Bootstrapping cases & samples

ksh5046 · Post by **ksh5046** » Mon Feb 17, 2014 2:44 pm

Hello

I am a beginner in smartpls.
for getting path coefficient i tried to do bootstrapping but
i had to fill up the parameters 'cases' and 'samples'.
Im very not sure how much should i give...

increasing samples seem to make it stable but, increasing cases seems to increase t statistics.
i have timeseries panel data about banks(yearly, appx.2000~2012)
in this case, should i put 13 for bootstrapping cases? or total no. of banks? or another?

Thank you for consideration.
Regards.

Hengkov · Post by **Hengkov** » Mon Feb 17, 2014 10:17 pm

Hi,

Put total bank in time-series data. Example, number of bank = 20. Total bank = 13*20 = 260 (you put this vaue in case bootstrapping).

Regards,

ksh5046 · Post by **ksh5046** » Tue Feb 18, 2014 2:22 am

Thanks a lot

it would be greatly helpful.

Regards,

kerim · Post by **kerim** » Mon Mar 03, 2014 1:13 pm

Hello,
does mean value of case in bootstarap procedure is equal to sample size? XWhen I have a sample of 106 observations, the bootstrapping case should be programmed for a value of 106?

Regards

Hengkov · Post by **Hengkov** » Thu Mar 20, 2014 6:42 am

Hi,

Yes.

kerim · Post by **kerim** » Fri Mar 21, 2014 2:05 pm

Thanks very much