Long G M, Tuck J P
Department of Psychology, Villanova University, Pennsylvania.
Am J Optom Physiol Opt. 1988 Jan;65(1):37-48. doi: 10.1097/00006324-198801000-00007.
In four studies, the comparability of several measures of contrast sensitivity functions (CSF's) as well as their individual test-retest reliabilities were determined with observers drawn from a college population. The relation between CSF scores obtained by different tests or by different psychophysical procedures with the same test was consistently low and frequently nonexistent. These low intertest reliabilities were found to be due to the lack of reliability of some of the individual tests. In particular, with both stationary and flickering gratings, the von Bekesy tracking procedure produced unacceptable variations in contrast estimates. Fortunately, the reliabilities of other procedures, particularly the simple adjustment method, were much higher (greater than 0.70). CSF scores obtained from the wallchart method exhibited insufficient variability among observers in the college sample to permit similar reliability estimates to be determined. Implications of the test characteristics are discussed for users of the CSF.
在四项研究中，从大学生群体中选取观察者，确定了几种对比敏感度函数（CSF）测量方法的可比性及其各自的重测信度。通过不同测试或同一测试的不同心理物理学程序获得的CSF分数之间的关系一直很低，且常常不存在。发现这些低的测试间信度是由于一些个别测试缺乏信度。特别是，对于静态和闪烁光栅，冯·贝凯西跟踪程序在对比度估计中产生了不可接受的变化。幸运的是，其他程序，尤其是简单调整法的信度要高得多（大于0.70）。从壁图法获得的CSF分数在大学样本的观察者中表现出的变异性不足，无法确定类似的信度估计。讨论了测试特征对CSF使用者的影响。
