original sample estimate

Frequently Asked Questions about the SmartPLS software. Please refer to this section first, if you have problems with the software.
Post Reply
andersonjanuarwijaya
PLS Junior User
Posts: 3
Joined: Mon Dec 05, 2011 3:34 pm
Real name and title:

original sample estimate

Post by andersonjanuarwijaya » Tue Jan 31, 2012 2:41 am

can i ask what is the definition of original sample estimate in bootstrapping output? can we tell the most effecting variable from the largest original sample estimate value? please give me some answer.. thanks a lot..

User avatar
Hengkov
PLS Super-Expert
Posts: 1618
Joined: Sun Apr 24, 2011 10:13 am
Real name and title: Hengky Latan
Location: AMQ, Indonesia
Contact:

Post by Hengkov » Tue Jan 31, 2012 2:49 am

Hi, Januar
Prosedur bootstrapping membutuhkan resampling yang besar, minimum standar 100, tapi beberapa literature menyarankan > 200, bahkan bisa sampai 5.000. Tapi tidak ada beda jika kita menggunakan resampling < 100 hasilnya tetap sama.
Silahkan lihat paper Hair et al 2011 dan Hair et al. 2012.
Regards,
Hengky

andersonjanuarwijaya
PLS Junior User
Posts: 3
Joined: Mon Dec 05, 2011 3:34 pm
Real name and title:

Post by andersonjanuarwijaya » Tue Jan 31, 2012 2:54 am

sebelumnya terima kasih sudah mau membalas post saya, tapi maksud saya, apa kita bisa melihat adanya pengaruh variabel A ke B dari nilai original sample estimate hasil keluaran bootstrapping? soalnya setelah saya lihat di salah satu forum, ada yang menyatakan bahwa nilai original sampel estimate yang paling besar berarti bahwa variabel A tersebut memiliki pengaruh yang lebih besar terhadap B dibanding variabel lain..
thanks,
Anderson

User avatar
Hengkov
PLS Super-Expert
Posts: 1618
Joined: Sun Apr 24, 2011 10:13 am
Real name and title: Hengky Latan
Location: AMQ, Indonesia
Contact:

Post by Hengkov » Tue Jan 31, 2012 3:05 am

Hi, Januar,
untuk melihat pengaruh antar variabel, lakukan prosedur bootstrapping dan lihat nilai T statistic-nya saja, jika T > 1.96, berarti variabel A berpengaruh signifikan terhadap variabel B. Untuk melihat GoF lihat pada R square dan AVE.
Regards,
Hengky

andersonjanuarwijaya
PLS Junior User
Posts: 3
Joined: Mon Dec 05, 2011 3:34 pm
Real name and title:

Post by andersonjanuarwijaya » Tue Jan 31, 2012 3:16 am

ah, berarti forum yang saya lihat itu salah? misalkan saya punya beberapa variabel (A, B, C, D) dihubungkan terhadap E, di sini saya ingin mengetahui variabel mana yang paling berpengaruh terhadap E. apakah ada cara untuk melihatnya? kalau yang saya tangkap dari forum tersebut, untuk melihat yang saya tanyakan di atas, mereka menggunakan nilai original sample estimate. apa benar begitu?
thanks,
Anderson

User avatar
Hengkov
PLS Super-Expert
Posts: 1618
Joined: Sun Apr 24, 2011 10:13 am
Real name and title: Hengky Latan
Location: AMQ, Indonesia
Contact:

Post by Hengkov » Tue Jan 31, 2012 5:42 am

Hi Januar,
Itu mudah saja, kamu tinggal lihat saja nilai T nya, misalkan lebih signifikan P 0,01 = T 2.58. Maka jika ada variabel yang mempunyai nilai T > 2.58 maka variabel tsb mempunyai pengaruh yang lebih besar dari variabel eksogen lainnya terhadap variabel endogen.

Tapi sesungguhnya untuk melihat besarnya pengaruh variabel eksogen terhadap endogen dilihat dari nilai R-Square dan Q Predictive Relevance.
Regards,
Hengky

jmbecker
SmartPLS Developer
Posts: 1018
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Post by jmbecker » Wed Feb 08, 2012 10:44 pm

Hi Folks,
it would be nice if you could keep the language of your postings in this forum to English so that everyone can follow your problems and advice.

yosat
PLS Junior User
Posts: 2
Joined: Sat Jun 05, 2010 8:42 am
Real name and title:

Bootstrapping

Post by yosat » Sat Apr 21, 2012 11:58 am

Dear pak Hengky
I have start learning SmartPLS using for small group. case=21, samples set to 100 in Bootstrapping
If I run Bootsrtapping using setting no sign changes. I got bad t, below 1,96.

But, if I set the setting = individual changes. Result are good.

Your comments will be appreciated very much.

Regards

JosAndyT


Hengkov wrote:Hi, Januar
Prosedur bootstrapping membutuhkan resampling yang besar, minimum standar 100, tapi beberapa literature menyarankan > 200, bahkan bisa sampai 5.000. Tapi tidak ada beda jika kita menggunakan resampling < 100 hasilnya tetap sama.
Silahkan lihat paper Hair et al 2011 dan Hair et al. 2012.
Regards,
Hengky

User avatar
Hengkov
PLS Super-Expert
Posts: 1618
Joined: Sun Apr 24, 2011 10:13 am
Real name and title: Hengky Latan
Location: AMQ, Indonesia
Contact:

Post by Hengkov » Mon Apr 23, 2012 2:15 am

Dear, J.M. Becker.
Sorry hehehehe...... ;-)

Hi Jos Andy
Standard Resampling => Bootstrap: No Sign change. This option may be very conservative as it may yield very high standard errors and, consequently, low t-ratios.
Another options => Individual sign changes and Construct level changes: For individual sign changes, the signs in each resample are made consistent with the signs in the original sample without ensuring a global coherence and Construct level changes for mode B.

yosat
PLS Junior User
Posts: 2
Joined: Sat Jun 05, 2010 8:42 am
Real name and title:

TQ

Post by yosat » Mon Apr 23, 2012 7:49 am

Dear pak Hengky

NOTED

Thank You for the prompt reply.

User avatar
Hengkov
PLS Super-Expert
Posts: 1618
Joined: Sun Apr 24, 2011 10:13 am
Real name and title: Hengky Latan
Location: AMQ, Indonesia
Contact:

Post by Hengkov » Thu Apr 26, 2012 6:48 am

Ok, your welcome.
Regards,
Hengky

brainwalk
PLS Junior User
Posts: 1
Joined: Thu May 24, 2012 5:17 am
Real name and title:

Post by brainwalk » Thu May 24, 2012 11:42 am

hi pak

apakah teman2 mempunyai kasus data real pada PLS?
terima kasih pak
saya sangat membutuhkan

need real data cases

thanks
brainwalk

lutfifbrn
PLS Junior User
Posts: 1
Joined: Fri Jul 26, 2019 8:38 am
Real name and title: Lutfi Febrian Wibawa

Re: original sample estimate

Post by lutfifbrn » Fri Jul 26, 2019 8:43 am

hai, how to know the correlation between independent and dependent variable
Like use SPSS, can we know the correlation using R2 ?

Thank you

Post Reply