error of Difference
PCA spline 0 0
PCA -52 18
Spline -111 18
Linear -157 26

Poserior predictive checks of admissions factors models

Posterior predictive checks (PPC) simulate y values - bar passage rates - from the models. Our checks simulate 200 distributions of bar passage rates based on 200 model simulations. We can then compare these distributions to the actual distribution of rates. Ideally, they align.

Figure 5 shows the PPC plots. All models have the same problem: they overpredict at the 75% to 90% bar passage rates and underpredict at the tails. This is shown by the dark line (actual distribution) being lower than the light lines (simulated distributions) at the 75% to 90% range and higher than the light lines outside this point. The takeaway from the misfit models is that there will be too many predictions in the 75% to 90% range and not enough at each tail.

Posterior predictive check for all models. A problem in all models is that they fail to accurately predict at the most likely bar passage rates. We know this because at the highest points in the curve, the predicted passage rates for schools is higher than the actual rates.

Figure 5: Posterior predictive check for all models. A problem in all models is that they fail to accurately predict at the most likely bar passage rates. We know this because at the highest points in the curve, the predicted passage rates for schools is higher than the actual rates.

Add attrition as a predictor

Step two involves using the best model in the previous section, PCA with a spline, and adding attrition as a predictor. We’ll incorporate attrition using the natural logarithm of attrition. The natural logarithm is used because attrition rates have a right skew: most schools have attrition rates between 0% and 15%, but some schools have rates greater than 30%. Using the natural logarithm removes much of the right skew and makes the distribution more symmetrical.

We’ll create two different models with attrition, each an extension of the PCA spline model. The first adds the natural logarithm of attrition, centered, as an individual predictor. The second also includes the natural logarithm of attrition, but wraps it in a spline to model non-linearity.

The table below compares each model based on the ELPD using leave-one-out cross validation. We see little difference between the two models with attrition added, but these two models are better than the model without attrition.

Model Difference in ELPD Std. Error of Difference
PCA spline with attrition spline 0 0
PCA spline with attrition -2 8
PCA spline w/o attrition -11 12

Moving on to the model check, the plot of posterior predictive checks in figure 6 shows that the models with attrition have the same problem we encountered in figure 5. More predicted pass rates fall between 70% and 90% than would be expected given the distribution of actual bar passage rates.

Despite the non-ideal fit, the PCA spline model with attrition will be used to predict bar passage rates of schools. This model performs almost identically to the PCA spline model with attrition added as a spline, but is simpler. All things equal, we’ll opt for the simpler.

Posterior predictive checks. The attrition model cehcks reveal the same problem as the models without attrition - too many predictions in the 75% to 90% range.

Figure 6: Posterior predictive checks. The attrition model cehcks reveal the same problem as the models without attrition - too many predictions in the 75% to 90% range.

Comparing actual and predicted bar passage rates for all US law schools

With technicalities behind use, let’s get to the results. Figure 7 compares each law school’s predicted and actual bar passage rates from 2014 to 2019. The left plot shows the 95% credible intervals for each school’s predicted bar passage rate (in red) and its actual rate (blue). The right plot, meanwhile, displays the 95% credible interval for each school’s residual. The residual is the actual rate minus the predicted rate. Finally, the percentages on the right represent the probability that the school’s actual rate is higher than its predicted rate. Stated differently, it’s the probability that the residual is positive.

Left plot reveals actual (blue dot) and predicted (red dot) bar passage rates. Right plot shows residuals: actual minus predicted bar passage rates. The percentage on the right is the probability that the actual bar passage rate is higher than the predicted rate.

Figure 7: Left plot reveals actual (blue dot) and predicted (red dot) bar passage rates. Right plot shows residuals: actual minus predicted bar passage rates. The percentage on the right is the probability that the actual bar passage rate is higher than the predicted rate.

Three problems with the methodology

What do we make of the plots? For starters, simply ranking each school by its residual is ill-advised for three reasons.

  1. Schools have a lot of uncertainty in their residuals. A school’s ranking is it’s true residual plus uncertainty in the residuals; the dreaded error terms in models. The higher the uncertainty, the more that the ranking for specific schools will simply reflect randomness. Here, the uncertainty is large enough to where rankings will reflect more randomness than we are comfortable incorporating. Figure 7 illustrates this. It shows the University of Kentucky’s residual ranking among all 201 law schools based on 500 simulations of the model. The meidan simualted ranking is 84 and the 90% credible interval stretches from 28 to 148. This range is too wide to be helpful.
The University of Kentucky's residual rankings from 500 model simulations. There is a wide degree of variance in the rankings, making them unhelpful.

Figure 8: The University of Kentucky’s residual rankings from 500 model simulations. There is a wide degree of variance in the rankings, making them unhelpful.

