One-tailed Hypopheses testing. F-statistic.

Hello. I have some misunderstanding. For example we have to populations with associated S1 n1 and S2 n2 standard deviation and sample size correspondingly. If i want to test wether S1^2 is greater S2^2 how shold I define null and alternative hypotheses and calculate appropriate F-statistic. Please explain the logic. Thanks.

Always put the bigger one on top and halve your alpha…

Does it mean that the form of Hypopheses condition does not matter? I just put the bigger variance in the numirator, then find upper critical value of F statistic using numerator degree of freedom in horisontal and denumerator degree of freedom in vertical values of F distribution table?

Right. Remember that by putting the bigger one on top, you need to halve alpha to look up critical value.

" you need to halve alpha to look up critical value" alpha - level of significance?


Am i right that it is stated in condition of the test? e.g. “with the 5% level of significance…”

do you think/know if L1 candidates are expected to know the hypothesis testing formulas (computing the test statistic etc)? should I expect -for this particular part of quant - some “qualitative” questions (interpretation) or proper calculations? thanks

What does the LOS say this year? They are usually very specific about this question.

they don’t say “calculate” but mostly “define”, “identify” or “interpete”. Hence I don’t expect computations on the test