Multicategorical spline model for item response theory

Additional information contained in incorrect responses calls for a multicategorical rather than a binary analysis of multiple choice data. A nonparametric divided-by-total model for joint maximum likelihood estimation of probability-of-choice functions (for particular responses) and of latent ability is proposed. The model approximates probability functions by rational splines. Some illustrative examples of real test data analysis and the results of a Monte Carlo study are presented.The research in this paper was supported by the National Sciences and Engineering Research Council of Canada Grants OGP0105521 and APA 320 awarded to the first and the second author, respectively. The authors are indebted to R. Melzack and A. Baker for making available the data analyzed in this paper. We would also like to thank J. McKenna and B. Cont for their assistance in editing this paper.     

中小学生语文阅读能力结构的发展特点

本研究对小学六年级、初中三年级、高中三年级三个年龄切面的学生的阅读能力结构构成进行分析,揭示该能力结构的发统特点。结果表明,语文阅读能力结构随年龄的增长而发生变化。一方面表现为因素的数量逐步增加,另一方面表现为比较复杂、层次较高的因素在整个结构中的相对地位不断增强。研究结果进一步指出,该能力结构因素增加的实现方式不同于西方心理学界所认为的分解模式,而是表现为“接替”模式,亦即通过“基础能力因素”对活动的影响作用为新的因素所接替而实现。     

Deficits in set-shifting ability in nonclinical compulsive checkers

Recent studies suggest deficits in set-shifting ability in patients with obsessive-compulsive disorder (OCD) as measured by tasks such as the Wisconsin Card Sorting Test (WCST; Berg, 1948). The present study examined the extent to which these deficits were demonstrated by nonclinical subjects selected on the basis of their checking scores on the Maudsley Obsessional-Compulsive Inventory. A multivariate analysis revealed that frequent checkers performed significantly worse on the WCST than noncheckers, making more total errors and more perseverative errors and requiring more time to complete the test. A follow-up analysis on a subset of the original sample found similar performance deficits in checkers. While the poorer performance of checkers could be statistically accounted for by anxiety during the first administration of the WCST, anxiety could not explain the relationship between checking status and WCST performance at follow-up. Therefore, there does appear to be some relationship between checking status and WCST performance beyond what can be explained by affective variables.     

Concurrent activities and instructed human fixed-interval performance.

Two experiments explored the effects of two types of concurrent activity on human fixed-interval performance. Eight adult subjects were given access to either reading material or a working television set across three fixed-interval values (60 s, 300 s, and 600 s). During Experiment 1, 2 subjects produced "scalloped" patterns and reported no verbal regulation (e.g., counting) in the presence of the reading material, but shifted to low-rate patterns and reported verbal regulation when the reading material was withdrawn. The 2 other subjects in Experiment 1 produced consistent low-rate performances and reported verbal regulation during access to reading material. However, when these subjects were given access to a working television set, they produced scalloped patterns and reported no verbal regulation. During Experiment 2, 4 experimentally naive subjects showed consistent scalloped patterning and no verbal regulation across fixed-interval values when they were allowed to watch television. When access to the television was denied, subjects reliably reported verbal regulation, and low-rate patterns emerged. These behavioral effects focus our attention on the contingencies that control human performance on fixed-interval schedules.     

聋童与听力正常儿童分类能力的比较研究

本研究通过不同性质的分类实验,比较了聋童和听力正常儿童在年龄和教育程度相同或相近的前提下表现出的分类能力的差异,研究结果表明:(1)听力正常儿童的分类水平高于聋童。(2)教育程度的提高对二者的分类水平均有明显的促进作用。(3)就本实验结果看,约在13、14岁间,二者的分类水平均趋于平稳。(4)具体语词的掌握对二者的分类水平均有明显的促进作用。(5)二者对具体材料的分类均优于对抽象材料的分类。     

Effects of day care and maternal teaching on child educability

Two groups of matched mother-child pairs, a day care sample (N = 38) and a home reared sample (N = 38) participated in a study testing mothers for degree of decentred teaching and children for degree of decentred educability. The children were seven years old. Mother and child were tested in separate settings, thus preventing interdependency among scores. The study used a matched-pair design with control of five background variables; age of the child, mother's educational level, mother's occupational status, family SES level, and type of family. The results showed strong positive effects of both day care experience and maternal teaching on child educability. Day care experience improved the educability of children independent of the mother's teaching strategy and gender of the child.     

