Consider a correlation coefficient of .095 between variables A and B. The correlation coefficient was calculated from a sample of 1100 respondents and is statistically significant, p < .0001. Also, consider a correlation coefficient of .891 between variables C and D. The correlation coefficient was calculated from a sample of 15 observations and is statistically nonsignificant, p > .10.

So, when should marketing analysts consider a correlation strong enough to even consider given the issue raised above?