Quantitative techniques

Descriptive analysis

Direct and cross sorting

Multivariate analysis

  Nominal variables Numeric variables
Nominal variables

2 variables:
Chi Square Test

More than 2 variables
Correspondence Analysis (AC)

2 variables:
Analyse of Variance (ANOVA)
t-test

Numeric variables

2 variables:
Analyse of Variance (ANOVA)
t-test

2 variables:
Correlation test
Simple regression analysis

More than 2 variables:
Principal Component Analysis (PCA)
Cluster Analysis
Multiple Regression Analysis

Principal Component Analysis (PCA)

What it is.

Principal Component Analysis provides a means of displaying both respondents and variables on a two-dimensional map. It is conceptually similar to Correspondence Analysis, but applies to numeric variables (continuous or scale). Combined to Cluster Analysis, Principal Component Analysis allows to visualize the clusters of respondents on a perceptual mapping.

What it is used for.

Principal Component Analysis is used in Attitudinal Segmentation studies to display on a two-dimension map the clusters of respondents who share the same beliefs and practices.

Principal Component Analysis also displays the relationships between variables (or correlation rates).

Back to list

Correspondence Analysis (AC)

What it is.

It is conceptually similar to principal component analysis, but applies to nominal variables rather than numeric variables. In a similar manner to Principal Component Analysis, it provides a means of displaying or summarising a set of data in two-dimensional graphical form.

What it is used for.

Correspondence Analysis is used to make perceptual mapping illustrating graphically relationships between variables. It allows to place on a two-dimensional map respondents categories (gender, age categories, specialties, place of practice, etc.) and respondents beliefs & practices (What they think or do).

Back to list

Cluster Analysis

What it is.

Cluster Analysis (or clustering) classifies respondents into homogeneous sub groups. The most commonly used technique in market research is the hierarchical cluster analysis which applies to a series of numeric variables (continuous or scale).

What it is used for.

Cluster Analysis is used for every segmentation study (Attitudinal Segmentation and Needs based Segmentation). This segmentation studies identifies and describe homogeneous market segments in order to best adapt the drug communication and services offering.

Back to list

Multiple Regression Analysis

What it is.

Multiple Regression Analysis calculates the relationship between one dependent variable Y (which we need to explain and predict) and several independent variables X1, X2, X3, etc. explaining Y. Multiple Regression Analysis can be linear or non linear (logarithmic, exponential, etc.). In any cases, Multiple Regression Analysis provides a regression function (eg. Y=aX1 + bX2 + c) and a correlation coefficient r ranging from 0 to 1 estimating the efficacy of the regression function to predict Y value from X1, X2, X3, etc.

What it is used for.

Multiple Regression Analysis is used in Attitudinal Segmentation study to provide the scoring system which can individually allocate customers to the identified and described market segments.

Back to list