would the null hypothesis that the 2000 mean hourly earnings of all workers was $16 be rejected at the 10% significance level in favor of the two-sided alternative? What about the null hypothesis that the mean was $15?