0

I get an error message if I try to use fisher.test:

fisher.test(carprices$insurance_symboling, carprices$weight)

Error in fisher.test(carprices$insurance_symboling, carprices$weight) : FEXACT error 6 (f5xact). LDKEY=427 is too small for this problem: kval=317844542. Try increasing the size of the workspace.

Ben Bolker
  • 211,554
  • 25
  • 370
  • 453
Lara
  • 1
  • 1
  • Could we please have a [mcve]? Did you look at any of the existing answers on this topic, e.g. https://stackoverflow.com/questions/17052639/fisher-test-error-ldstp-is-too-small/17054348#17054348 ? – Ben Bolker Jul 18 '22 at 22:19
  • The `fisher.test` is applicable only to count/frequency data. Are both `insurance_symboling` and `weight` factors? The error message suggests that the crosstabulation of these two variables produces a very large table. – dcarlson Jul 19 '22 at 03:03
  • Weight is a metric factor and insurance_symboling an ordinal factor. The dataset has 205 entries – Lara Jul 21 '22 at 15:35

0 Answers0