Algorithms for Reliability Estimate as a Test – Quality Indicator (Published)
This study examined the algorithms for reliability estimate as a test-quality indicator. It was discussed along the methods of estimating reliability such as: Test-retest as measures of stability, Equivalent or Alternative –forms reliability as measures of equivalence and stability, while the measures of internal consistency are Split-half, Kuder-Richardson 20 and 21, Coefficient alpha, Hoyt’s analysis of variance, Scorers (Judge) reliability and Inter-rater reliability. Also, using the reliability coefficient as a Test quality indicator was addressed and variables that affect reliability estimate are itemize as test length, test content, test difficulty, item discrimination, group heterogeneity, student motivation, students testwiseness, time limit and security precautions. Therefore, this paper recommends that, in order to demystify the course at any level of our educational system, specialist in the field of Tests, Measurement and Evaluation should be strictly allowed to handle the course and every professional teachers should be abreast to the procedural ways of estimating reliability of test in the classroom examination as a quality indicator in the teachers’ made test.
Keywords: Algorithms, Estimate, Reliability, test quality and indicator.
A GENERALIZED METHOD FOR ESTIMATING PARAMETERS AND MODEL OF BEST FIT IN LOG-LINEAR MODELS. (Published)
In this article, we proposed generalized method and developed algorithms for estimation of parameters and best model fit of log linear model for dimensional contingency tables. For purpose of this work, the method was used to provide estimates of parameters of log –linear model for four- dimensional contingency table. Parameters of higher dimensional tables can in like manner be estimated. In estimating these parameters and best model fit, computer programs in R were developed for the implementation of the algorithms. The iterative proportional fitting was used to estimate the parameters and goodness of fits of models of the log linear model. A real life data was used for illustration and the result obtained showed the best model fit for four dimensional contingency table is [BSG, BGA]. This showed that the best model fit must have sufficient evidence to fit the data without loss of information and must have the highest p-value and least likelihood ratio estimate.
Keywords: Algorithms, Categorical Data, Contingency Table, Generalized Method, Iterative Proportional Fitting., Parameters