24 December 2013

Confirmatory Factor Analysis SmartPLS


Confirmatory Factor Analysis (CFA) is a method used to get value from several factors that affect data and whether the data itself is approriate to do next analysis. Many softwares can be used to perform CFA. But, if your data are less than 200, then it's better to use SmartPLS. If more, you can use SPSS or AMOS. Usually, many research used more than one software to perfom quantitave analysis.

This time, I want to make tutorial on hoq to perform CFA using SmartPLS. It is a free software which can be downloaded from its website. The free version is only the 2.0 version. The 3rd version is not free because you have to buy the license, even for the students the functionality are limited. 

We start with creating a new project

Click FILE - NEW - CREATE NEW PROJECT


You will see a dialog box to input your project name ... then click NEXT


Choose your project file. The type file that associate with SmartPLS is only .csv ... So, prepare the .csv file. If you want, you can convert your data from SPSS to .csv file... Then, click OK.


You will see in the PROJECTS area your new project which have two things... .splsm and .csv file... .csv file is your raw data which need to be analyzed...


and .splsm is the file which you can use for making your data model and do analysis...


Look at the menu in the top of the window.... Tere is a blue circle with + in the corner... It is used for insertion mode to build several factors in your model.


Click on the .splsm window and there will be several circles which you can use for factors in your model...


You also can rename the circle...just do the right click - rename variables or press F2 on your keyboard...



Let's see in the left of the window...there are several indicators from your raw data...drag your indicators to its factor variables....


You can change the position of the indicators with right click on the variable and click Align Top, Bottom, Left, or Right


Then, how to make the relationship? look beside the circle with + icon... there is a two circle icon which related each other using a line... click that icon...


And make relationship between variables... the orange circle will be changed to blue color...and ready to be analyzed...


For the first analysis... click on the PLS Algorithm in the menu....


Then you will see a dialog box...just click Finish....


Then you can see the result... in the relationship line, you can see there are several values. That values represent the factor loadings of the data...


If you want to see the full report, click on the report on the menu....


You will see a new window open that represent the result...but how to read it?


To see the factor loadings, click on the PLS- Calculatin Result - Outer Loadings
There will be several correlation and you can see the value of each factors. Sometimes, the value need to be more than 0.7 or 0.65... if you have an item that the value is less than 0.7 or 0.65, you need to eliminate it from your model for further analysis because it's not significant enough.


Then, you can also see the AVE value, composite reliability, R square, cronbach alpha, and other... in the Overview...



So..you can see all of the things that need to be seen in the CFA... maybe yes it's not easy to see...you need to move those values to excel and move it to appropriate tables than everyone can see your result easily..

Hope this tutorial help you and if you have any questions, feel free to ask me...i'll be happy to help you.. ^^

Intan Web Developer

A Wife and PhD candidate to-be in National Taiwan University of Science and Technology. Dreamer, Writer, Traveller, and Tech Addict. Like to travel everywhere and experience anything.

2 comments:

  1. Wonderful web journal. I appreciated perusing your articles. This is really an awesome read for me. I have bookmarked it and I am anticipating perusing new articles. Keep doing awesome!

    confirmatory data analysis

    ReplyDelete
  2. Great article and explanation. But using such a pale color as the font color was really not the best.

    ReplyDelete

Anda bisa memasukkan komentar tentang postingan di sini...Terima Kasih ^.^