- Through its reported probability value, it tells you the likelihood that you would find the observed relationship, if the null hypothesis were really true for the population.
- That is, how likely is it that this relationship we found is simply due to sampling error?
- Or, how probable is it, given our findings, that the null hypothesis is really true?
- A probability value of less than .05 is the social science standard for rejecting the null hypothesis.
- This logic also applies to the probability values associated with other measures of significance.
NOTE: Chi-square is not a measure of the strength of a relationship. All it does it tell you is you can infer this relationship to the population.Three factors influence the value of Chi-square:
- The strength of the relationship
- The size of the sample
- The size of the table