A statistical model is given for representing the several components of variability present in measurements (e.g., DMFS scores) given by examiners to patients. Methods for making inferences about the intraclass correlation coefficient of reliability are presented and illustrated on a real set of data. The proper analysis of data from a reliability study is shown to depend on the planned design and analysis of the clinical or field trial to be conducted following the reliability trial.
Overall, these results support the view that dental-facial esthetics and self-perceptions of occlusal appearance, as well as attitudes toward malocclusion and orthodontic treatment, are important factors in the individual's decision to obtain orthodontic treatment. The data indicate that psychosocial variables not directly related to occlusal problems and treatment, do not add significantly to the description of differences between those who seek treatment and those who do not. While these analyses did not include a number of variables--such as socioeconomic factors or attitudes of significant others--which may also be important in predicting individual response to malocclusion, they do suggest strongly that psychosocial factors directly related to need for treatment are probably most productively assessed in terms of self-perceptions of occlusal appearance and attitudes regarding dental malrelations and their treatment.
In an examination which incorporates penalties for incorrect responses, does the guessing strategy followed by an examinee have any effect on his test score? Several studies (Sherriffs and Boomer, 1954; Slakter, in press;Votaw, 1936) have indicated that when examinees are requested to respond to items previously omitted, their mean test score will increase. In other words, an examinee following a conservative guessing strategy would tend to increase his test score by answering all items, even though the usual penalty for incorrect responses is in effect.However, two recent experiments investigating the effect of guessing strategy on test score report somewhat conflicting findings. Hammerton (1965) found that when examinees were instructed to answer all questions, they scored higher on vocabulary tests than when instructed to leave out questions if not confident of the answer. On the other hand, Taylor (1966) studied three different instruction groups (do not answer unless certain; do as well as you can; answer all questions), and found no differences in mathematics test scores due to instructions. These contradictory results by Taylor might be due to the tendency for partial information to be less of a factor in mathematics examinations than in other types of examinations; e.g., vocabulary examinations. However, the contradictory findings may also be explained by reference to a table presenting the number of omitted and unfinished items for each of the three treatment groups (Taylor, 1966, p. 4). Reanalysis of the data in the table appears to indicate that there was little difference in guessing among the groups. The Taylor results, therefore, may be misleading because of the failure to manipulate guessing effectively with the directions.The present experiment, in order to study the effect of guessing strategy on test score more thoroughly, consisted of two parts. The first part was concerned with conventional directions; the second with Coombs type directions (Coombs, 1953; Coombs, Milholland, & Womer, 1956); i.e., the examinee is instructed to select those alternatives that are incorrect, with resultant losses and gains in test points depending on whether or not the correct (or keyed) alternative is selected, and how many alternatives are selected. Specifically, the following guessing strategies were examined:1. For conventional do-not-guess directions, which of the following strategies results in a higher average test score? a. You will receive one point for each correct answer. For each incorrect answer, you will lose 1/3 of a point. If you omit a question, 1The assistance of Roger Koehler and Carl Terranova is gratefully acknowledged.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.