PSYCHOLOGICAL ASPECTS OF MAN’S ENVIRONMENT 1029
energy change required to perceive a change in the stimulus.
Because the amount of energy related to that required for a
perceived change is often related to the absolute value of the
stimulus, the different limen is often given as a fraction of
the absolute value.
Since the range of intensity for most of the senses is very
large, it is customary to use a logarithmic scale in expression
of the stimulus values. In audition, the decibel is a common
unit. Intensity ( I ) is given in decibels (dB) byIPP^10 log / dB,^1 rwhere P 1 is the stimulus power, and P r the reference level.
Measurement in decibel units is less common for the other
senses, but some type of logarithmic scale is frequently used.
The measurement of thresholds presents certain techni-
cal problems, due to variability in the response, and the pos-
sibility of response bias. Two major methods were devised to
deal with the problem: the method of limits and the method
of average error. In the method of limits, series of stimuli
of varying intensity are presented and the subject is asked
to make a response, according to whether the subject does
or does not perceive the stimulus, or stimulus change. Since
threshold values obtained typically depend on whether the
approach to the threshold is up (from unperceived to per-
ceived) or down (from perceived to unperceived), several
series of trials in both directions may be averaged. The
method of average error is an adjustment method. The sub-
ject is asked to adjust a variable to match a standard, and the
variability of the response is used as a measure of the limen.
It is most suitable to difference limen measurements.
Recently, the concept of a sensory threshold has been
questioned, and the most popular view at present is to view the
internal process as continuous, with a criterion applied to pro-
duce a dichotomous response. The view derives from the signal
detection theory of sensory processes. The extent to which this
model of the threshold differs from older ones is arguable; the
arbitrary nature of the threshold concept has always been rec-
ognized. However, it has led to a more systematic treatment
of the problems of noise and bias in threshold measurements.
Procedurally it stresses the measurement of thresholds by
yes/no judgments in which the probability of a false alarm is
explicitly estimated, and by forced choice techniques. Receiver
operating characteristic (ROC) curves are used to analyze the
data. The resulting measure of sensitivity is usually referred
to as d. From the point of view of reliability, either the classic
methods or new signal detection procedures will usually lead
to satisfactory results.
The sensory dimensions have been divided into two
types: intensive dimensions and qualitative dimensions.
Examples of intensive dimensions are the intensity of light
and sound. Examples of qualitative dimensions are color
and pitch, which often have recurrences in them, such as the
scale relation in pitch, which complicate analysis. Sensory
intensity normally increases logarithmically with increases
in the physical stimulus. Qualitative dimensions may have
more complex relationships with their physical correlates.The measurement of sensory thresholds may be extended to
animals by means of conditioning procedures to be described
in the next section.
Several thresholds other than the absolute and differen-
tial limens are important for environmental studies. Often,
concern will lie in thresholds for discomfort, and for toler-
ance. It may also be desired to obtain information on the
degree of discomfort produced by different environmental
conditions. Direct judgments of discomfort or tolerance tend
to be highly variable; individual tolerances, or at least state-
ments about these tolerances, differ from individual to indi-
vidual. One technique which is sometimes used is to obtain
comparison judgments about the relative discomfort caused
by different stimuli. One method of measuring pain thresh-
olds is based on this principle; the discomfort caused by var-
ious stimuli is compared with the pain caused by a known
exposure to infrared light on the forehead.
A fixed standard need not be used; cross comparison
of the set of stimuli of interest is possible. Techniques have
been developed for taking such paired judgments, and using
them to scale the stimuli along the dimension of judgment.
Variations of this technique can be applied to ranked data:
ranking is less time consuming than a full set of paired
comparison judgments, but it requires more assumptions to
generate a scale. Multidimensional scaling techniques are
usefully applied to preference or similarity judgments, how-
ever obtained. They are especially useful where the relevant
dimensions of eth stimulus set are not clear, since they can
be analyzed to obtain dimensions of judgment. It is also pos-
sible to analyze such judgments for clusterings among the
subjects.
Classical and Operant Conditioning Classical condi-
tioning is a type of learning pioneered by I. Pavlov, and still
extensively studied by psychologists. In its simplest form,
the procedure consists of presenting some stimulus which
consistently elicits a response, known as the unconditioned
stimulus (UCS) in combination with some other stimulus
which does not normally evoke the response, known as the
conditioned stimulus (CS). After a number of trials, the CS
will be found to elicit the response in the absence of the UCS.
Such a response is known as a conditioned response (CR).
Various complexities can be introduced into this para-
digm and it has been the subject of extensive study and
theoretical discussion. Its primary importance for the envi-
ronmental field lies in the fact that the technique can be used
to determine sensitivity to various stimuli. This can occur in
two ways.1) The effect of various physiologically active stim-
uli on the course of acquisition and extinction of a
known CR can be studied. Changes in the course
of learning and extinction are evidence of the
effect of the stimulus.
2) The procedure can be used to determine threshold
sensitivity to a stimulus. If a CR can be established
to a given concentration or intensity of a stimu-
lus, it is evidence that the organism is sensitive to
the stimulus, though the reverse is not necessarilyC016_012_r03.indd 1029C016_012_r03.indd 1029 11/18/2005 11:03:53 AM11/18/2005 11:03:53 AM