1
$\begingroup$

I have some sample data with age and I'd like to put them into bins (for example, ranges 20-24 25-30, etc.). The resulting variable would constitute values from 1-8.

What I'm confused about is what type of variable is the new variable?

$\endgroup$
2
  • $\begingroup$ Why does its type matter to you? $\endgroup$
    – whuber
    Commented May 30, 2013 at 22:27
  • 1
    $\begingroup$ Unless you mean you're also subsequently turning it into a factor with labels 1 to 8, binning it just results in the bins previously indicated. If your bins are all constant-width, why use 1-8 rather than a number related to the bins, like the left ends or the centers? $\endgroup$
    – Glen_b
    Commented May 30, 2013 at 22:47

1 Answer 1

0
$\begingroup$

From the point of view of defining the variable's category in the data editor it is somewhere between ORDINAL and SCALE, I would tend towards defining it as ordinal, since you will have lost most of any refinement for analysis by the binning. However, binning could be be a good idea for the purpose of displaying the data in charts, which I guess is your reason for doing it.

$\endgroup$
3
  • $\begingroup$ Binning for visual display does not change the meaning or possible uses ("type") of the underlying variable, whereas binning-qua-numerical re-expression of the variable for further analysis can change its meaning. Ultimately, as with almost all questions referring to Stephens' typology of data, the question of "type" just gets in the way of the more important issues concerning how to analyze and present the data in the first place. $\endgroup$
    – whuber
    Commented May 30, 2013 at 22:52
  • $\begingroup$ +1. Stephens here is a typo for Stanley Smith Stevens. $\endgroup$
    – Nick Cox
    Commented May 30, 2013 at 23:37
  • $\begingroup$ Thanks @Nick. Getting the spelling correct helps with searches on our site. :-) Incidentally, here is a thoughtful and sympathetic account of Stevens' typology. $\endgroup$
    – whuber
    Commented Jul 3, 2013 at 18:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.