Question

How can I interpret a regression statistics table in Excel?

Answer 1

I assume you mean this:

http://cameron.econ.ucdavis.edu/

The "Coefficients" are the slope or y-intercept in this case. "HH SIZE" refers to the Slope, and of course, Intercept is the y-intercept.

If you multiply the Standard Error by `1.96`, you get the Associated Error for either the Intercept or the Slope. The Associated Error is basically the uncertainty you have.

For example, in a standard physics lab course, bare minimum, here's what you would need to know:

• Slope
• Intercept
• Slope Standard Error (`SE_"slope"`)
• Slope Associated Error (`AE_"slope"`)
• Intercept Standard Error (`SE_"int"`)
• Intercept Associated Error (`AE_"int"`)

The sample standard deviation is:

`s = sqrt(1/(N-1) sum_(i=1)^N (x_i - barx)^2)`

where `N` is the number of trials, `x_i` is each individual value, and `barx` is the average of said values.

The Standard Error is:

`SE = s/sqrt(N)`

where `s` is the standard deviation above, and:

`AE = 1.96*SE`

Here is an example of an Ohm's law analysis I did using a similar regression statistics table:

Oftentimes, even in a quantitative analysis course, you only need to further know the coefficient of determination `R^2`. The closer it is to `1`, the better it is, but it is only for a linear fit line.

Other than that, I have not had to use any other quantity on the regression statistics table in my 7 University semesters.