Tourism Logistics - โลจิสติกส์สำหรับการท่องเที่ยว

Home เคล็ดลับวิจัย Simulations and SEM Bluebird Perspective: On the paper Trick on Structural Equation Model (LISREL)



Journal EEQEL


Bluebird Perspective: On the paper Trick on Structural Equation Model (LISREL) Print E-mail
By Pairach Piboonrungroj
26  June 2009
I wrote this article for sharing my personal idea on an article titled "Trick on Structural Equation Model (LISREL)" which was just published on this website.  Please consider this article as an exchange of idea and not a fight.

In Structural Equation Model (SEM),  we have six major stages (Hair 2001)as you can see in my paper about methodological implication in TSCM research. Researchers who use SEM are those who try to test their conceptual model with empirical data. The conceptual model should be derived from literature review or exploratory studies. We can not sit in front of the PC/Mac and thinking what the model should be without any reasonable or logical supports. I mean previous research.
Then, after a hard working on conceptualise the model, we design questionnaire and get the data.
What?s next?

Yes, we input the data into any computer packages for SEM analysis. There are LISREL, Amos or EQS. The result will be the same. The choice of computer package is based on own preference and availability whether you have a licence to use them. You may try a student and trial version of LISREL in
So what?
If the result is not what we want, what will you do?
To me, I will check my data if any thing goes wrong.
However, we need to clean the data before running whatever the model is.
The point is that, unlike regression in SPSS, LISREL offer the solutions to improve your model. However, every step that you take, do make sure you have a theoretical support. Never just do it because of the statistical sake!
In the first year of PhD study here in Cardiff, I had to study Quantitative Analysis that is at the most advance just about Multiple regression analysis.
I was thinking ?why I have to study this module as I know them all??.
I found that I almost know nothing.
?You can not just ??.?
The above is the typical comments I?ve from my first coursework. I just make everything works statistically ? but NOT logically or NOT sufficient academically and logically supports and procedure.
I?ve got 70% for this assignment. It made me so upset and disappoint as I think I should done it better. 70% is the distinction level but mark means nothing to me. The comments from the lectures hurt me much.
In the second semester, I put all my effort to make it perfect, at least no comment like ?this is not making sense? or ?you canNOT just test it, include this variable.
I did a lot better but a lot of thing need to be improved.
However, every step I take I know what I was doing.
You may download my assignment in the second semester at the end of this article.

Finally, I am writing this article to say that

?we use computer to assist our calculating competence limitation. But we need to conceptualise, analyse by our self. If computer always right, there is no room for econometricians nor analysts.?

Back to menu



ผู้เขียนไม่หวงห้ามที่ท่านจะคัดลอกบทความ บนเว็บไซต์นี้ไปใช้ในรายงานของท่าน  

แต่ขอความกรุณาเพื่อนนักวิชาการ เพื่อนผู้ทำเว็ปไซต์ 
น้อง ๆ นักเรียน นักศึกษา ทุกท่าน