f test result interpretation

[h21]

so if anova table show us an F result, how do we use that to understand the significance of the result? I am talking about reading 9 example 17,anyone help?

 

[tickersu]

If the calculated F is beyond the critical value (or a p-value less than the preselected alpha) this indicates statistical significance at the corresponsing significance level for the hypothesis test of Ho: b1=b2-=…=bn=0 vs. Ha: at least 1 bi not equal to zero (in short, reject Ho in favor of Ha). This indicates that the model has statistical utility and should be evaluated practically (via [adjusted] R-Squared, SER, and some other metrics) to determine if it’s worth using.
For the test, just know how to determine significance, how to set up the null and alternative hypotheses, and that a signficant F-test indicates these variables are significant as a group, and at least one of the slopes is nonzero (at least one of the independent variables is a statistically significant predictor of the dependent variable).

 

[krokodilizm]

True but in this case I think the F test just carries out a decorative function because there is no p-value to test its significance, even though its magnitude is in the hundreds.

 

[tickersu]

krokodilizm wrote:
True but in this case I think the F test just carries out a decorative function because there is no p-value to test its significance, even though its magnitude is in the hundreds.

In general, if it’s that large, it’s probably significant at most practical levels of significance and degrees of freedom (but you’re right, they would need to give you at least a snip from an F-table or a p-value and significance level). From that table alone, they’re more likely to ask you to calculate the F-stat or to calculate a missing value in the table from manipulating the F-statistic.