Temporal differentiation of response duration in children of different ages: developmental changes in relations between verbal and nonverbal behavior.

Children aged 4.5, 7, or 11 years received an experimental session in which a contingency was placed on button-press duration. Each discrete trial was followed by a brief verbal probe asking a question about the contingency requirement. Other groups of children received an identical task followed by a postexperimental interview. Level of adaptation to the duration contingency tended to increase with age in subjects receiving posttrial verbal probes, but not for those who were interviewed. Eleven-year-olds in the verbal probe condition showed a strong correlation between accurate temporal differentiation and number of verbalizations relating to response duration or timing. The younger subjects, with one exception, showed no association between timing-related verbalizations (which were almost totally absent) and response duration differentiation. This developmental difference occurred even though the younger subjects verbalized after almost every trial. The results suggest that although 11-year-old children apparently produce rule-governed behavior under verbal control as adults do, the behavior of younger children may be controlled directly by reinforcement contingencies even when their verbal repertoires are highly developed.     

Acquisition and maintenance of trusting behavior

This study determined whether a two-person exchange situation contained natural contingencies for trusting behavior or whether external contingencies were necessary. Pairs of college students worked matching-to-sample problems for money. On each trial there was one problem and the subjects determined which of them would solve it. Trusting behavior was defined as an increase in the number of consecutive problems each subject allowed his partner to work during sessions that also ended with an equitable distribution. Simply, trust was a temporary deviation from equity. A subject could give the problem to the other person (cooperate), or not respond and let the other person take the problem (share). Other possibilities were for both subjects to try to take the problem (complete), or for neither subject to respond and thereby let the person who worked the last problem also work the next one (passive trust). When only four lever pulls were required to distribute a problem (no external contingencies to reach either equity or trust) subjects reached equity, but only minimal trust (strict alternation of single problems) developed in 18 sessions. When 30 or 60 lever pulls were required to distribute a problem (smaller response requirement for passive trust and therefore a contingency for trust), trusting behavior developed after a few sessions (fixed ratio 30) or after several trials of the first session (fixed ratio 60) and it ordinarily expanded gradually to 10 to 15 consecutive problems through passive trust. The aversiveness of the inequity involved in trusting appears to necessitate a contingency for acquisition. Once trust develops, however, this aversiveness is reduced as subjects learn the inequity is only temporary (e.g., once trust was acquired at fixed ratio 60 it was maintained at fixed ratio 4, which would not initially produce it), and the direction of the inequity appears to become of questionable importance (e.g., being behind was alternated over rather than within sessions and usually not in a systematic manner).     

A COMPARISON OF THREE STRATEGIES FOR TEACHING OBJECT NAMES

Researchers in applied behavior analysis have been charged to provide large-scale demonstrations of the outcomes of evaluations. In this research, three experiments were conducted to examine the relative efficacy of three methods of presenting stimuli in object naming tasks. Stimuli were introduced successively, simultaneously, or using a combination of the two procedures. College adults, mentally retarded children and adolescents, and preschool children were taught to produce the names of five Hebrew letters, English words, or American coins, respectively. Presentation method was a between-subjects treatment in a factorial design. Results from the series of systematic replications were consistent in showing better posttest performance for subjects in the Simultaneous and Combined conditions. Further, follow-up data in Experiment III showed that retention was also superior for subjects trained by the Simultaneous or Combined methods. Although the acquisition criterion was met in fewer trials by subjects in the Successive condition, only several minutes more training time was required by the Simultaneous and Combined conditions. From a cost-effectiveness point of view, either of the latter two techniques should be favored over the Successive procedure for teaching verbal naming skills.