Goodness-of-fit test - observed vs. expected (some expected values < 5)

Question

[Lucy]

Hi,

I would like to do a goodness-of-fit test in R to compare observed vs. expected values for a vector of length 6. The expected values are small so I have read that it is not recommended to perform a Chi-squared test, however I am struggling to find a good alternative. Some articles seem to suggest that I could use a Fisher's exact test while others suggest otherwise.

Help would be much appreciated.

Best wishes,

Lucy