For finding significance of parameter estimates in smartpls , should the t-value presented in the ouput report of smartpls be taken (after bootstrapping is done ) . OR by bootstrapping procedure as mentioned Shrout and Bolger (2002) i..e if bootstrap is run for 1000 samples , then after putting the sample values in ascending order and taking 25th and 975th element as the confidence interval and if the mean of the bootsratp sample lies in it, then it is considered significant at 95% confidence level
Which method is appropriate to see the significance of the direct paths i..e estimates of the direct paths.
bootstrapping-significance-
-
Diogenes
- PLS Super-Expert
- Posts: 899
- Joined: Sat Oct 15, 2005 5:13 pm
- Real name and title:
- Location: São Paulo - BRAZIL
- Contact:
Hi,
They are two different ways to use the bootstrap procedure, see:
Good, P. I. Introduction To Statistics Through Resampling Methods And Microsoft Office Excel. New Jersey: John Wiley & Sons, 2005.
Percentile Bootstrap (no assumption)
Parametric Bootstrap (with assumption of normality of the bootstrap results).
If you want to test the assumption of normality of the bootstrap results, just copy/paste the results (usually 1000 repetitions by path) to a statistical package and try a histogram or Kolmogorov-Smirnov test.
Personally, I have used the t-values from SmartPLS (parametric bootstrap).
Best regards,
Bido
They are two different ways to use the bootstrap procedure, see:
Good, P. I. Introduction To Statistics Through Resampling Methods And Microsoft Office Excel. New Jersey: John Wiley & Sons, 2005.
Percentile Bootstrap (no assumption)
Parametric Bootstrap (with assumption of normality of the bootstrap results).
If you want to test the assumption of normality of the bootstrap results, just copy/paste the results (usually 1000 repetitions by path) to a statistical package and try a histogram or Kolmogorov-Smirnov test.
Personally, I have used the t-values from SmartPLS (parametric bootstrap).
Best regards,
Bido
if there's no value on my t-statistic
i do bootstrapping to my inner model. but the result like this:
O M STDEV STERR T Statistics (|O/STERR|)
-1,000000 -1,000000 0,000000
is it significance? my R2 is 1 (100%). is there multocolinearity? should i use principle component analysis (PCA) to eliminate the multicolinearity?
O M STDEV STERR T Statistics (|O/STERR|)
-1,000000 -1,000000 0,000000
is it significance? my R2 is 1 (100%). is there multocolinearity? should i use principle component analysis (PCA) to eliminate the multicolinearity?
I need PLS to analyze my cases
-
Hengkov
- PLS Super-Expert
- Posts: 1599
- Joined: Sun Apr 24, 2011 10:13 am
- Real name and title: Hengky Latan
- Location: AMQ, Indonesia
- Contact:
Hi,
R-square = 1 and T-statistic = 0.000 indicate you have same indicators/same value in excel sheet for exogenous to endogenous, check your model and also data set again!
For references see My books in www.pls-pm.blogspot.com.
Best Regards,
Hengky
R-square = 1 and T-statistic = 0.000 indicate you have same indicators/same value in excel sheet for exogenous to endogenous, check your model and also data set again!
For references see My books in www.pls-pm.blogspot.com.
Best Regards,
Hengky