The many null distributions of person fit indices

This paper deals with the situation of an investigator who has collected the scores ofn persons to a set ofk dichotomous items, and wants to investigate whether the answers of all respondents are compatible with the one parameter logistic test model of Rasch. Contrary to the standard analysis of the Rasch model, where all persons are kept in the analysis and badly fittingitems may be removed, this paper studies the alternative model in which a small minority ofpersons has an answer strategy not described by the Rasch model. Such persons are called anomalous or aberrant. From the response vectors consisting ofk symbols each equal to 0 or 1, it is desired to classify each respondent as either anomalous or as conforming to the model. As this model is probabilistic, such a classification will possibly involve false positives and false negatives. Both for the Rasch model and for other item response models, the literature contains several proposals for a person fit index, which expresses for each individual the plausibility that his/her behavior follows the model. The present paper argues that such indices can only provide a satisfactory solution to the classification problem if their statistical distribution is known under the null hypothesis that all persons answer according to the model. This distribution, however, turns out to be rather different for different values of the person's latent trait value. This value will be called ability parameter, although our results are equally valid for Rasch scales measuring other attributes.As the true ability parameter is unknown, one can only use its estimate in order to obtain an estimated person fit value and an estimated null hypothesis distribution. The paper describes three specifications for the latter: assuming that the true ability equals its estimate, integrating across the ability distribution assumed for the population, and conditioning on the total score, which is in the Rasch model the sufficient statistic for the ability parameter.Classification rules for aberrance will be worked out for each of the three specifications. Depending on test length, item parameters and desired accuracy, they are based on the exact distribution, its Monte Carlo estimate and a new and promising approximation based on the moments of the person fit statistic. Results for the likelihood person fit statistic are given in detail, the methods could also be applied to other fit statistics. A comparison of the three specifications results in the recommendation to condition on the total score, as this avoids some problems of interpretation that affect the other two specifications.The authors express their gratitude to the reviewers and to many colleagues for comments on an earlier version.     

时序信息的加工:自动还是控制

从长时记忆的角度出发,以故事的形式为实验材料,对时序信息的加工方式和通道效应进行了研究。结果表明,时序信息的加工存在视听通道效应,通道效应的机制源于记忆,且与加工方式有关。时序信息三个属性的加工方式不同：顺序属性倾向于自动加工;位置属性,就视觉信息来说倾问于自动加工,听觉在有顺序标码情况下倾向于自动加工、而在无顺序标码情况下则是一控制加工过程;间隔特性是一个控制加工过程。     

Response-independent Events In The Behavior Stream

The metaphor of the behavior stream provides a framework for studying the effects of response-independent food presentations intruded into an environment in which operant responding of pigeons was maintained by variable-interval schedules. In the first two experiments, response rates were reduced when response-independent food was intruded during the variable-interval schedule according to a concomitantly present fixed-time schedule. These reductions were not always an orderly function of the percentage of response-dependent food. Negatively accelerated patterns of key pecking across the fixed-time period occurred in Experiment 1 under the concomitant fixed-time variable-interval schedules. In Experiment 2, positively and negatively accelerated and linear response patterns occurred even though the schedules were similar to those used in Experiment 1. The variable findings in the first two experiments led to three subsequent experiments that were designed to further illuminate the controlling variables of the effects of intruded response-independent events. When the fixed and variable schedules were correlated with distinct operanda by employing a concurrent fixed-interval variable-interval schedule (Experiment 3) or with distinct discriminative stimuli (Experiments 4 and 5), negatively accelerated response patterns were obtained. Even in these latter cases, however, the response patterns were a joint function of the physical separation of the two schedules and the ratio of fixed-time or fixed-interval to variable-interval schedule food presentations. The results of the five experiments are discussed in terms of the contributions of both reinforcement variables and discriminative stimuli in determining the effects of intruding response-independent food into a stream of operant behavior.     

Data Smoothing: Prediction of Human Behavior, Detection of Behavioral Patterns, and Monitoring Treatment Effectiveness in Single-Subject Behavioral Studies

Data-smoothing can be particularly useful in predicting human behavior, detecting behavioral patterns, and monitoring treatment effectiveness in highly variable single-subject behavioral experiments that cannot be determined by only visual inspection of their graphs. Using an example from the applied behavior analytic literature, the use of moving-average and exponential data-smoothing aided the detection of the unique behavioral patterns of a child with autism across different treatments. Furthermore, the utility of the data-smoothing procedures to monitor and control the effectiveness of an intervention is discussed. The ease of their calculations suggest use of data-smoothing by behavior analysts whenever the effects of particular interventions are questionable.     

