Aporoved for oublic retease; dstrnbutton unkmited.
.,ZThis report. as submitted by thee contractor, has been cleared fcoa release to Deense Technical Information Center"(OTIC) to comply with regulatory requirer"Sns. it has been given no Primary distribution other than to DTIC and well be available only thro'ugh DTIC or other referencet tervices such at the National Technical Information % * S4rvlce (NTIS). The wiws. cn-p. n, anelor fending. contaiined in thit report are tho-se of the authorlel and ishould not be consttoue as an Oticom; .epajernei of the Army position policy, or decision. unlest to des-gnal t"" Approved for public release distribution unlimited. This research note indicates that novices judge importance in texts according to sentence type category (definitions, facts, equations, etc.). Subjects varying in expertise judged the importance of sentences in physics texts when < these sentences were presented in one of two forms: 1) definitions or facts, 2) equations or verbal formulae. Experts and subjects without physics training judged these variants similar or equal in importance. However, beqinninq physics students judged definition and equation versions as more important. " (OVEP) Novice RulesList of Tables r Table 1: Sample target sentences 38 Table 2: Parameter estlma!es, standard errors, and ratio of estimates to standard 40 errors for a loglinear model of the ratings data from the experimental groups (Experiment 1).Asterisks indicate coefficients that differ from 1"zero by more than 2 standard errors. Table 3: Parameter estimates, standard errors, and ratios of estimates to 42 standard errors for a loglinear model of the ratings data from the control groups (Experiment 1). Asterisks Indicate coefficients that differ from zero by more than 2 standard errors. Table 4:Parameter estimates, standard errors, and ratios of estimates to 43 standard errors for a logistic regression model of the sentence selection data from the experimental groups (Experiment 1). Table 5:Parameter estimates, standard errors, and ratios of estimates to 44 _ standard errors for a logistic regression model of the sentence selection data from the control groups (Experiment 1). Table 6:Parameter estimates, standard errors, and ratio of estimates to standard 45 errors for a loglinear model of the ratings data from the experimental groups (Experiment 2).Asterisks indicate coefficients that differ from zero by more than 2 standard errors. Table 7:Parameter estimates, standard errors. and ratios of estimates to 47 standard errors for a loglinear model of the ratings data from the control groups (Experiment 2). Asterisks indicate coefficients that differ from zero by more than 2 standard errors. Table 8: Parameter estimates, standard errors, and ratios of estimates to 49 standard errors for a logistic regression model of the sentence selection data from the experimental groups (Experiment 2). Table 9:Parameter estimates, standard errors, and ratios of estimates to 50 standard errors for a logistic regression model of the sen...