The classical Psychophysical Methods

由於我的研究有用到Just Noticeable Difference (JND)找出某個力回饋裝置能提供多少層次的力回饋組合。然而傳統的JND實驗是請受試回答兩次刺激之中何者的感受較大，但我的實驗設計是兩者是相同(same)還是不同(difference)，為了深入了解JND的原始做法，在此紀錄1997這本心理物理學的書籍片段。

This reference from PSYCHOPHYSICS THE FUNDAMENTALS 3th, 1997

Difference Thresholds

The method of constant stimuli can also be used to measure difference thresholds. The observer's task is to examine pairs of stimuli and to judge which of the pair is given a fixed value and its called the standard stimulus (St). Th value of the other stimulus, called the comparison stimulus (Co), is changed from trial to trail, being sometimes greater than, sometimes less than, and sometimes equal to the value of the standard stimulus. Usually five, seven, or nine values of the comparison stimulus, separated by equal distances on the physical scale, are employed. The values of the comparison stimuli are chosen so that the stimulus of greatest magnitude is almost always judged less than the standard. There are usually an equal number of comparison stimulus values above and below the value of the standard stimulus. In a random sequence, each of the comparison stimuli is paired several times with the standard stimulus, and the observer reports which stimulus has the grater sensory value.

Time Error

A time error may also confound experimental results when the standard and comparison stimuli are presented successively. In one form of the time error, the proportion of times for which the comparison stimulus is judged grater than the standard stimulus is found to be higher when the comparison stimulus is presented second than when it is presented first (Hellstrom, 1985). Successive presentation makes it necessary for the observer to compare the second stimulus with the memory image of the first. In one interpretation of the time error, it is assumed that since the memory image may rapidly fade, the first stimulus may be judged less than the second stimulus, even when the physical intensities of the two stimuli are identical. Again, since the aim of the discrimination experiment is to study the ability of observers to detect differences in stimuli, certain precautions must be taken to eliminate the biasing effects of time errors. The most common procedure is to present the standard stimulus first on half of the trails and second on the other half of the trails. The method of counter-balancing spatial location or temporal order of standard and comparison stimuli is based on the assumption that, when the results from all of the trails are combined, the effects of the space or time errors will cancel, providing an unbiased estimate of the DL.

Difference threshold

Because the point of subjective equality (PSE) represents a complete lack of discrimination, and because 0 or 1.0 greater than responding is perfect discrimination, the intermediate proportion points of .25 and .75 have been used to find the DL. It is possible to determine two DL's, an upper and a lower. The two are often averaged to give on DL for a particular standard stimulus.