Effects of a variable-ratio conditioning history on sensitivity to fixed-interval contingencies in rats.

We investigated the possibility that human-like fixed-interval performances would appear in rats given a variable-ratio history (Wanchisen, Tatham, & Mooney, 1989). Nine rats were trained under single or compound variable-ratio schedules and then under a fixed-interval 30-s schedule. The histories produced high fixed-interval rates that declined slowly over 90 sessions; differences as a function of the particular history were absent. Nine control animals given only fixed-interval training responded at lower levels initially, but rates increased with training. Despite differences in absolute rates, rates within the intervals and postreinforcement pauses indicated equivalent development of the accelerated response patterns suggestive of sensitivity to fixed-interval contingencies. The finding that the histories elevated rates without retarding development of differentiated patterns suggests that the effective response unit was a burst of several lever presses and that the fixed-interval contingencies acted on these units in the same way as for single responses. Regardless of history, the rats did not manifest the persistent, undifferentiated responding reported for humans under comparable schedules. We concluded that the shortcomings of animal models of human fixed-interval performances cannot be easily remedied by including a variable-ratio conditioning history within the model.     

Biasing the pacemaker in the behavioral theory of timing

In the behavioral theory of timing, pacemaker rate is determined by overall rate of reinforcement. A two-alternative free-operant psychophysical procedure was employed to investigate whether pacemaker period was also sensitive to the differential rate of reinforcement. Responding on a left key during the first 25 s and on a right key during the second 25 s of a 50-s trial was reinforced at variable intervals, and variable-interval schedule values during the two halves of the trials were varied systematically. Responding on the right key during the first 25 s and on the left key during the second 25 s was not reinforced. Estimates of pacemaker period were derived from fits of a function predicted by the behavioral theory of timing to right-key response proportions in consecutive 5-s bins of the 50-s trial. Estimates of pacemaker period were shortest when the differential reinforcer rate most strongly favored right-key responses, and were longest when the differential reinforcer rate most strongly favored left-key responses. The results were consistent with the conclusion that pacemaker rate is influenced by relative reinforcer rate.     

Temporal discrimination learning of operant feeding in goldfish (Carassius auratus)

Operant temporal discrimination learning was investigated in goldfish. In the first experiment, there was a fixed daily change in illumination. Eight subjects were trained to operate a lever that reinforced each press with food. The period during which responses were reinforced was then progressively reduced until it was 1 hr in every 24. The final 1-hr feeding schedule was maintained over 4 weeks. The feeding period commenced at the same time each day throughout. The food dispensers were then made inactive, and a period of extinction ensued for 6 days. The pattern of responding suggested that the fish were able to exhibit temporal discrimination in anticipation of feeding time. This pattern of responding persisted for a limited number of days during the extinction procedure. The second experiment produced evidence that operant temporal discrimination could develop under continuous illumination.     

时序信息提取机制的探索

对时序信息加工的经典研究是采用新近性判断范型,结果发现其提取机制是以新近性为基础的逆向串行搜索过程。本研究采用新近性判断范型和早远性判断范型对时序信息的提取机制进行了深入的考察,结果表明：（ｌ）提取时序信息既存在逆向串行搜索,又存在顺向串行搜索;（２）早远性判断和新近性判断任务对不同部分时序信息恢复的效应不同,其中早远性判断易化早远部分,新近性判断易化新近部分──表现为对相应部分辨别力的提高和正确反应潜伏期的缩短;（３）在不同的时序信息提取任务中,被试会根据具体条件进行反转反应。     

Temporal control of behavior: schedule interactions

In Experiment I the response that terminated the postreinforcement pauses occurring under a fixed-interval 60-second schedule was reinforced, if the pause duration exceeded 30 seconds. The percentage of such pauses, rather than increasing, decreased. There were complex effects on the discriminative control of the pause by the reinforcer terminating the previous fixed interval, depending on whether the fixed interval and the added reinforcer were the same or different. In Experiments II(a) and II(b), each reinforcement initiated an alternative fixed-interval interresponse-time-greater-than-t-sec schedule, the schedule values being systematically varied. When the response following a pause exceeding a given duration was reinforced, fewer such pauses occurred than when they were not reinforced, i.e., on the comparable simple fixed-interval schedule. There was no systematic relationship between mean interrinforcement interval and duration of the postreinforcement pause. The pause duration initiated by reinforcement was directly related to the dependency controlling the shortest pause at that time, regardless of changes in mean interreinforcement interval.