Hypothesis Testing

Hi, I think I have this down pretty good (finding confidence intervals, etc.), except for one problem. I always get the opposite answer than what Schweser’s are. I get the same interval and test statistic but I choose the opposite reject/fail to reject decision. For instance, there was one question where the test statistic asked if the mean score on a test was greater than 100. I came up with a confidence interval of 1.66 and the test statistic they supplied was something like 3.5. I figured 3.5 is greater than 1.66 so you should fail to reject. However, they say 3.5 is greater than 1.66 so you should reject. I can’t figure out why. Can anyone help?

hi, if you hypothesize H0 <= X, you are implicitly saying/expecting that the test statistic should lie below the critical value (1.66 in your case). a test statistic of 3.5 is not less or equal to 1.66, so this goes against your expectation and you must reject your hypothesis… supporting your H1. remember, you state your H0 contrary to your “true” belief and hope to be able to reject it.

a sample stat outside of your confidence interval = reject. find a graph showing the standard normal parabola and that should help it sink in.

Try using the search function. I think someone answered this is detail last fall…

Can any body tell me if CFA Institute provides the complicated formulas or we have to memorize them? What about statistical tables?

Masood - you have to memorize the formulae. As far as statistical tables go, they’ll give you the info you need, but know what std dev corresponds to 68% confidence, 90%, 95% and 99% (1, 1.645, 1.96, and 2.58) respectively.