BIOL 373 Lecture Notes - Lecture 8: Ratio Test, Biostatistics, Test Statistic
Document Summary
2 s1 (whichever is larger, largest should be in the numerator) Compare to critical value f0. 05(2),df1,df2 or f0. 05(2),df2,df1. Hypotheses: for two-tailed tests: h0 = s1, for one-tailed tests: Note: don"t forget to report the actual p value for your calculated test statistic, not just whether or not it exceeds the critical value for the test statistic. Comparing f values to the f distribution. F tables relate critical f values with the probability of being in the curve tail: critical values are used to make additional inferences. If assumptions of the t test are severely violated: try data transformation (13, use a nonparametric test! (8. 9-8. 10, these do not use parameter estimates, and do not make assumptions about the form of the population distribution. Log-transformed data: x"=log(x) [or x"=log(x+1), to avoid x=0, taking the log spreads out lower values. Does not cause any significant statistic differences.
