Hi all.
I want you guys to meet my special friend. He is a very generous with his knowledge. I came to know him before i decided to further my studies. He explained the research terminology in a very 'everyday talk' sort of speak. Very handsome fella. "Sudahlah handsome, bijak pula tu!".
His name is Othman Talib.
His famous nickname is drotspss.
----
Oskies..enough with the introduction. Now, back to our discussion about sample size. In class, Dr. Aziah have talked about Krejcie and Morgan. And during the QA session at INTAN, this is the only reference that we all coined in out mock RP. So... while facebooking, I stumbled across his entry. Which is awesome because I always wonder... there must be other researcher who must have came up with something else for calculating sample size since 1970's, right?
Actually I have found it. Yaynesss! Here is the part of the entry by Dr. Othman, talking about sample size. Please note that I've pasted the original form (Malay) because i do not want to misinterpret it somehow.
- Jadual Krejcie and Morgan (1970) jika anda tahu populasi anda (target ataupun accessible). Pengiraan dalam jadual ini tidak mengambilkiran statistical power dan effect size dan saya berpendapat sesuai jika anda ingin membuat kajian survey (tak perlukan effect size)….dan populasi anda besar serta diketahui bilangannya.
- Jadual Cochran (1977) yg berasaskan formula margin of error bagi membolehkan penyelidik menerima error item seperti 5% bagi data kategori dan 3% margin error bagi data continuous. Ini adalah error yang sengaja diadakan utk diterima oleh penyelidik kerana setiap pengukuran mesti membawa error. Ini termasuklah menerima error dengan α = .05 (Type 1 error). Oleh itu Cochran telah menyediakan dua formula iaitu bagi data continuos dan data categorical. Anda tidak perlu tahu saiz populasi apabila menggunakan formula Cochran.
- Jadual Cohen (1988) berasaskan statistical power dan effect size (ES). Katakan anda membuat ujian-t kumpulan tidak bersandar, maka anda tetapkan kesan saiz yang (mengikut Cohen) katakan bernilai d = .70. maka sampel anda adalah sebanyak 33 orang bagi setiap kumpulan. Ini bermakna, kajian eksperimen memerlukan anda menganggar nilai ES sebelum kajian bermula. Jika anda yakin dapat mengumpulkan lebih ramai sample, anggarkan sahaja ES yang lebih sederhana, katakan 0.6, maka sampel anda bagi satu kiumpulan adalah 45. Ini dinamakan prior to research sampling method.
----
Please download the documents by clicking:
Credit to : Niteshkewat
----
Please download the documents by clicking:
----
So, the next time you think of sample size, you can opt for something that is not Krejcie and Morgan. Yay for Dr. Othman! Thank you for sharing this precious info.Anywhooo.
And then, I do myself some googling. Here are some of theinfo that might be useful in my future research:
- Chua (2006) states that most studies are conducted using Cohen’s statistical power analysis as the guideline for estimating the desired sample size. The sample size calculated using the Cohen’s Statistical Power Analysis would be more meaningful and acceptable. Chua also mark that Krejcie and Morgan’s sample size would be too large a number. Conducting a study, which involved too many subjects than what is actually necessary would mean that valuable time and resources were not efficiently utilized. Source: Jurnal Penyelidikan IPBL.
- According to Cappelerri and Darlington (1994), Cohen’s Statistical Power Analysis is one of the most popular approaches in calculating sample size especially in the behavioural science. According to his method, (unlike the straight forward approach by Krejcie and Morgan), there are few factors that were put into considerations: significance level, effect size, desired power, estimated variance and sample size.
- The larger the sample size, the more precise the result would be. (Note: As you can see, this statement contradicts with the Chua's standings.) Source: PRC.
- In order to determine an adequate sample size, the value of significance level, effect size, power and estimated variance have to be pre-determined.
Till next time.
Toodles!
:D
No comments:
Post a Comment