t检验（T test）.docVIP

t检验（T test） [statistics] T test, F test and statistical significance (P value and SIG value), to understand the difference can also see the 10:451 on Friday October 12, 2007, the origin of T test and F test Generally speaking, in order to determine the sample (sample) statistical results to the overall error probability inference, we use some statistical methods developed by statisticians, statistical verification. By a statistical test to the values obtained with statisticians established the probability distribution of random variables (probability, distribution) were compared, we can know how many% chance will get the results. If, by comparison, the probability of this result is very few, that is to say, in the few opportunities, very rare circumstances; that we can have confidence that this is not a coincidence, is statistically significant (the use of statistical terms, is to be able to reject the null hypothesis of null hypothesis Ho). On the contrary, if the comparison result, the probability is very high, it is not rare; not we are confident at this is not a coincidence, perhaps coincidentally, perhaps not, but we can not determine. F-measure and t value is the value of statistical test, the probability distribution corresponding to the F distribution and t distribution. Statistical significance (SIG) is the probability of this sample results. 2, statistical significance (P value and SIG value) The results of the statistical significance is the result of the true extent (to represent overall) an estimation method. Professional, P value as a result of the decline index, P value is greater, the more we can not think of variables in the related sample is a reliable index for each variable in the overall association. The p value is the observed result that error probability that is representative of the overall effective. If p=0.05 suggested variables associated with 5% of the samples may be due to accidental causes. That was not related to any variables in gene