ERIC Number: ED053394
Record Type: RIE
Publication Date: 1971-Jun-28
Reference Count: 0
Some Mathematical Models of Individual Differences in Learning and Performance. Psychology and Education Series.
With the advance of computers, extensive work has been undertaken in the field of programmed instruction. Much effort has been invested to devise schemes of optimal instruction with respect to suitable criteria. Yet, what is needed is a theory which prescribes how learning can be improved, i.e., a theory of instruction. The present study is motivated by the absence of adequate formalization of individual and item differences. The results of the study demonstrate unequivocally that the One-Element Model (OEM) with the heterogeneity provision is still a fairly accurate model. More significant is the observation that individual differences have a first order effect on the predictive power of simple stochastic models. In addition, the hypothesis was confirmed that the heterogeneity assumption increases the predictive power of simple learning models and has a sizable effect on their learning properties. Finally, in the context of computer-assisted instruction in elementary mathematics, results demonstrated that asymptotic performance data can be accounted for successfully by probablistic automation models with few parameters. (Author/TA)
Publication Type: N/A
Education Level: N/A
Sponsor: National Science Foundation, Washington, DC.
Authoring Institution: Stanford Univ., CA. Inst. for Mathematical Studies in Social Science.