Paul W. Fox

Do Laboratory Findings on Test Expectancy Generalize to Classroom Outcomes

We conducted a meta-analysis of both classroom and laboratory studies of the effects of expecting a recall, recognition, essay, multiple-choice or true-false test on students’ subsequent achievement. In laboratory studies, studying with a recall set produced strong positive effect sizes for both discrete and prose materials. However, studying with a recognition set produced no effects with discrete materials and small negative effects with prose materials. In contrast, results from classroom studies indicated that students achieved most when preparing for the type of test they received. These results run counter to standard wisdom in the college study skills area and lead us to challenge the assumption that laboratory studies on expecting tests of recall and recognition provide a useful analog to test expectancy effects involving essay and multiple-choice tests in the classroom.

Unexceptional spatial memory in an exceptional memorist

Rajan Mahadevan evidences an exceptional memory for arrays of digits. We tested whether Rajans spatial memory was likewise exceptional. Eight control Ss and Rajan were instructed to remember the position and orientation of 48 images of common objects shown either to the left or the right of fixation and facing either left or right. Rajans accuracy for judging whether the position and orientation of these pictures had changed when they were shown in a different sequence was lower than that of control Ss for both judgments. Rajans exceptional memory capacity apparently does not extend to spatial relations.

Patterns of stability and change in behaviors of free association

Several basic questions are raised concerning data generated by tests of free association: What is the upper limit of normative correspondence across associative collections? What is the temporal reliability of the individual S s free association behavior? Do cultural norms predict temporal changes in association by the individual S ? Associations collected from the same S s on two occasions 60 days apart revealed great normative stability. However, examination of each S s associates showed that over 50% of the responses differed from one occasion to the next, even though the group data comprising the cultural “norm” were nearly identical. Patterns of temporal change in association by the individual S were lawfully related to the strength of the cultural primary response.

Age Differences in Hypermnesia: Word Gain Versus Word Loss

A hypermnesic task was administered to 82 younger adults (ages 27-39), 63 middle-aged adults (ages 40-59), and 119 older adults (ages 60-87). Previous research suggests that relational encoding prevents loss of items and item-specific encoding promotes item gains in a hypermnesic task (Klein et al., 1989) and that there are age differences in relational but not item-specific encoding (Luszcz et al., 1990). This information provided the basis for three predictions: (a) There are age differences in hypermnesia, (b) there are age differences in word losses in a hypermnesic task, and (c) there are no age differences in word gains in a hypermnesic task. In order to manipulate type of encoding, a list of words with high association strength (to evoke relational encoding) and words with low association strength (to evoke item-specific encoding) was constructed. The results of this investigation provide support for the encoding manipulation and for all three predictions. In addition, the nature of the age differences in word loss observed suggests that although older adults may be capable of relational encoding, this form of encoding is not as effective at preventing word loss for them as it is for younger adults.

Recall and misrecall as a function of cultural and individual word association habits and regulation of the recall environment

Data are presented regarding (a) experimentally induced changes in word-association strength; (b) recall as a function of cultural associative probabilities; and (c) dependencies between association and recall responses of individual Ss in particular recall environments. Subjects were trained by a brief exposure to two 5-word lists composed of the second most frequently occurring cultural responses (R2s) to Kent-Rosanoff stimuli. During the final 60 sec of the 3-min retention interval, experimental Ss free-associated to stimuli related to the five R2 words of List 2. Control Ss either free-associated to stimuli unrelated to the words of training or continued a vowel-cancelling activity. All Ss were then tested for recall of List 2 in the presence or absence of the corresponding stimulus terms. Major findings include: (a) substantial and systematic changes in R2 associative strength following item pre-exposure; (b) variation of recall and misrecall according to trends in normative association data; and (c) recall interference or facilitation dependent upon the individual Ss associative hierarchy and the stimulation conditions regulating the recall environment.

The NRC booklet: A group technique for studying word arousal phenomena as a function of Ss

A booklet has been designed which can be used with large, intact classes of Ss yet it permits the programming of idiosyncratic training material in experiments on learning and retention.

Free association, free recall, and stimulated recall compared

Three primary methods of recall (free association, free recall, stimulated recall) and two modifications of them (modified free association, modified stimulated recall) were defined by use of three variables: (a) presence or absence of prior laboratory training, (b) presence or absence of E controlled test stimulation, and (c) test instructions to free associate or to recall Data obtained from approximately 300 Ss and an earlier set of 600 revealed the differential effectiveness of all three primary methods in recall and misrecall.