HYBRID CODING METHOD FOR PLANNED COMPARISON OF TREATMENTS IN REGRESSION ANALYSIS.

Author: Etaga Harrison Oghenekevwe
Department: Statistics
Affiliation: Nnamdi Azikiwe University Awka

Coding in regression analysis has been widely used with varying applications. Various
coding structures have been developed in the past. Coding structures such as Dummy, Simple, Hermert, Reverse Hermert, Forward Differencing and Backward Differencing were developed to deal with specified (simple) hypotheses. A mixture of Dummy, Simple and Contrast coding was considered and studied to yield a hybrid coding which when applied correctly will not only yield the results of simple, dummy, or contrast coding but can also give other results. Existing data were used to illustrate the new method. Live Data on Rainfall, GDP, Oil and Non-Oil exports, Government expenditures obtained from the Central Bank of Nigeria Statistical Bulletin of year 2009 were subjected to the various methods for effective comparison of the existing methods. The results obtained showed that; Coding in regression analysis is superior to the chow-test as it was able to simultaneously test for the shift in intercept as well as in slope of given models. Regression was also better than the conventional ANOVA and t-test as it expressed clearly the magnitude of any observed differences. The new HCM was applied and the result obtained showed that it should be used to test specific hypotheses under planned comparison as it compared favourably with the dummy.

For full copy of this document please visit the digital library help desk.

Advertisements

Tags: , , , , , ,

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: