The greatest lower bound to the reliability of a test and the hypothesis of unidimensionality

To assess the reliability of congeneric tests, specifically designed reliability measures have been proposed. This paper emphasizes that such measures rely on a unidimensionality hypothesis, which can neither be confirmed nor rejected when there are only three test parts, and will invariably be rejected when there are more than three test parts. Jackson and Agunwamba's (1977) greatest lower bound to reliability is proposed instead. Although this bound has a reputation for overestimating the population value when the sample size is small, this is no reason to prefer the unidimensionality-based reliability. Firstly, the sampling bias problem of the glb does not play a role when the number of test parts is small, as is often the case with congeneric measures. Secondly, glb and unidimensionality based reliability are often equal when there are three test parts, and when there are more test parts, their numerical values are still very similar. To the extent that the bias problem of the greatest lower bound does play a role, unidimensionality-based reliability is equally affected. Although unidimensionality and reliability are often thought of as unrelated, this paper shows that, from at least two perspectives, they act as antagonistic concepts. A measure, based on the same framework that led to the greatest lower bound, is discussed for assessing how close is a set of variables to unidimensionality. It is the percentage of common variance that can be explained by a single factor. An empirical example is given to demonstrate the main points of the paper. The authors are obliged to Henk Kiers for commenting on a previous version. Gregor Sočan is now at the University of Ljubljana.     

The role of personality factors in predicting the reported physical health symptoms of Turkish college students

This study investigates the role of optimism, health control beliefs, perceived health competence, and medical help-seeking variables in predicting the frequency of reported physical symptoms. A total of 345 college students (207 male and 138 female) were presented with the Life Orientation Test, Multidimensional Health Locus of Control, Perceived Health Competence Scale, and Physical Symptom Checklist. Separate stepwise multiple regression analyses were applied to the data obtained from males, females, and the total group of students. Results showed that optimism, medical help seeking, chance health locus of control, and internal health locus of control predicted the reported physical symptom scores of the total sample. Also revealed were the different patterns for males and females. While optimism and internal health locus of control best predicted physical health for males, optimism and chance health locus of control best predicted the physical health symptoms of females.     

Effects of bilingual and monolingual presentations and time on concept- and language-matching tasks

Balanced Turkish-English bilingual participants viewed word pairs, presented both monolingually (English-English or Turkish-Turkish) or bilingually (English-Turkish or Turkish-English) and both for short and long durations. They made decisions on whether the simultaneously presented words in a pair were in the same language or not, or whether they denoted the same concept or not. In the short presentation condition, we found no evidence for subliminal processing. In cases in which both words were consciously identified, participants were more accurate, although not faster in the long than in the short presentation condition for both language and concept decisions. In the long presentation condition, language decisions were more accurate than concept decisions, although not faster. In addition, language decisions were not affected by whether the words were synonyms (concept identity), and concept decisions were not affected by whether the presentation was monolingual or bilingual (language identity), although in the monolingual conditions, "same" decisions were faster but not more accurate, and in the bilingual conditions a speed-accuracy trade-off was observed in that "same" decisions were faster but "different" decisions were more accurate.     

Perceived self‐society moral discrepancies predict depression but not anxiety

Discrepancies between one's own beliefs, standards and practices and the standards expected by others are associated with increased vulnerability to depression and anxiety. Perhaps the most important personal standard is morality, one's standard of acceptable behaviour. We therefore reason that perceived discrepancies between one's own moral standards and those of society predict anxious and depressed moods. We tested this hypothesis, for the first time, in a sample of 99 female Turkish students. Moral discrepancies were assessed using an adapted moral foundations scale: participants were asked how much payment they would require to perform a series of potentially immoral acts, and how much payment they thought the average person in society would require. Participants also completed standard questionnaire measures of depression and trait anxiety. Results show that perceived self‐society moral discrepancies were significantly related to depression scores, but not to anxiety scores. Furthermore, only discrepancies related to the moral dimensions of respect for ingroups and avoiding harm were related to depression. We argue that perceiving a discrepancy between one's own standards of behaviour and those of society can increase vulnerability to depression, much as other kinds of self‐other discrepancies can; however, the specific moral standards which predict depression may vary with culture and the characteristics of the sample.     

Is retrieval the key? Metamemory judgment and testing as learning strategies

Re-reading is the most common learning strategy, albeit not a very efficient one. Testing is highly efficient, but not perceived by students as a learning strategy. Prospective judgment-of-learning (JOL) reflect the learner's impression of subsequently being able to retrieve the ongoing learning in a cued-recall task. Estimating JOL involves attempting to retrieve the information, as in testing. The few studies that have explored the potential mnemonic benefit of JOL have yielded contradictory results. Our aim was to compare JOL and testing with re-study and to examine the impact of these strategies according to the relative difficulty of the material (cue-target association strength) in two experiments. After a first encoding phase, participants re-studied, provided JOL, or took a test. Forty-eight hours later, they participated in a final cued-recall test, during which their confidence level judgments were collected. The main result was that delayed JOL behaved in the same way as testing, and both yielded better performances than re-study when material was of moderate difficulty. The easy or very difficult material revealed no differences between these strategies. JOL is proposed as an alternative to testing when faced with difficult material.     

Is Event-Based Prospective Memory Resistant to Proactive Interference?

Proactive interference builds up rapidly in recall with the presentation of successive lists comprising items from a single category (e.g., animals). In two experiments, we used a similar build-up paradigm in prospective memory. Interestingly, in Experiment 1, there was no evidence of proactive interference in prospective memory, although build-up emerged in a similar recall task. In Experiment 2, by also showing proactive interference build-up in a recognition task analogous to our prospective memory task, we ruled out the possibility that it was the recognition processes inherent in the prospective memory task that might make the task easier and prevent proactive interference. We suggest that, under normal conditions, prospective memory is resistant to build-up of proactive interference and propose that this resistance is a function of the strengthening afforded by a cue-to-intention association that is a part of prospective memory tasks. We discuss the finding using the classical paired-associate learning framework.     

Age-based discrimination of rival males in western bluebirds

Adaptive social behavior frequently involves discriminating between classes of individuals such as relatives versus non-relatives, older versus younger individuals, or individuals of different status. In the absence of spatial cues, this discrimination may be based on signals that correlate with fitness-related traits (e.g., older or high-status males may sing higher performance songs) or with identity, for example, when receivers distinguish and classify signalers based on their unique signal structure. Here, we examine vocal age-based discrimination in western bluebirds (Sialia mexicana), a North American songbird in which older males have a significant advantage in obtaining extra-pair fertilizations, and therefore pose a significantly higher threat to paternity than younger males. We asked whether western bluebird males showed a higher response to playback of songs of older males compared to younger males relative to their own age. We prepared song stimuli by removing three potential signals of age that have been identified as important in other species: (1) note consistency (which was achieved by playing a single instance of each note repeatedly), (2) note repertoire size, and (3) singing rate (the latter two were equalized across conditions). Even in the absence of these potential signals of age, young males responded more strongly to playback of older males’ songs than to young males’ songs, suggesting that they are able to discriminate between age classes relative to the threat they pose. Further research is required to determine whether this discrimination is based on individual recognition or signal features that are correlated with age.