Michael S. Humphreys

The Interactive Effect of Personality, Time of Day, and Caffeine: A Test of the Arousal Model

The personality dimension of introversion/extraversion is one of the few personality dimensions that can be reliably identified from study to study and investigator to investigator. The importance of this demension within personality theory is due both to the stability of the trait and the influential theory of H. J. Eysenck. The basic assumption in Eysencks theory of introversion/extraversion is that the personality differences between introverts and extraverts reflect some basic difference in the resting level of cortical arousal or activation. Assuming that there is a curvilinear relationship (an inverted U) between levels of stress and performance leads to a test of this arousal theory. That is, moderate increases in stress should hinder the performance of introverts who are presumably already highly aroused. However, the same moderate increase in stress might help the performance of the presumably underaroused extraverts. Revelle, Amaral, and Turriff reported that the administration of moderate doses of caffeine hindered the performance of introverts and helped the performance of extraverts on a cognitive task similar to the verbal test of the Graduate Record Examination. Assuming that caffeine increases arousal, this interaction between introversion/extraversion and drug condition supports Eysencks theory. This interaction was explored in a series of experiments designed to replicate, extend, and test the generality of the original finding. The interaction between personality and drug condition was replicated and extended to additional cognitive performance tasks. However, these interactions were affected by time of day and stage of practice, and the subscales of introversion/extraversion, impulsivity, and sociability, were differentially affected. In the morning of the first day, low impulsives were hindered and high impulsives helped by caffeine. This pattern reversed in the evening of the first day, and it reversed again in the evening of Day 2. We concluded that the results from the first day of testing require a revision of Eysencks theory. Instead of a stable difference in arousal between low and high impulsives, it appeared that these groups differed in the phase of their diurnal arousal rhythms. The result is that low impulsives are more aroused in the morning and less aroused in the evening than are the high impulsives. A variety of peripheral or strategic explanations (differences in caffeine consumption, guessing strategies, distraction, etc.) for the observed performance increments and decrements were proposed and tentatively rejected. It seems probable that some fundamental change in the efficiency with which information is processes is responsible for these performance changes.

Item and relational information: A case for context independent retrieval

The probability of recognizing a member of a word pair tested with the pair intact was shown to equal the probability of recognizing a single word plus the probability of recalling an unrecognized word. A reanalysis of earlier results ( Humphreys, Memory and Cognition , 1976 , 4 , 221–232) showed that the probability of failing to recognize both words in an intact study pair was not less than the probability of failing to recognize the two words in a rearranged pair. A second experiment confirmed this finding. These results support the hypothesis that two types of information (item and relational) underlie these recognition judgments. They also support assumptions about the retrieval processes being independent of context.

Relational information and the context effect in recognition memory

A role for relational information was examined for the paradigm in which recognition-memory performance on items tested in the same context in which they were studied is compared with performance on items tested in different contexts. Over a series of five experiments, randomly formed pairs were used to manipulate the context of high-frequency English words. Comparisons were made between instructional manipulations designed to influence the use of relational information, and between yes/no, confidence rating (both between- and within-subject), and forced-choice tasks. There was a context effect not due to the use of inappropriate response strategies. However, high-criterion subjects resembled those subjects who were specifically instructed to use relational information, while low-criterion subjects showed little or no context effect. A model specifying the relationship between item and relational information and how relational information influences decisions in recognition-memory paradigms was proposed.

Phonemic-similarity effects in good vs. poor readers.

Experiments 1–4 examined immediate serial recall of rhyming and nonrhyming items by normal and poor readers in Grades 2–4. Children with generally low achievement were excluded from the poor-reader groups, so that the achievement deficit of the poor readers was centered in reading. The poor readers did not differ from the normal readers in their susceptibility to phonemic similarity either with letter lists or with word lists. Children low in both achievement and intelligence were included in Experiment 3, and they also showed normal susceptibility to phonemic similarity, except that phonemic-confusion effects were reduced when task-difficulty levels were high. Experiment 5 further demonstrated that the serial-recall task is relatively insensitive to phonemic-similarity effects when difficulty levels are high. Previous results suggesting that poor readers are relatively insensitive to phonemic similarity in such tasks may have been an artifactual consequence of marked differences in overall task difficulty for the groups compared. Implications of variations in sample-selection procedures also are discussed.

