0
$\begingroup$

I have dataset which is markedly left-skewed, and I wonder if it will be inappropriate to use the normal distribution curve to analyses the data given it's highly skewed?

Picture of my data

$\endgroup$
  • $\begingroup$ More importantly, you data seems to be positive. Normal distribution is almost never suited for such cases. $\endgroup$ – Federico Nov 21 '18 at 16:19
  • $\begingroup$ Do you why that is? :) $\endgroup$ – pete415 Nov 21 '18 at 16:24
  • $\begingroup$ What? Why your data is positive? Or why the normal distribution is not suited for positive data? I think both are pretty clear... $\endgroup$ – Federico Nov 21 '18 at 16:26
  • $\begingroup$ How exactly do you want to analyze your data? $\endgroup$ – lisyarus Nov 21 '18 at 16:27
  • $\begingroup$ Can you tell us more about the underlying data? Is it continuous or discrete? Bounded from below/above? It's not clear from the picture $\endgroup$ – Federico Nov 21 '18 at 16:31
1
$\begingroup$

Your data seems to be non-negative. The normal distribution is almost never well suited for such case. You might want to try a gamma distribution, possibly just an exponential distribution.

Moreover, you should consider if your data is continuous or discrete. In the latter case, you might prefer Poisson distribution or geometric distribution.

If your data is also bounded from above, then beta distribution and binomial distribution are an option.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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