Quick: What’s the formula for Standard Error?
b/t
ha, what’s b/t? coefficient/t-value
se= 1/root of number of observations
Standard error for what exactly?
You said quick dude. No time to spell it out for you. But yes, that’s what I meant.
Swet that’s the standard error for the autocorelation of residuals. The standard error for multiple regressions is some ridiculous formula. They usually give you the coefficient and either T, or the coefficient and SE and you find T
standard error of the estimate, that’s what I was looking for originally because I forgot the formula, and was reviewing quant what is the formula for SEE? something like square root of MSE, whcih is SSE/n-2 is that right? been a long time
b/t is the standard error of the coefficient. Standard error of estimate (SEE) = (unexplained variation/(n-2))^(1/2). Unexplained variation is MSE which is SSE/(n-k-1) from Anova table. This is used to explain how well a regression describes a relationship between x and y.
So to summarize?.. For autocorrelation of residuals = SE = 1/sq root of number of observations For standard error of multiple regressions is usually given or can be calculated as b/t and there is also some really complicated formula that is not worth our time at this point
Most of the time they will give you T, Because if they give the standard error and you F that up you cant answer questions later on about multicollinearity.
Just to add
t-statistic of Auto-correlation is = Autocorrelation/ SE
From the dead it rose!!
4 years, dude. That’s not very quick…
we dont need to know how to calcualte Std Error