691 / 762
Chapter 15: Chi-Square Tests
691
.: For a significance level 0.01, reject the null hypothesis if the
computed test statistic value,
= 45.9854 >
= 9.210.
Conclusion
: Since 45.9854 > 9.210, reject the null hypothesis. That is,
at the 1 percent level of significance, there is enough sample evidence to
claim that the criterion (criteria) used in purchasing a smart phone is (are)
dependent on the gender of the person making the purchase.
Figure 15-25
shows the rejection region. Observe that the test statistic falls
way in the right of the rejection region. A test statistic for the test being so
far to the right in the rejection region demonstrates the significance of the
test.
Figure 15-25:
Rejection Region for
Example 14-6
Figure 15-26
provides a visual of the contributions of the individual terms
in the
test statistic value. Observe that both the contributions for the
appearance and features criteria for the different gender seem to be
significantly different while the contribution for the cost criteria, seem to
show no significant difference. These observations would support the