The data from problem 13 show no significant relationship between the type of game and helping behavior. To construct the following data, we simply doubled the sample size from problem 13 so that all of the individual frequencies are twice as big. Notice that the sample proportions have not changed.

a. Test for a significant relationship using a 5 .05. How does the decision compare with the decision in problem 13? You should find that a larger sample increases the likelihood of a significant result.
b. Compute the phi-coefficient for these data and compare it with the result from problem 13. You should find that the sample size has no effect on the strength of the relationship.

Prosocial Violent Neutral 14 16 30 24 16 14 30 54 Helped Did not Help 6 36 30