Bivariate research techniques

Bivariate research techniques consist of a variety of statistical testing methods used in market research to analyse the relationship between two variables. One variable is frequently labelled as the independent variable, which is usually demographic, geodemographic or behavioural in nature, and the other is known as the dependent variable.

There are countless uses for bivariate research techniques. One example could be within education market research, where it is possible to analyse the relationship between a child’s gender and their performance in certain exams.

There are many different statistical methods within the general field of bivariate analysis. One of the most common methods employed during market research is bivariate regression analysis, also known as linear regression.

Naturally, different forms of bivariate research techniques are suited to different types of variables. The choice of analysis method also depends greatly on the desired level of measurement of the variables. Examples of other types of bivariate analysis are probit regression, logit regression, rank correlation coefficient, ordered probit, ordered logit, simple regression or vector autoregression.

There are certain limitations with bivariate research techniques. For example, this only takes in to account the relationship between two variables. In order to see how a third variable affects the others, multivariate analysis techniques are needed. Bivariate analysis usually does not factor in how a variable could influence the other, and therefore cannot give an explanation for the relationship between the two variables, but only provide a description. Explanatory analysis is necessary to infer cause.

Within the range of bivariate research techniques, there is such a thing as bivariate explanatory analysis, which as mentioned previously, is needed to explain a cause behind any identified relationship of variables. As before, deciding on which bivariate explanatory analysis technique depends on the level of measurement of variables. There are two types of measures of influence; these are asymmetric measures (where a direction of influence exists) and symmetric measures (where no direction of influence exists).