THE TESTING COLUMN - Issue? - Understanding MBE Score Differences Using SD Units

Sep 21, 2026

Uploaded by Aditya

THE TESTING COLUMN explains how to interpret Multistate Bar Examination (MBE) scores using standard deviation (SD) units. It presents a normal “bell” curve model, mapping SD units to MBE scale points, and describes how examinees are distributed across score intervals. Using examples from 2001–2002 and February–July administrations, it shows how mean score changes (e.g., 5.9 points) translate into fractions of an SD and how that magnitude relates to expected impact on pass/fail rates, while noting that SD units do not identify the causes of score shifts.

LanguageEnglish

CategoryExam

FormatPDF

Pages3

Reading Time8 min

AccessFree to download

The Testing Column

Normal “Bell” Curve and SD Units

Mapping MBE Scale to SD Units

Interpreting Score Differences Across Administrations

SD Units and Pass/Fail Rate Impact

Originally appearing in THE BAR EXAMINER, Vol. 72, No. 2, May 2003. Reprinted with permission of the National Conference of Bar Examiners. All rights reserved. THE TESTING COLUMN by Susan M. Case, Ph.D. that MBE scores for most examinees are between 95 and 185, a range of 90 points (6 x 15) . Figure A below shows a typical bellshaped curve. The horizontal axis has two labels. The top label (“SD units”) shows the points on the scale in terms of standard deviations, ranging from-3.0 SDs to +3.0 SDs. The height indicates the percentage of examinees who scored at a particular level. The highest point of the curve is at “0.0” which indicates the mean of the distribution of scores. Approximately one-third of the examinees have scores between the mean and-1 SD below the mean (the section labeled C); FIGURE A: Normal “Bell” Curve THE TESTING COLUMN 43 approximately one-third of the examinees have scores between the mean and +1 SD above the mean (section D) . Approximately 95 percent of the examinees have scores within 2 SDs above and below the mean (sections B, C, D, and E) . The second label for the horizontal axis (“MBE scale”) represents a typical set of MBE scores, with a mean of 140 and an SD of 15. The label “140” corresponds to the “0.0” on the SD scale, indicating that 140 is the mean of theMBE scores. The following table shows the translation of MBE scores into SD units. | Scale in SD units | Scale in MBE units | | --- | --- | | -3.0 SDs | (140) - (3 X 15) = 95 | | -2.0 SDs | (140) - (2 X 15) = 110 | | -1.0 SD | (140) - (1 X 15) = 125 | | 0.0 | (140) - (0 X 15) = 140 | | +1.0 SD | (140) + (1 X 15) = 155 | | +2.0 SDs | (140) + (2 X 15) = 170 | | +3.0 SDs | (140) + (3 X 15) = 185 | Figure A shows that the one-third of examinees who are included in Section C had MBE scores between 140 (the MBE mean) and 125 (the MBE mean minus 1 SD); another one-third of the examinees who are included in Section D had scores between 140 and 155 (140 + 15). About 95 percent of the examinees had scores between 110 and 170 (the mean minus 2 SDs to the mean plus 2 SDs); these examinees who fall within plus or minus 2 SDs ofthe mean are included in the middle four sections of the graph (labeled B, C, D, and E). Most of the remaining 5 percent of examinees are included in the tails ofthe curve in the sections labeled A and F. This is a lot of background, included to make a point about how to interpret differences in scores from one examination to another. Generally, a difference of 0.15 SDs or more would be considered meaningful. For the MBE, 0.15 SDs equals 2.25 scale points (0.15 x 15=2.25), so a mean change of slightly over 2 scale points may be considered a meaningful difference. The normal curve shown in Figure Amay also be used to illustrate the impact of a decline in scores on the pass/fail rate. If the pass/fail standard is near the mean, at the peak of the curve, a small change in mean score is likely to have a large impact on the pass/fail rate. This is a result of the fact that, at the peak of the curve, a large proportion of examinees have the same score. On the other hand, if the pass/fail standard is away from the mean in either direction (near either tail of the distribution), a small change in performance will have a much smaller impact on pass/fail rates since few examinees have scores in that range. FIGURE B: Distribution of MBE Scores for February and July 2002 44 THE BAR EXAMINER, MAY 2003 Figure B shows the 2002 MBE results for the February and July administrations. In this graph, the horizontal axis indicates the scaled score, and the vertical axis indicates the percentage of examinees who obtained each scaled score. Vertical lines are drawn at 135 and 141 to show the mean for February (135) and the mean for July (141) . Note that the July curve is somewhat to the right of the February curve, illustrating graphically the fact that July scores tend to be higher than February scores. The difference in

FAQs

What are SD units and how do they relate to MBE scores?

SD units express points on the score scale in terms of standard deviations from the mean. The article maps SD units to MBE scale values using the mean of 140 and an SD of 15.

How is the normal “bell” curve used to interpret examinee performance?

The article uses the bell curve to show where examinees fall relative to the mean, including one-third between the mean and -1 SD, another one-third between the mean and +1 SD, and about 95% within 2 SDs above and below the mean.

Why can small changes in mean scores affect pass/fail rates differently?

When the pass/fail standard is near the peak of the curve, a small mean change can move many examinees across the threshold. If the standard is near a tail, fewer examinees are in that range, so the pass/fail impact is smaller.

Sep 21, 2026

Uploaded by Aditya