  1. Schools with high LSAT and undergrad GPAs will also have high predicted bar passage rates. Since neither actual nor predicted rates can exceed 100%, this leaves schools with high LSAT and undergrad GPAs little room to outperform their predictions. When you are already at the top, you can’t go much higher.

  2. The predictions are only as good as the model. As figure 6 shows, the model is not ideal. It potentially misfits by different amounts depending on the predicted bar passage rate.

Comparing similair schools

Taking reasons two and three together, a better approach might be to compare similar schools. For example, with might compare schools with similar admissions factors and where the majority of bar takers are from the same state. To highlight this method, we’ll compare three North Carolina schools with similar admissions factors: Campbell University, Elon University, and North Carolina Central University.

Figure 9 shows the comparison’s results. It’s the difference between each school in their difference of residuals, only incorporating North Carolina bar exam results. To account for uncertainty, the entire probability distribution is shown for these differences.

Here’s how it works. Let’s say that Campbell’s actual NC bar passage rate is 70% and their predicted rate is 60%, for a difference of 10 percentage points. Elon’s actual NC rate and predicted rate are both 60%, for a difference of 0. In this situation, Campbell’s residual - difference between actual and predicted bar passage rate - is 10 percentage points higher than Elon’s (10 - 0).

Difference in residuals between comparitive law schools in North Carolina. Predictions and residuals are limited to the NC bar exam.

Figure 9: Difference in residuals between comparitive law schools in North Carolina. Predictions and residuals are limited to the NC bar exam.

OK, so this might be useful. But, I’m still not comfortable with it due to the problems previously mentioned. The best course of action is to scrap the whole enterprise. But, in the spirit of learning from what goes wrong, I figured I would post it here for posterity.

Addendum: the final rankings that have little value

If you ignored this post’s advice and want to know the actual modeled ranking, here you go. Again, you shouldn’t make much of them; I’m not. And you especially shouldn’t make much of them once you see the wide 90% credible intervals. Also, going back to problem two mentioned above, notice that almost all of the top ranking schools have low LSAT / GPA combinations and low bar passage rates.

