Browsing by Author "Adepoju, A.A"

Now showing 1 - 3 of 3

An alternative technique to ordinal logistic regression model under failed parallelism assumption
(2009-12) Adeleke, K.A .; Adepoju, A.A
Maternal health status is often measured in medical studies on an ordinal scale but data of this type arc generally reduced for analysis to a single dichotomy. Several statistical models have been developed to make full use of information in ordinal response data, but have not been much used in analyzing pregnancy outcomes. The authors discussed two of these statistical models the ordinal logistic regression model and the multinomial logistic regression model. Logistic regression models are used to analyze the dependent variable with multiple outcomes which can either be ranked or not. In this study, we described two logistic regression models for analyzing the categorical response variable. The first model uses the proportion-! odds model while the second uses the multinomial logistic regression model. The fits of these models using data on delivery from a Nigerian State hospital record/database were illustrated and compared to study the pregnancy outcomes. Analyses based on these models were carried out using STATA statistical package. The Multinomial logistic regression was found to be an important alternative to the ordinal regression technique when proportional odds assumption failed. The weight of the baby and the mother's history of disease (treated or not treated) were found to be important in determining the pregnancy outcome.
Ranking of simultaneous equation techniques to small sample properties and correlated random deviates
(Science Publications, 2009) Adepoju, A.A; Olaomi, J.O
Problem statement: All simultaneous equation estimation methods have some desirable asymptotic properties and these properties become effective in large samples. This study is relevant since samples available to researchers are mostly small in practice and are often plagued with the problem of mutual correlation between pairs of random deviates which is a violation of the assumption of mutual independence between pairs of such random deviates. The objective of this research was to study the small sample properties of these estimators when the errors are correlated to determine if the properties will still hold when available samples are relatively small and the errors were correlated. Approach: Most of the evidence on the small sample properties of the simultaneous equation estimators was studied from sampling (or Monte Carlo) experiments. It is important to rank estimators on the merit they have when applied to small samples. This study examined the performances of five simultaneous estimation techniques using some of the basic characteristics of the sampling distributions rather than their full description. The characteristics considered here are the mean, the total absolute bias and the root mean square error. Results: The result revealed that the ranking of the five estimators in respect of the Average Total Absolute Bias (ATAB) is invariant to the choice of the upper (P1) or lower (P2) triangular matrix. The result of the FIML using RMSE of estimates was outstandingly best in the open-ended intervals and outstandingly poor in the closed interval (- 0.05
Sensitivity of estimators to three levels of correlation between error terms
(2010-05) Adepoju, A.A; Iyaniwura, J.O.
A Monte Carlo simulation is employed to investigate the sensitivity of simultaneous equation different levels of correlation between random deviates. Three arbitrary levels of correlation between pairs random deviates were assumed. Three small sample sizes were used in this experiment, N = 15, N=25 and N 40 each replicated 100 times. A number of factors should be taken into account in choosing an method Although system methods are asymptotically most efficient in the absence of violation of independence of random errors, system methods are more sensitive to any kind of error than single equation methods. In practice, models are never perfectly specified nor are they completely free of correlated random deviates. It is a matter of judgment whether the correlation is strong enough to warrant avoidance of system methods. As sample size increases, the TAB for all the estimators decreased consistently except for FIML. OLS, 2SLS, LIML and FIML are remarkably insensitive to the choice of triangular matrices (P1 and P2) when using TAB to judge their performances. Best RSS estimates of 2SLS, LIML, and 3SLS found in the feebly correlated region