STIMULUS-INSTRUCTION SITUATION
In fact, this varies along a number of parameters:
(1) To react or not react. Generally, reaction to the stimulus is required. This is not always so, as for
example in tests of pain sensitivity.
(2) Restricted versus unrestricted variety of response. Completely unrestricted response is unknown
in testing. It is assumed that subjects (as Cattell and Warburton argue) will not eat the test materials
or destroy them. However, responses can be relatively free, as in the Rorschach, or restricted, as in
reaction-time measures.
(3) Inventive versus selective responses. The meaning is obvious here. The true-false response mode
typifies the latter, the description of a Thematic Apperception Test (TAT) picture the former.
(4) Single versus repetitive responses.
(5) Sequence of responses: ordered versus unordered. This is concerned with the extent to which
subjects are required to give responses in a given order.
(6) Homogeneous versus patterned responses. Some tests require responses of the same kind;
others necessitate a variety of different responses.
(7) Natural versus limited responses. This refers to the instructions, such as 'Work at your own
speed', compared with 'Do not spend long over each question' and 'Work as fast as you can'.
(8) Concluding reaction versus reaction to reaction. Either the subject reacts to the test material or
he reacts to his own reaction, for example he may have to evaluate it or associate to it or remember
it. Cattell and Warburton (1967) consider these to be the main parameters of variation in the
construction of test materials, concerning which there is little cause for argument. However, three
further, more subjective dimensions are presented which, for the construction of objective tests,
would appear to be particularly important.
(9) Immediate meaning versus referent meaning. In some tests there is no meaning beyond the test,
as in reaction times. However, in certain tests (e.g. where opinions are asked) there is considerable
symbolic reference. This is a critical parameter since most forms of distortion and problems in the
subjective meanings of words and concepts occur with tests which have referent meaning.
(10) Itemized versus global presentation. This parameter is concerned with the distinction between
tests consisting of items and those which consist of a single task.
(11) Nature of the psychological decision required in the task. This concerns the question of whether
the response demands (a) cognition (e.g. correct judgement), (b) judgement of feeling or (c)
judgement of familiarity or recognition. Two other possibilities might be, Cattell and Warburton
(1967) argue:
(12) Variation in the motives for accepting the test situation
. (13) Variation in freedom to leave the test situation
Response-scoring parameters
(1) Objective versus self-evaluative. The point here concerns mainly those tests using self-report
items. Is the subject scored as he would expect from the instructions or on something unknown to
him? This does not involve deception of subjects. For example, an objective test is the Critical
Evaluations Test (T8, Cattell and Warburton, 1967). Here subjects are asked to state whether a given
performance - for example, a waitress taking ten minutes to bring six meals to a table - is very good,
good, poor or very poor. What inter alia is scored here is the number of critical evaluations. The
subject matter of the items is irrelevant to the scorers. Thus this inventory-like test is, in fact, entirely
objective, as defined at the beginning of this chapter. Of course, all the tests to be discussed in this
chapter are objective in terms of this first parameter.
(2) Overt behaviour (total organism) versus physiological response (part organism). Typical
physiological responses might be blushing or tremor.
(3) Dimension of one possible response versus classification among a variety of responses
(parametric versus non-parametric). The parametric measures a dimension of the response: time,
errors, repetitions. The non-parametric reveals the number and variety of classes of response. In this
sense many creativity tests are scored non-parametrically.
(4) Total quantity or number of responses versus fraction meeting a criterion. This cuts across the
third category above, since, as Cattell and Warburton point out, a variety score could be of either
category here.
(5) Single homogeneous score verses patterned relational score. The single score is possible only if
the test is scored as a whole, as is the case with personality inventories. A patterned relational score
can take many forms: difference in time for completion of first and second parts; memory for
material under standard and under distracting conditions. Finally, Cattell and Warburton actually add
a sixth category:
(6) Normative versus ipsative scoring. However, this seems to be a classification of a radically
different type from those above, since it affects all tests and is more properly relegated to the sphere
of standardization than test construction. If we were to include this category, there would be a
possible 26 types of response score