Rank 90% Credible Interval
Lower bound Upper bound
Regent University 1 1 7
Northern Illinois University 2 1 11
California Western School of Law 3 1 10
Syracuse University 3 1 11
Western State College of Law 3 1 13
Washington and Lee University 6 1 19
Liberty University 7 1 108
Duquesne University 8 3 20
Florida International University 8 3 19
University of La Verne 10 1 46
Belmont University 11 2 35
Concordia Law School 11 1 105
University of Baltimore 11 7 29
Southern University 14 4 47
McGeorge School of Law 15 6 43
University of Massachusetts Dartmouth 16 1 77
Loyola Marymount University-Los Angeles 17 10 36
UNT Dallas College Of Law 18 1 110
Lincoln Memorial University 19 1 149
Seton Hall University 20 8 48
Widener Commonwealth 20 3 79
Campbell University 22 5 56
Georgia State University 22 10 49
Mitchell Hamline School of Law 24 1 111
Quinnipiac University 25 5 79
St. John's University 25 13 51
University of Richmond 25 8 72
University of Southern California 28 13 63
Texas A&M University 29 15 70
Pennsylvania State University-Dickinson School of Law 30 1 194
University of Georgia 30 17 69
City University of New York 32 11 88
Florida Coastal School of Law 32 13 82
University of San Diego 34 14 74
North Carolina Central University 35 10 105
Capital University 36 12 98
University of California-Davis 37 17 89
Thomas Jefferson School of Law 38 11 94
Gonzaga University 39 8 110
William and Mary Law School 39 17 83
Albany Law School of Union University 41 18 93
Florida A&M University 41 11 104
Indiana University - Bloomington 41 14 92
University of New Hampshire 41 5 124
University of Illinois 45 16 86
John Marshall Law School 46 24 93
Santa Clara University 47 19 97
University of Nebraska 48 16 110
University of California-Irvine 49 13 109
New York University 50 44 75
Ohio Northern University 51 1 168
South Texas College of Law 52 32 99
University of Tulsa 52 12 123
Mercer University 54 15 117
Cornell University 55 33 95
Willamette University 55 12 144
Texas Tech University 57 28 104
Baylor University 58 26 108
Fordham University 58 37 90
University of Utah 58 12 118
University of Arkansas-Little Rock 61 13 143
St. Mary's University 62 23 113
University of Michigan 62 38 89
Vanderbilt University 64 29 109
St. Thomas University (Florida) 65 18 122
University of Iowa 65 22 120
University of Dayton 67 12 158
Cleveland State University 68 22 133
Mississippi College 68 8 148
University of Pennsylvania 68 46 103
Pace University 71 32 129
Roger Williams University 72 14 156
Boston University 73 45 110
University of Kansas 73 23 131
University of Virginia 73 47 103
William Mitchell College of Law 73 17 164
Faulkner University 77 9 156
University of California-Berkeley 77 46 109
Saint Louis University 79 30 120
University of Missouri 79 29 126
University of South Carolina 79 23 130
University of Oklahoma 82 37 123
University of Kentucky 83 28 148
Boston College 84 51 122
Louisiana State University 85 34 132
Chapman University 86 34 137
Ohio State University 87 48 127
Villanova University 87 40 137
Arizona Summit Law School 89 26 155
Drake University 89 18 151
Duke University 89 52 120
University of Alabama 92 46 138
University of Montana 93 15 170
Howard University 94 20 160
University of Denver 95 41 139
University of New Mexico 96 20 148
University of Oregon 97 36 168
University of North Dakota 98 17 184
Hamline University 99 16 193
New England Law | Boston 99 35 147
University of Pittsburgh 99 42 152
Arizona State University 102 48 146
Nova Southeastern University 103 52 151
Columbia University 104 83 124
University of Cincinnati 105 52 163
Brooklyn Law School 106 65 139
University of Maryland 107 56 155
University of Washington 108 66 153
University of South Dakota 109 17 195
Florida State University 110 69 154
University of Louisville 111 40 175
Stetson University 112 75 165
University of Colorado 112 61 161
Ave Maria School of Law 114 30 184
Pepperdine University 115 52 162
University of St. Thomas (Minnesota) 115 46 168
Washburn University 115 35 173
University of North Carolina 118 73 161
Harvard University 119 98 130
University of Missouri-Kansas City 120 58 162
University of Chicago 121 88 141
Yale University 122 98 144
Georgetown University 123 100 142
Samford University 123 55 182
University of Maine 125 32 191
West Virginia University 126 38 194
Case Western Reserve University 127 60 170
University of Hawaii 128 39 189
Catholic University of America 129 57 186
Loyola University-Chicago 129 78 163
University of California-Los Angeles 131 94 152
Touro College 132 73 177
Southern Methodist University 133 90 161
University of Akron 134 68 177
Drexel University 135 74 180
Western New England University 135 48 191
Loyola University-New Orleans 137 74 179
Michigan State University 137 86 174
Stanford University 137 101 160
Widener University-Delaware 137 66 187
University of Nevada - Las Vegas 141 55 190
Cardozo School of Law 142 96 160
Seattle University 143 82 175
Lewis and Clark College 144 87 188
Indiana University - Indianapolis 145 78 175
University of Miami 146 100 170
University of Notre Dame 147 88 182
Chicago-Kent College of Law-IIT 148 96 169
Northeastern University 149 91 175
George Mason University 150 88 182
University of Connecticut 151 83 183
Southern Illinois University-Carbondale 152 77 195
University of Memphis 153 77 192
University of Idaho 154 69 195
Brigham Young University 155 97 186
Pennsylvania State University-Penn State Law 155 34 198
University of Texas at Austin 157 126 166
Charleston School of Law 158 80 197
University of Florida 159 122 175
University of Wyoming 159 51 198
Wake Forest University 159 102 191
University of Mississippi 162 87 192
Texas Southern University 163 100 191
Charlotte School of Law 164 80 196
Barry University 165 110 193
George Washington University 165 139 174
Western Michigan University 165 117 184
University of Toledo 168 96 197
Temple University 169 130 184
University of Tennessee 169 121 193
University of Arkansas-Fayetteville 171 92 197
Wayne State University 171 131 193
Creighton University 173 96 198
University of Arizona 173 115 195
University of Houston 175 138 185
Washington University 176 144 185
Southwestern Law School 177 142 192
University of Minnesota 177 145 188
Vermont Law School 177 95 199
New York Law School 180 147 190
Golden Gate University 181 122 195
Northwestern University 182 159 188
Valparaiso University 183 121 198
Depaul University 184 160 194
University of Buffalo-SUNY 184 155 195
Northern Kentucky University 186 148 199
Suffolk University 186 159 193
Emory University 188 166 193
Oklahoma City University 189 145 198
University of Detroit Mercy 190 157 198
Rutgers University 191 164 198
Whittier Law School 191 157 198
University of California-Hastings 193 179 197
Elon University 194 167 200
Atlanta's John Marshall Law School 195 176 199
Tulane University 195 182 197
Hofstra University 197 190 198
Appalachian School of Law 198 30 201
University of San Francisco 199 197 201
American University 200 199 201
District of Columbia 200 198 201
Shane Orr
Shane Orr
Education Analyst

Spent six years playing Army, another three as a lawyer, and finally settled into being a data and programming guy.

Related