Enhancement of Survey Research: A Statistical Perspective
Surveys are carried out to study the characteristics of a population by studying a sample a sub-set of population. An accurate sample which is representative of population is essential to gather representative data to draw statistical inferences and conclusion for decision making using the outcome of survey. However, there is a lacuna on part of selection of the type study variables— continuous, nominal, ordinal or ratio and on the needed statistical test which can be run on the gathered data. This situation leads to a wrong representation results and improper decision making. Therefore, in this review the researchers elaborate on the selection of appropriate statistical test for analyzing the survey data and interpretation of results. This manuscript cover most commonly used statistical stets for survey data association, reliability, predictions, differences between groups and calculation of sample size for a study. The researchers also provided on the use of binomial, ordinal and multinomial logistic regressions and the authors assume that researchers carrying out a study is well-versed with sampling methods, and emphasizing the sampling methods is out of scope of this paper. The researchers believe that the students, research scholars, and other research community who undertake the survey research will immensely benefit from our article.
Keywords: Survey, continuous, ratio, variable, ordinal logistic regression, lacuna, population.