Supposed BP test = 1.80 p-value = .649 are there hetroskedasticy at 1% significant? (p -value is the lowest level of significant null can be rejected.) I know if t-stats > critical value then there is heteroskedasticy, but i dont know how to interpret the p-value here. I supposed that given that, .649 > .01 then we canot reject null therefore hetroskedasticy is not signiifcant !?

i think you need to move the deciaml 2 places over… i p value of .649 means 64.9% thus we can not reject the null p value is too high… conclude = 0 = no heteroskedasticty

Null hypothesis for BP test - Heteroskedasticity is not present. Smaller p values for regression < 5% i.e. 0.05, suggests that the null hypothesis can be rejected with 95% confidence. Lower the p-value, better are the chances of rejecting the null hypothesis. So in this case, p-value of 0.649 (high more than 0.01) will mean that we can’t reject the null hypothesis, so heteroskedasticity is present.

damm, am i wrong? i set up the test wrong?

Sorry, Nikko, you are correct, missed “not” in last line. We can’t reject the null hypothesis “Heteroskedassticity is not present”, which means heteroskedasticity is not present. A bit confusing :-(. Null hypothesis for BP test - Heteroskedasticity is not present. Smaller p values for regression < 5% i.e. 0.05, suggests that the null hypothesis can be rejected with 95% confidence. Lower the p-value, better are the chances of rejecting the null hypothesis. So in this case, p-value of 0.649 (high more than 0.01) will mean that we can’t reject the null hypothesis, so heteroskedasticity is “not” present.

yeah, the null hypothesis is opposite to what we generally set up. H0 = is heteroscedastic