Effects of training to increase self-monitoring accuracy

We examined the effects of teaching children to monitor and record their attending behavior accurately. On the basis of their low levels of attention to task and agreement with an observer about their attending, we selected four students from a class of 11, all of whom were practicing self-recording. The classroom teacher trained the students to make their judgments about their attention to task correspond with hers. The accuracy training, which was applied according to a multiple baseline (across subjects) design, produced salutary increases in levels of attention to task. But changes in accuracy of the students' self-recording were small and unstable. Although the procedures used in this study produced high levels of reactivity, they do not support the position that accuracy in self-monitoring is required for reactivity.     

个体和群体方式绩效评估的准确性研究

本研究通过对 2 1 0名大学生以个体和群体方式对两个学科评估结果的研究 ,分析了个体与群体评估方式以及不同的评估标准对评估过程和评估结果的影响。研究结果表明 :1群体评估方式有利于减少评估中的晕轮效应 ;2群体方式有利于提高对评估信息的利用水平 ;3群体方式有利于提高评估者评分倾向 ,即代表性认知的一致程度 ;4群体方式有利于提高评估者理解绩效信息意义的一致程度 ;5群体方式评估时评估者对绩效信息结构的理解更为合理 ;6提供较为详细的评估标准有利于提高对信息的利用水平 ,有利于提高评估者评分倾向的一致程度以及评估者对学科差异的敏感性。     

Hick's Law in a Stochastic Race Model with Speed-Accuracy Tradeoff

We present an analytic solution for a race model of n stochastic accumulators for multiple choice reaction time. We show that to maintain a constant level of accuracy, the response criterion needs to be increased approximately logarithmically with n, to compensate for the increase with n in the likelihood that an incorrect alternative will be most active after any fixed amount of time accumulating information. Assuming that participants monitor and maintain a constant level of performance can then explain the logarithmic dependency of the response latency on n as specified by Hick's law. Moreover, we show that for short time intervals, the Shannon information that observers extract from a stimulus, is predicted to increase linearly with processing time.     

理解评估与成绩预测：两种不同的元理解监测形式

探讨理解评估与成绩预测与各种强化元理解监测线索的认知任务的关系。结果发现,理解评估与成绩预测的判断值偏离标准测验成绩的程度受监测线索强化方式的调节;主动强化监测线索比被动强化更能提高理解评估和成绩预测的精确性;精确的理解评估或成绩预测所需的线索不同,利用同一线索评估理解或预测成绩,其精确性也不同。这一结果挑战了元理解监测的一维观,表明理解评估与成绩预测涵盖了元理解监测不同方面的心理特征     

重复学习判断中的低自信效应

学习判断是元认知判断的一种重要形式,是对未来回忆成绩的预测。重复学习判断中的低自信效应是一个有趣而反常的现象,指在重复学习测验中,学习判断往往低估了回忆成绩的增加,出现低自信的一种现象。目前,大部分实验证据支持重复学习判断中的低自信效应是普遍存在的,但也有一些实验结果表明该效应是受一定条件限制的。对于这种低自信效应的作用机制,研究者提出了不同的理论模型（如,锚定调整假说、记忆偏向消除理论、过去测验记忆的启发式假说等）进行了深入探讨。最后,文章还指出了此领域现有研究的局限和有待研究的问题     

新手考官的工作记忆和人格对其评分有效性的影响——以无领导小组讨论为例

相对于其它评价中心技术而言,在无领导小组讨论中考官因素对评分结果的影响尤为重要.本研究主要探讨无领导小组讨论中新手考官的工作记忆与人格对其评分有效性的影响.结果发现,首先,新手考官的评分者一致性较低,评分准确度较差.其次,工作记忆和人格的部分因素分别从不同方面影响新手考官的评分有效性,具体表现在：（1）利他性越强,新手考官评分总均值的准确性越高,且评分结果越宽松;（2）新手考官的决断性越强,对所有应聘者做出有效区分的准确性越高;（3）新手考官的沉稳性越高,对各维度的区分越有效;（4）注意转换和抑制能力对新手考官的晕轮效应及其在各个维度上进行区分的准确度有抑制作用.     

Memory load as a cognitive antidote to performance decrements in data entry

In two experiments, subjects trained in data entry, typing one 4-digit number at a time. At training, subjects either typed the numbers immediately after they appeared (immediate) or typed the previous number from memory while viewing the next number (delayed). In Experiment 2 stimulus presentation time was limited and either nothing or a space (gap) was inserted between the second and third digits. In both experiments after training, all subjects completed a test with no gap and typed numbers immediately. Training with a memory load improved speed across training blocks (Experiment 1) and eliminated the decline in accuracy across training blocks (Experiment 2), thus serving as a cognitive antidote to performance decrements. An analysis of each keystroke revealed different underlying processes and strategies for the two training conditions, including when encoding took place. Chunking (in which the first and last two digits are treated separately) was more evident in the immediate than in the delayed condition and was exaggerated with a gap, even at test when there was no gap. These results suggest that such two-digit chunking is due to stimulus encoding and motor planning processes as well as memory, and those processes transferred from training to testing.     

非参数认知诊断方法下诊断结果的概率化表征

非参数认知诊断分类方法非常适合课堂评估,其诊断结果采用0-1形式而缺乏概率化表征,不能精细地区分被试属性掌握程度的差异或变化,还缺乏可用于评价真实测验分类结果的信度和效度指标。要刻画被试属性掌握程度的差异,首要的问题是要为非参数认知诊断方法提供一种可以量化属性掌握概率的方法。针对此问题,基于二项分布和玻尔兹曼分布提出非参数认知诊断方法下诊断结果的概率化表征方法,并用于构建分类准确性和分类一致性指标。模拟研究与实测数据分析结果显示：概率化表征方法与非参数认知诊断方法的分类结果高度一致;概率化表征方法与认知诊断模型所得的属性掌握概率十分接近;概率化表征方法所得的属性（模式）掌握概率可用于计算属性（模式）分类准确性和分类一致性指标,在实际测验情景下可作为信度和效度指标,评价诊断结果的重测一致率和判准率。