Consider an ordinal variable as continuos for correlation with a continous variable

by Titus Pullo   Last Updated May 16, 2018 14:19 PM

I need to analyse the results of a survey. Among the different questions I have: "How happy did you feel last month"?. There are 5 possible answers:

  1. Not at all
  2. Not very
  3. Somewhat
  4. Very
  5. Completely

Now, I would like to measure if there is a correlation between age and happiness. One option is to simply plot the age on the x-axis (from 18 to 99), the happiness values (from 1 to 5) on the y-axis and then check for a possible trend. However, I'd like to have a more rigorous statistical test. Given that happiness is an ordinal variable, would it be wrong to treat it as a continuous variable and calculate either Pearson or Spearman correlation coefficient? If this approach is correct, can it be done for binary variable as well? E.g gender: Male or Female?



Related Questions





Pearson correlation binary vs continuous

Updated October 22, 2017 15:19 PM



Cache file /home/queryxchang/questarter.com/apps/frontend/config/../cache/-q-consider-an-ordinal-variable-as-continuos-for-correlation-with-a-continous-variable-20-346541-html could not be written