Sequential testing effects and the relationship between recognition and recognition failure

Tulving and Wiseman (1975) reported that there was a systematic relationship between the proportion of words recognized and the proportion of recallable words recognized. This relationship indicates a moderate positive covariation between recognition and recall across subject items, when each subject is given both types of test, recognition followed by recall. In this paper it is shown that the theoretical enterprise of trying to account for this relationship is fruitless unless the data are corrected for sequential testing effects. Evidence on the existence of these effects is reviewed, and then it is shown how they introduce a measure of dependency between recognition and recall. When the data are corrected, the theories proposed by Begg (1979) and Flexser and Tulving (1978) are shown to be poorly supported. The utility of the general enterprise of determining the relationship between recognition and recall by these means is also questioned.

Interpretation of the two-stage analysis of paired-associate memorizing

Abstract Four groups were run with response difficulty and stimulus difficulty varied factorially. A two-stage Markov model fit the data adequately. The parameter associated with the first stage depended on stimulus difficulty as well as response difficulty, refuting an interpretation of the first stage as response learning. The learning parameters associated with the second stage seemed to depend only on stimulus difficulty. The results suggest that the first stage of learning involves storage of the stimulus-response pair in memory, and the second stage involves learning to retrieve the item reliably.

Individual differences in diurnal rhythms and multiple activation states: A reply to M. W. Eysenck and Folkard

Three issues raised by M. W. Eysenck and Folkard are discussed. These include (a) just what individual difference variable is mediating the time of day and caffeine effects; (b) what the difference is in the diurnal rhythms of low and high impulsives; and (c) whether it is necessary to postulate multiple activation states. Suggestions for future research are then given.

Impulsivity, neuroticism, and caffeine: Do they have additive effects on arousal?

Abstract The personality dimensions of impulsivity and neuroticism have been linked to differences in basal arousal. The hypothesis tested was whether these personality variables and caffeine have additive effects on arousal. All subjects received three paired-associate trials on each of two neutral control lists, two lists with semantically similar stimuli, and two lists with acoustically similar stimuli. Half of the subjects received caffeine and half placebo. Although significant interactions with personality and drug condition were obtained, the ordering of the conditions was inconsistent with the assumption that the arousal performance curve is single peaked. Further, S. Schwartzs ( Journal of Research in Personality , 1975 , 9 , 217–225) hypothesis that in paired-associate learning high arousal and low arousal subjects process semantic and physical information differentially was not supported. The general issue of how to determine whether individual differences in performance are caused by differences in arousal is discussed.

Children's cued recall: Developmental differences in retrieval operations

Abstract Second and fifth grade children were aurally presented a study list containing eight target items plus primacy and recency buffers, followed first by free recall, then by cued recall for items not produced in free recall. Two additional buffer items were used at Grade 5 in an effort to equate the two age groups in free recall of targets. The cues were either of high or of moderate associative strength, and equal for the two age groups. Free recall of targets did not differ with age, suggesting equivalent storage of targets. At Grade 2 the cued recall level matched expectations based on a simple associative account of cuing effects. However, cued recall at Grade 5 was much higher than the normative association values, implicating strategic retrieval operations, the possible nature of which was discussed.

The Derivation of Endpoint and Distance Effects in Linear Orderings from Frequency Information.

Both endpoint and distance effects have been observed in the protocols of subjects learning linear orderings. It is suggested that these could be produced by subjects learning the frequency with which words occur as the greater member of a pair or relationship. In two experiments when subjects learned verbal discrimination lists constructed from a five term linear ordering, both effects were obtained. The pattern of errors appeared to be essentially the same on Trial 1 as it was on all trials, and did not differ as a function of whether or not the subjects were informed about the linear ordering. Furthermore, the end of the order, which was most difficult, varied in accordance with frequency predictions. A modification of the original frequency analysis was made to include the learning of the frequency with which words occur as the lesser member of a pair or relationship.