Questionnaire: Chocolate consumers

Authors: Caroline DOUX and Charlotte LE COZ

We decided to study the different kinds of chocolate's consumers with these principal directions: how often do you eat and buy chocolate? How do you eat it? How do you see chocolate? We focused on chocolate bars.

We built our questionnaire looking for different profiles. We expected to find some addicts, some greedy persons and others who do not especially like chocolate. However, we will see that we had little surprises during our analysis.
Our dataset is made of 120 individuals and 24 variables.

We first used a Multiple Correspondence Analysis and studied the first two axes.

  • The first component opposes people who can absolutely deal without chocolate, never eat chocolate and rarely buy it (positive coordinates) to people who cannot deal without chocolate, are almost addicted and eat and buy it very often (negative coordinates).
    Our first axis is thus about chocolate addiction. We notice that "keep in the cupboard" has a negative coordinate whereas "keep in the fridge" has a positive one. People who rarely eat chocolate put it in the fridge to keep it safe.
  • The second axis opposes people who eat chocolate often and alone and accompany it with nothing (positive coordinates) to persons who sometimes eat almond chocolate with juices during their snack (negative coordinates).

Then we performed a hierarchical clustering. We obtained 3 clusters:

  • In the first one, we have addicted persons: they buy and eat chocolate very often and they cannot deal without it.
  • People in the second clusteris mostly made of women who buy and eat chocolate occasionally. They eat chocolate with almonds and accompany it with tea or juices.
  • The third cluster gathers people, especially boys, who scarcely buy and eat chocolate. They eat it in bed or when working and keep it in the fridge

As a conclusion, we can say that we almost obtained what we expected: addicted persons, occasional consumers, and people who do not eat chocolate.

Find here the data set and the R script:

The results of the ENMCA() function are the following:

Beware the results of HCPC() and ENMCA() are not quite the same since the data are ventilated and the method of classification is a bit different.