What Is Regression Analysis?
Regression analysis usually involves identifying the relationship between a dependent variable and one or more independent variables. Simple linear regression is a case where we have one independent variable while multiple linear regressions refer to a case where we have more than one independent variable under study (Sharma, 2005). A regression model is thereafter obtained and statistical tests are conducted to check for the validity of the model. An example of a case where multiple linear regressions can be preferable is when we are interested in obtaining the exam scores of college students given the data on daily hours of study, participation in games, and the psychological status of the student during the exam. We may for instance conduct our analysis and find that: exam score=45+1.5hours of study+0.2participartion in games+0.1psychological status. This would be our multiple linear regression model and we may conduct various tests to check for the significance of the model and the individual predictors (Archdeacon, 1994).
What Is Correlation Analysis?
Correlation analysis usually helps us to measure the strength of correlation between two variables. We obtain the correlation coefficient which is usually a value that ranges between -1 and 1. The closer the coefficient is to 1, the stronger the correlation while the closer the coefficient is to -1, the weaker the correlation between the variables. A correlation of 0.89 is a strong positive correlation between the involved variables.
What Does Reliability Mean?
Reliability refers to the measure of the consistency of results produced by a system. This can be gauged by computing the correlation coefficient or in some cases the reliability coefficient (Graff, 1995). The reliability coefficient is the ratio of the true value of variance to the value of the total variance of the test scores. Reliability coefficient=1-Error Variance Observed Variance
For this study, we conducted item analysis using reliability analysis to determine which items to include or exclude from a scale. The items to be selected are those that yield higher summed scores than the rest. We are expected to compute the total correlations between the item scores for a dimension and the total scores. We assume that the items are linearly related to a single test plus a random measurement error. The analysis could for instance give us the output below
The Cronbach’s alpha is the measure of internal consistency and the high the value, the more reliable the item is.
Validity refers to the credibility and believability of the results from a test. The credibility and reliability can be increased by ensuring that the goals and objectives of the test are clearly written down (Shuford, 1967). We may also compare our measures with the existing measures to enhance the validity and reliability of results.
Graff, D. (1995). Highway safety: Reliability & validity of dot crash tests. Place of publication not identified: Diane Pub Co.
Sharma, A. K. (2005). Textbook of correlation and regression. New Delhi: Discovery Publishing House.
Shuford, E. H. J., & SHUFORD-MASSENGILL CORP LEXINGTON MASS. (1967). How to shorten a test and increase its reliability and validity. Ft. Belvoir: Defense Technical Information Center.
Archdeacon, T. J. (1994). Correlation and regression analysis: A historian's guide. Madison, Wis: University of Wisconsin Press.
If you are the original author of this essay and no longer wish to have it published on the SuperbGrade website, please click below to request its removal: