Question

For the following scenario, identify whether a correlation coefficient can be accurately calculated and interpreted – if there is an issue, indicate the most relevant reason why we should be cautious when interpreting the correlation coefficient. You can assume all measures are quantitative.

Shawna is interested in seeing whether there is a relationship between a person’s openness to new experiences and their creativity. She collects a random sample from her community and records their scores on two questionnaires that measure these two variables. She then calculates a Pearson’s correlation between the two sets of scores.

Please read carefully

Group of answer choices

There is an issue, outliers are likely to affect the correlation

There is an issue, there is likely a non-linear relationship between the variables

There is an issue, there is likely a restriction of range in the sample

There is an issue, there are likely heterogeneous samples in the data

There is an issue, correlation should be not used to imply causation

There is no issue, a correlation coefficient can be accurately calculated and interpreted

Answer 1

A Correlation coefficient signifies the degree of relationship between the two variables. But the correlation can be affected if outliers are presents and sample size is not large enough and also it is also not necessary that correlation would imply causation (cause and effect of the variables)

Therefore, the answer choices are:

There is an issue, outliers are likely to affect the correlation

There is an issue, correlation should be not used to imply causation