Degrees of freedom in bootstrap: n-1 or samples -1

khauschildt · Post by **khauschildt** » Mon Mar 08, 2010 8:10 pm

Dear all,

which is the appropriate number for degrees of freedom when computing p values for t-statistics, e.g., in Excel?

The number of bootstrap samples - 1 (e.g., 499, 999),as here and here, or
the number of data points - 1 as posited in this article on p. 19 (formula) and, e.g. p.27 in the subsript for the t-value.

I am a bit confused and would be grateful for more references. Maybe I have just missed some simple point or misread something.

Thank you!
Kristina

stephan.kramer · Post by **stephan.kramer** » Tue Mar 09, 2010 9:16 am

Hi Christina,

the df's for your t-statistic are the number of bootstrap-samples - 1.

Best regards, Stephan

khauschildt · Post by **khauschildt** » Tue Mar 09, 2010 9:20 am

Hi Stephan,

thanks for your reply! I am tending towards that answer as well, but do you perhaps have a reference for that? Because, in the linked article, the authors explicitly state that the correct number of freedom is n-1 and also use this consistently to report their results. That's what's confusing me.

Thank you!

stephan.kramer · Post by **stephan.kramer** » Tue Mar 09, 2010 3:30 pm

Hi Kristina,

e.g. it is stated at the following article, p. 305-306: "Henseler, J., Ringle, C.M., & Sinkovics, R.R. (2009). The Use of Partial Least Squares Path Modeling in International Marketing. Advances in International Marketing 20."

Best regards,

Stephan

khauschildt · Post by **khauschildt** » Tue Mar 09, 2010 3:37 pm

Perfect! Thank you, I found it and will just cite that then.

:)