Sunday, November 1, 2009

Moving to the next phase...

To start the second phase of my research, I need to do the pilot study. Have I done any pilot study before? The proud answer is YES. Next question, Do I really understand why I need to do it? Sincerely , NO.

I believed PhD is a whole journey of learning (that is why it sometimes become a very lonely and boring journey in your life). You can successfully grab your Master degree without really understand what you are doing BUT I don't think you could do the same in your PhD. As for me, things become tougher as I got my Master degree more than 10 years ago. Even if I have any knowledge about doing good research it must have been rusted by now.

Actually, this posting is just to brush up my knowledge. To the readers, if you are in the advanced level of doing research, this posting might not benefit you. Just untuk saya memaham apa yang kononnya saya faham and untuk golongan yang "blurrrrr" cam saya...heeee

What is pilot study? It can refer to 2 things. First it could be a " small scale version" done in the preparation for the actual study; second it could also be the pretesting or trying out of a particular research instrument.

Why do we need pilot study?
Many reasons. In short, it could give advance warning about where the actual research could fail. As for me , my pilot study is more on testing the adequacy of my research instruments.

How to check the adequacy of research instruments?
Must check the validity and reliability.

What is validity?
Refer to a degree to which a test measures what it is supposed to measure. 3 types; content validity, construct validity and criterion -related validity.
Content validity- measures the degree to which the test items represent the domain of the trait being measured. Researcher need to identify the overall content of the construct to be measured. Not an easy task, so usually we end up asking the experts in the particular field to identify or evaluate the content of the test/construct to be developed.

Construct validity- concerns the degree to which the test measures the construct it was designed to measure. 2 parts in evaluating construct validity. First, examine the theory underlying the construct. I believed when you reached this stage, you have successfully done this. If not, you won't be able to get your candidacy approval. Second,evaluate the adequacy of the test to measure the construct. Check the convergent( if convergence exists, construct validity is supported) and discriminant validity ( demonstrate that the construct can be differentiate from other similar construct).Usually, researcher use factor analysis with VARIMAX rotation to get these reading.

Criterion -related validity- concern to detect the presence or absence of one or more criteria considered to represent traits/ constructs.

What is reliability?
Concern with the accuracy of the measurements. Refers to the degree to which measures are free from error and yield consistent result. Usually, we need to find the Cronbach Alpha. Rule of thumb, the alpha should be >0.70. If failed to get this result, then we should identify which item/ statement that caused the problem. Should exclude this item and check again on the alpha. If the reading improve, may be we need to make further investigation on this item. May be something wrong with the "translation process" (as I need to translate my questionnaire). Kalau teruk bebenor reading, may be need to think of doing the second pilot study after making all the necessary adjustments.

What to report in the thesis?
1) Detail explanation on how we carry out our pilot study.
2) Provide results of the pilot study:
- descriptive analysis
- explanation on the validity and reliability

No comments: