Page 1 of 1

IPMA PERFORMANCE

Posted: Sun Mar 27, 2016 12:03 pm
by juliosolano
Sr.

Yo estoy confundido con el concepto PERFORMANCE en IPMA. Alguien me puede ayudar a comprender.

Gracias.

------------------------
Mr.

   I am confused with the concept PERFORMANCE IPMA. Someone can help me understand with some example ?

Thank you.
Julio

Re: IPMA PERFORMANCE

Posted: Mon Mar 28, 2016 8:57 pm
by janschreier

Re: IPMA PERFORMANCE

Posted: Sun Apr 03, 2016 2:59 pm
by Suhair
Dear Ringle,

I'm a PhD student, doing my research in the field of construction management. I'm using SmartPLS 3.0 for data analysis
my model has HOC reflective-reflective where I decided to use two stage approach because the indicators (items) for each LOC are different in number (i.e LOC1 has 2 indicators, LOC2 has 4 indicators..etc.)
I followed the instructions by the advocates in PLS where I tested my model via two approaches (first draw all possible paths at only low order components, testing the quality of the measurement model (reliability and validity), then using the latent variable scores of LOC as indicators for the HOC) (the second approach is a combination of repeated indicators and two stage approach) both approaches yielded same results

I also decided to use Important-Performance Analysis (IPMA) which I found very relevant for my research which mainly aims at developing performance measurement and improvement framework for small and medium contractors in developing countries

my question is which model I should use to test IPMA? i.e should I use the the fisrt step model (i.e. either the first model which include all LOC in the two stage approach) or the second step model (the model that has latent variables scores in the two stage approach)??

my understanding that IPMA uses latent variable scores to generate the performance index, if I'm going to use the second stage (step) model which alrady has latent variable scores from the first stage (step) then how SmartPLS going to treat this in particular when I set the ranges??

I was thinking of using the first step model which has the original raw data (original liker scale 1-5) am I right?

best wishes
Suhair