Equating the MBE
bar exam design, psychometrics
"The only constant is change. In the high-stakes testing world, we have to change almost everything each time we administer a test. Because the stakes are high for the bar examination, there are any number of examinees who would dearly love to see items on the MBE before the test is administered. Thus, we wait for a few administrations before we reuse any question, or item, and we use any given item for only a few test administrations before we quit using it forever. Against this backdrop, we also have to ensure that a score we produce retains its meaning across time and space. A score earned in 2013 in New York should have the same meaning as a score earned in 2014 in Guam. We rely on standardized administration of the MBE to ensure that where an examinee sits for the exam (space) does not affect his or her score. To ensure that whether an examinee takes the MBE in 2009 or 2015 or July or February (time) is not a factor, we rely upon equating.
"Each time NCBE builds an MBE test, we do our best to choose items that will work the same way as items have in the past. The exam is built according to a detailed subject matter blueprint and statistical criteria that ensure comparability of what is measured across time. (One recent update is the addition of Civil Procedure to the February 2015 MBE, which was a change in content that will continue going forward.) However, with so many items and so much content to cover, it is nearly impossible to build a test that has exactly the same level of difficulty as those previously administered. The overall difficulty of an examination will be slightly different each time. Equating is the process of statistically adjusting scores to account for these differences in difficulty.
"In this article, I will describe the process we use at NCBE to develop the MBE so that it can be equated, as well as the process we use in scaling and equating the examination." (29)