Factors influencing test reliability

Numerous factors are listed, with the evidence for their importance. The following factors in the construction of tests affect the reliability: A test is apt to be more reliable if it has a large number of items, if it takes a long time, if it has a narrow range of difficulty, if its scaling is even, if it has few interdependent items, if it is objective in scoring, if it is scored accurately, if the factor of chance enters little into securing the correct answers, if the positions of the correct items among the others are carefully guarded against constant trend on part of subjects to mark more frequently items in certain positions, if the material is homogeneous, if the material is common to subjects, if it is given late in the school year, if it has little dead material, if it has few catch questions, if it contains subtle factors. The following factors in the variability among the individuals taking the test also affect the reliability: Constancy of the speed of working, variability in accuracy, incentive or effort, the obtrusion of competing ideas, illness, worry, excitement, accidents during examination, and cheating.