Talk:Morale/Fatigue/@comment-24269960-20140503071717/@comment-61.127.97.155-20140509045540

Ugh pressed enter before I finished.

You use two tailed test when you are unsure what difference you are testing to see. The general speculation is line abreast is evasive and so the alternative hypothesis to test is unarguably "line abreast increases evasion". Besides, in such a case you generally would use the 5% SI if this was 2 tailed, but it definitely is not.

I'm not sure how the 'cumulative probability' has become  0.458 and is not 0.39. What are you cumulating?

Using no approximation, line ahead is shown to be L ~ B(100,0.39) and P(L>52)= 0.00312642175108502, at 1% SI the critical value is 51; reject null. If you choose to use a 2 tail, then you cannot logically justify 1% SI.