Creating a correlation matrix for conjoint simulations

Creating a correlation matrix for your simulations is easy.

  1. Run your preference and revenue share simulations by inserting the product names, attributes and levels that you want.
  2. Click Show advanced settings then View correlation Matrix

Now you can see your correlation matrix where the header corresponds to the name of products/attributes/levels combinations you inserted earlier.

You may also export your correlation matrix to CSV or Excel format.

Correlation matrix for simulations

How to interpret correlation?

Pearson correlation coefficient, ρ measures the degree of relationship between two variables. The range of correlation coefficient will be from -1 to +1, where:

  • ρ = +1 indicates strong relationship between two variables. Example: If respondents like Coca-Cola, they also tend to like Pepsi.
  • ρ = 0 indicates that there is no relationship between two variables. Example: Price of an orange and price of a table have zero correlation with each other. If price of an orange increases, nothing can be inferred about the price of a table because they are not dependent on each other.
  • ρ = -1 indicates that when one variable increases, the other variable decreases (moves in opposite direction). Example: As you climb a mountain (increase in height), the temperature decreases.

Take our Claims Test for Yogurt for example, we can see that:

  • Claim 1 and claim 10 have high positive correlation (ρ = 0.7). This suggests that respondents who like claim 1 tend to like claim 10 as well.
  • Claim 1 and claim 7 have zero correlation (ρ = 0). This suggests that if respondents like claim 1, it is uncertain if they will like claim 7.
  • Claim 8 and claim 11 have high negative correlation (ρ = -0.8). This suggests that respondents who like claim 8 tend to dislike claim 11.
Example of the correlation matrix for simulations