Human and computational question answering

The ability to answer questions about a text is the strongest possible demonstration of text comprehension. The question-answering problems that arise in a story-understanding system are discussed. Some solutions are described and illustrated by a computer program, which reads stories, and answers questions.     

Local and Global Sources of Control in Pigeon Delayed Matching-to-sample Performance

The present study employed a behavioural detection approach to investigate the combined effects of sample duration and sample presentation frequency on delayed matching accuracy of pigeons. Experiment 1 showed that when both samples were exposed at each of the two possible durations within a two-alternative, delayed matching session, discriminability was higher to the longer duration sample than to the short-duration sample, as found when sample duration is varied between sessions. Experiment 2's asymmetrical procedure increased bias toward the more frequent of the two samples but had no influence on discriminability. Initial discriminability was higher to samples exposed for longer durations, irrespective of stimulus presentation frequency. The results suggest qualitatively different effects of the two sources of stimulus control under consideration: Sample duration (a local or within-trial manipulation) exerted its effect on discrimination of the stimuli, whereas sample presentation frequence (a global factor) exerted its major effect on response bias. An interpretation of the data in terms of Blough's (1996) analysis of errors in matching tasks suggests that the amount of behaviour under control of the sample stimulus markedly changed with different sample durations. The analysis also showed that the biasing manipulation exerted most of its effect on the portion of behaviour outside of control by the critical stimulus. We argue that theoretical accounts of delayed matching performance need to consider both local and global factors as determinants of matching accuracy.     

COGNITIVE STYLE AND ATTITUDINAL CORRELATES OF THE PERCEIVED CAUSES OF RAPE SCALE

This study examines the relations between beliefs about the causes of rape and attitudinal and cognitive style (the tendency to think about social problems systemically, the view of people as complex and changeable, and an intellectual personality) measures in a sample of 270 community-college students. The Perceived Causes of Rape (PCR) Scale included the following subscales: Male Dominance, Society and Socialization, Female Precipitation, Male Sexuality, and Male Hostility. Beliefs about the causes of rape varied on three dimensions: individual versus sociocultural causes of rape, those causes that focus on the perpetrator versus those that focus on the victim, and rape myths versus feminist beliefs. The causes of rape identified as rape myths were associated with male sexuality stereotypes, a version of Burt's (1980) Rape Myth Acceptance Scale, attitudes toward feminism, and self-identification as a feminist. Agreement with the sociocultural causes of rape was associated with cognitive style measures and age. We suggest that belief in sociocultural causes of rape may require a predisposition to think systemically as much as an ideological stance.     

Marital status and quality in middle-aged women: Associations with levels and trajectories of cardiovascular risk factors.

The current study compared cardiovascular risk profiles and trajectories (i.e., within-person changes) of women who were married or cohabitating and who had high relationship satisfaction with those of women with moderate or low satisfaction and with those of women who were single, divorced, and widowed. Participants were 493 women from the Healthy Women Study, a prospective investigation of health during and after the menopausal transition. Risk factors were measured across more than 5 occasions and 13 years, on average. Data were analyzed using a multilevel modeling technique. Overall, women in relationships with high satisfaction had lower levels of biological, lifestyle, and psychosocial risk factors when compared with the other groups. In some cases, women in satisfying marriages also showed a lower risk trajectory on risk factors relative to other women. Hence, marriage appears to confer health benefits for women, but only when marital satisfaction is high.     

Occupation and subclinical carotid artery disease in women: are clerical workers at greater risk?

The current study examined cardiovascular risk factors and carotid atherosclerosis in 362 women (ages 42-50 years) who were working in clerical, blue-collar, or white-collar jobs or who were not employed. Risk factors were measured premenopausally and ultrasound measures of carotid atherosclerosis were obtained approximately 11 years later. Clerical and blue-collar workers had more atherogenic profiles on physical, behavioral, and psychosocial risk factors when compared with white-collar and nonemployed women. Clerical workers had greater carotid intima-media thickness relative to all other groups and more focal plaque when compared with white-collar workers. Risk factors and workplace characteristics did not account for the greater carotid atherosclerosis observed in clerical workers. Further research is needed to investigate why clerical work may increase cardiovascular risk.