AP Latin Multiple-Choice Item Writing Tutorial
Why Multiple-Choice Questions?
Knowledge of Latin
Colleges that award credit for the AP Latin Vergil and Latin Literature courses require a measure that shows that
students not only can write literary essays on the works they have read in class, but also have a knowledge of
Latin grammar and syntax that will allow them to take advanced courses. The multiple-choice section of the
examination provides this information. Two kinds of passages appear on this section of the exam: sight passages and
syllabus-based passages. The sight passages provide a measure for the colleges of the student's overall ability to
comprehend Latin, while the syllabus-based passages test mastery of a part of the AP Latin syllabus.
Reliability
Many more multiple-choice questions than free-response, or essay questions can be asked in a given time period. The
number of questions is an important factor when considering not only the efficient use of testing time but also the
reliability of a test, or the likelihood that students would obtain the same score on a different
version of the test. A test is designed to measure the ability level of the students, not the difficulty level of the
test itself. To give a true indication of a student's ability, different versions of a test should be of equal
difficulty. Reliability increases along with the number of questions, so a one-hour multiple-choice test with 50
questions is more reliable than a one-hour free-response test with four essay questions.
Fair Scores
The AP Program strives to award fair scores independent of the difficulty level of a particular
year's test. There will, of course, be variations in the difficulty level of individual AP test forms; the use of
multiple-choice questions allows the program to account for these variations when awarding scores. Each year's
multiple-choice test contains a set of questions that appeared on a previous test. The use of these repeated sets of
questions makes it possible to determine the relative ability level of the students taking a particular test form, and
to tie scores to the ability level of the student and not to the difficulty level of the particular test.
Elimination of Bias
In addition, the program tries to ensure that different performance on individual questions is based on differences
in ability, not gender or ethnic differences. Studies are conducted that identify questions where performance of
gender/ethnic groups differs when the groups are matched for ability. These studies are conducted on the
multiple-choice items that make up Section I of the tests. The questions are then examined carefully by a panel of
experts, in order to eliminate bias from the test.
previous | contents | next