Scaled-scoring is a method of score reporting that standardizes reported scores across exams, differing exam forms, and exam versions.
Instead of reporting exam results as a percentage of total items answered correctly and having different required passing percentages for each exam, all F5 exams are scored on a scaled-score basis, where your score will range from a possible 100-350 points; all F5 exams are calibrated for a passing score of 245 on that scale.
A scaled cut score remains stable across all forms and versions of the same exam. Exam forms containing different items are all aligned to the same specifications; however, when individual items differ in difficulty, forms may also differ in difficulty. Scaled scores take these differences into account and ensure the reported scores across exam forms and versions have the same meaning regardless of difficulty. Fair and consistent decisions can then be made about exam results regardless of the exam form or version.
Regardless of the number of items on an exam, an individual's raw score (number of correct items answered), or the required cut score percentage, the following conditions will be true:
1. A score of "0" will always be reported as a scaled score of "100";
2. The "cut score", or "passing score" will always be reported as a scaled score of "245"; and,
3. A perfect score will always be reported as a scaled score of "350".
The graphic below illustrates how this would work with two, 100 item exams, one with a 50% required passing percentage (blue) and one at 70% required passing percentage (orange).
In the case of a 50% required passing percentage (blue), a candidate would have to correctly answer 50 of the 100 questions in order to pass. In this case a raw score of 0 is set to a scaled score of 100, a raw score of 50 is set to 245 (the cut-score), and a raw score of 100 is set to 350. Raw scores between 1 and 49 are "scaled" to equally cover the 145 point difference between 100 and 245, which comes out at approximately 3. Each correctly scored item below 50 is calculated by taking the number of correct answers, multiplied by the scaled-score value (~3), and added to 100; a raw score of 20 would be approximately a scaled score of 160 (20 x 3, + 100), whereas a raw score of 40 would be approximately a scaled score of 220 (40 x 3, + 100). Raw scores between 51 and 99 are similarly scaled to equally cover the 105 point difference between 245 and 350, which comes out to approximately 2. Each correctly scored item above 50 is worth about 2 additional scaled score points. In this case, the scaled score is calculated by taking the raw score, subtracting the required passing raw score, multiplying by the scaled-score value (~2) and adding it to 245. A raw score of 60 would be an approximate scaled score of 265 (60-50, x 2, + 245); and, a raw score of 70 would be an approximate scaled score of 285 (70-50, x 2, + 245).
In the case of a 70% required passing percentage (orange), a similar calculation is performed, but this time the raw cut score is at 70, rather than 50. Again, a raw score of 0 is set to 100 and a raw score of 100 is set to 350; however, now a raw score of 70 is set to 245 because you must answer 70 of the 100 questions correctly to pass the exam. Instead of 50 questions scaled across the 145 difference between 100 and 245, we must scale 70 questions across that difference. Each of the first 70 questions are worth about 2 scaled-score points, and each of the 30 remaining questions are worth about 1.5 scaled-score points.
You will note that in the blue example (50% requirement) a raw score of 70 would result in a scaled score of 285, well above the cut-score of 245, but in the orange example (70% requirement), a raw score of 70 would result in a scaled score of 245, right at the cut-score line. Conversely, a raw score of 50 on the blue example results in a scaled-score of 245 (the cut score), while in the orange example, a raw score of 50 would only generate a scaled score of 200, well below the cut score. This is precisely what scaled-scoring is intended to do.