EPPP Pass Point
Establishing a pass point for a high stakes licensure exam such as the EPPP is essentially a matter of professional judgment. Procedures for establishing pass points depend on experts making judgments about the level of expertise necessary for minimally competent and ethical practice of the profession.
Historically the EPPP pass point was established by expert opinion of individual boards who used the exam. When ASPPB first adopted a recommended pass point of 70%, it chose this cut-off because it was the modal choice of jurisdictions. In fact, some jurisdictions have engaged in formal pass point setting exercises (typically variations of the Angoff procedure) and these have confirmed the 70% pass point as reasonable.
Pass points chosen by jurisdictions for doctoral level independent practice have generally been better agreed upon than pass points for masters level supervised practice. We are not aware of any Angoff procedures being done for supervised practice. Modal choices for jurisdictions with supervised practice provisions were either 60 or 65 percent.
One of the problems with setting an absolute score (e.g., 140) or percentage score (e.g., 70%) pass point is that exam forms vary slightly in difficulty. Candidates with equal ability might pass one exam form and fail another based on the characteristics of the exam. A consistent standard of knowledge is fairer to candidates and preferable to an absolute standard. This problem is diminished by statistically equating different forms of the exams . Equating tables were available to jurisdictions for their use with the paper-and-pencil exam. The computerized exam is equated at the independent practice pass point.
With the goal of acceptance of common standards based on equated pass points, when ASPPB moved to computerized delivery, it also went to scaled scores. Scaled scores allow the reporting of comparable scores for different exam forms. The computerized exam is pre-equated at the independent practice pass point. This is accomplished through the use of statistical techniques based on item response theory using items that have a history of use and established item statistics. The scaled score pass point was set at an arbitrary number of 500. The 500 score is equivalent to a 140 or 70% on the equating base form of the exam. The supervised practice pass point of 65% would be 450 and a 60% pass point would be 400 on the same scale. Neither of these is directly equated for difficulty level. In fact, not only do exams differ in difficulty level at the independent practice pass point, but they may differ at the supervised practice pass point in ways that are not exactly equivalent to the independent level. That is, exams vary in the distribution of scores so the difference between a scaled 140 and a scaled 130 may reflect different levels of difficulty on different exams. For example, a score of 450 might reflect the 30th percentile of the distribution of scores on one test and the 32nd on another.
Exam statistics can inform pass point setting and can describe the effects of setting different pass points. All equating is done using statistics for first time examinees taking the exam at the doctoral level. To include repeat examinees would be to use a biased sample with a lowered mean and an increased standard deviation. Also it should be noted that we have several categories of master’s level candidates. Statistics for that group are complicated by the fact that "All Master’s Candidates” include candidates who are actually finishing their doctorates but are able to sit for the exam based on their masters.
In setting a recommended pass point for supervised practice ASPPB considered the equivalents of either 60% or 65%, i.e., 400 or 450 on the new scale. A review of the scores on recent exams for first time doctoral and first time master’s candidates was conducted. On both paper-and-pencil and CBT administrations, means for doctoral and masters candidates were between one half and one full standard deviation apart. One half deviation below the scaled score of 500 corresponds to about 450 and one full standard deviation to about 400. ASPPB chose the more stringent score to recommend as the pass point. Jurisdictions may set their own pass points, but should be sure to articulate their rationale, especially if they do not use the ASPPB recommended pass points.
A summary table of passing scores by jurisdiction and license type is included in ASPPB's online Handbook on Licensing and Certification Requirements.