to price reductions by exploring how promotions are reflected in the demand for a
product. Or any other quantitative investigation of how different marketing mix
variables affect consumers’ brand behaviour.
Data and analysis in the economic approach
Because methods are mainly quantitative in the economic approach, large quan-
tities of data are preferred. The big data samples can be used to deduct correlations
between variables and are suitable because of the need for data and results to be
replicable; it is important to ensure that they are representative. The disadvantage
is that it is difficult to gain a sound understanding of why variables are correlated,
because the data are sampled broad instead of deep. As opposed to qualitative
research methods, where smaller samples deliver rich and descriptive conclusions,
the results of quantitative research designs are often expressed in tables or other
statistical representations of data.
The economic approach rests upon a positivist research ideal. This line of
thought is very much in opposition to the stream of qualitative research methods
that have become more and more dominant in marketing research in recent
decades. In the quantitative methods, objectiveness is important and phenomena
are presumed to be measurable. The objectivity of data is important for validity,
and closeness to the subject of research is not essential. Data like scanner panel
data from cash registers at supermarkets and laboratory experiments are
considered valid. The data are then subjected to different kinds of statistical
analysis that often consist of some sort of regression analysis.
The economic approach 41Box. 4.3 Regression analysis
Regression analysis is a statistical tool for the investigation of relationships
between variables. Any regression study sets out with a hypothesis that the
investigator formulates, e.g. when a brand is on promotion (the price is
lower than usual) then the demand will increase. After having formulated
the hypothesis the researcher assembles data on the variables of interest –
data on price levels over time and data of demand levels over time. The data
are then subjected to regression analysis that estimates the quantitative
effect or correlation between the variables. Hereby it is possible to ascertain
the causal effect of one variable upon another; hence how a price cut in the
shape of a promotion affects sales. The illustration reflects the correlation
between how price promotions affect demand (the xaxis reflecting the price
and the yaxis illustrating the fluctuations in demand). It is clear that
whenever the product has been on promotion – sold at a reduced price – the
demand for the product is higher.
Once a correlation between two variables has been established the statis-
tical significance of the estimated relationships can be seen. In figure 4.6 the