Consistency of individual exponents in cross-modal matching

Abstract

An important question about individual differences in the exponent of the psychophysical power law is how they should be interpreted. The differences may reflect permanent characteristics of individuals, and it has been argued that, if this is so, the range of these differences is so great as to identify the class of data as exceptional among the physical and biological sciences. Cited as evidence of such permanence has been the correlation between individual exponents obtained on two separate occasions. In a previous paper, we showed that increasing the time interval between occasions reduced the correlation to a nonsignificant level; we argued, therefore, that obtained individual differences in exponents, even though large, depended upon the operation of factors only incidentally associated with the particular observer. In a series of new studies of session-to-session correlation between individual exponents, we provide evidence that: (1) our original finding for magnitude estimates of visual size is repeatable, with the correlation dropping to nearly zero after 1 week; (2) when judged line length is matched to brightness, a delay of 1 week is sufficient to produce a nonsignificant correlation; (3) in contrast, magnitude estimates of loudness yield significant correlations after a week's delay; (4) but, when moduli are arbitrarily changed between sessions by the experimenter, these correlations for magnitude estimates of loudness drop to a nonsignificant level, even for a zero-delay condition. We conclude that, whereas in some scaling tasks the passage of time alone between sessions is sufficient to disrupt what appears to be the mnemonic basis for session-to-session correlation, in other (less familiar) tasks, more positive interference (in the form of a modulus change) is needed to achieve the same end. The evidence is consistent with the belief that enduring characteristics of the observer contribute only a small portion of the variability in individual power law exponents.

Effects of single-session repetitive judgments on magnitude estimation scales for lingual vibrotactile sensation

“…Eisler (1982) suggests that a critical element in both magnitude estimation and magnitude production is a paradigm that permits openended responses with minimal constraints imposed upon the subjects by the experimenter. If memory and learning is encouraged through repetitive response judgments, then subject spontaneity is eliminated (Teghtsoonian & Teghtsoonian, 1983). Warren (1981) indicates that one way of minimizing contextual effects when performing psychophysical intensity judgment experiments is to use only first judgments from large groups of subjects.…”

Magnitude estimation of conceptual data dimensions for use in sonification.

“…In other words, the relation is linear. Hence, from the subject's point of view (or, as we say, subjectively), number and line length are directly proportional" (Stevens, 1975, p.109 (Teghtsoonian & Teghtsoonian, 1983, Experiments 1 and 2). However, as I described in the previous section, the magnitude estimation procedure tends to slightly underestimate the "actual" value for the exponent, which may be why Stevens (1975) asserted a linear relationship between perceived and actual line length (i.e., a "real" exponent value of 1.0).…”

Consistency of magnitude estimations with conceptual data dimensions used for sonification

“…At the level of the individual, for example, M. and R. Teghtsoonian (1971Teghtsoonian ( , 1983) have looked at how stable magnitude estimation slopes are for a single participant, across multiple replications, with different inter-session intervals. Interestingly, they have found only weak intra-individual stability from session to session unless there is essentially no break between the sessions (see M. Teghtsoonian & Teghtsoonian, 1983 for a review). This is in contrast to the situation for magnitude estimation slopes determined from the data of groups of participants, where intergroup stability does seem to be evident.…”