Ecotrix plz help

[Neha Gupta]

[happy]

We need to test beta coefficient is statistically different from 1.

H0 : B = 1
H1 : B > 1
Test Statistic :  t = (B hat -1)/s.e.(B hat) = 0.8214

T calculated is less than t crit . So we do not reject null hypotheses.

Conclusion IBM is not volatile security

[Neha Gupta]

Got this and this is surely correct..but i also wanna ask whether F stat may also be used instead? Then Fstat is 211.896> 4.61, so reject the null according to that?

[happy]

No we can't use F Statistics given in the question, because by default statistical package calculate F statistics under null hypothesis that slope coefficient is equal to 0. But we want to test it is statistically different from 1.

This doesn't mean we can't use F test, we can use f test but we need to transform the model.

rit = a + 1*rmt + u

rit - rmt = a + u

r* = a + u  (Restricted model)

calculate RSS for the restricted model and unrestricted model, and then calculate F statistics.

But we can't do this with given amount of information.

[Neha Gupta]

okk, thankyou happy:)