I am currently facing a dilemma with the samples of a pilot study, and would appreciate your help and suggestions.
Originally, there were 39 students - in the total population (100%). Then the administrators changed their decision to allow only 18 students to participate. Now that I have resumed the study, I'm left with 10! The rest didn't attend due to reasons like no transport etc. The time, the venue and the classes were chosen by the administrators.
Will my results be valid? Should I find another school?
Dear Fatimah,
My stats lecturer always insisted that we use sample sizes above 35. For smaller samples you can use non-parametric stats, like Mann Whitney to replace t-test. But when we discuss, we don't claim too much for our research. Put in some qualitative discussions, like conversations/interviews. And put the constraint as a suggestion for future studies. But if you can, it may be better to get another school.
Dear Fatimah,
My stats lecturer always insisted that we use sample sizes above 35. For smaller samples you can use non-parametric stats, like Mann Whitney to replace t-test. But when we discuss, we don't claim too much for our research. Put in some qualitative discussions, like conversations/interviews. And put the constraint as a suggestion for future studies. But if you can, it may be better to get another school.
For statistical validity of your findings, you can use the bootstrap method in conjunction with statistical tests indicated @Miranda.
Thank you..sometimes it does feel like the end of the world! thanks Dr. Yeoh and Dr. Sylantyev Sergiy
Fatimah
This is what is called Mortality in research. Replacing the subjects is not the solution. You have to content with the results because they give only a partial information. That is the limitation of the study, and you should report as such. No generalization is possible with such a result.
Dear Fatimah, don't feel too bad about it. Because schools are not always so considerate to researchers like us. Let's think like Patrick: "Moving on, moving ooooon..."
Thanks Prof N. Woww... Im still so new in this. Dr. Yeoh, 150% agree even though they are our friends... yes, as Prof. Low's motto 'moving on'..
Dear Fatimah, when People dislike our presence, we should stay far far away.
Did you check the newer video recording I sent to the other thread of yours? It's corrected, I think. (Mustn't be too sure, we are prone to error.)
Dear Fatimah and Miranda
Don’t be too worried about the way people behave, as we have little control on that. Let us be clear about our goals and what we want to achieve in life. Continue with that irrespective of what happens. Help as much as possible.
Dear Fatimah
Accepting a difficult time is beginning to overcome it .Do not give up despite the difficulties that all of us face and will face to do a research
Thanks Prof. Elias, and thanks again Prof. N.
Dr. Yeoh, I'll download it later - hopefully its in this time. You could be a master of ceremony with that voice ;) ;)
@Fatimah, what is the purpose of your educational research, what is hypothesis and methods that you apply?
Anyway, you should report what happened with your samples, as @Kamal have mentioned!
Mortality in research is not so rare! @Nageswara is right!The problem of "representativeness" is the main issue!
Dear @Fatimah, maybe this UNESCO paper will help You! Good 90 pages for reading!
I think that you have to insist in N>30 size of your sample, due to the normal distribution approach.
Thank you Prof. Kamal, Prof. Ljubomir and Demetris. Its quantitative for the questionnaire, whether the intervention improves their interest / motivation towards integrated lessons with technology; and qualitatively to see whether their answers in the post test are better than before after the intervention.
You have an interesting dilemma. You state that the research is a pilot study. But any way you look at it, a sample of ten is too few if you want to pursue quantitative research. My suggestion would be to focus on qualitative issues. Use your informants to work out micro-level individual questions about your substantive questions. If in future you gain access to a larger sample the information gained in this type of exploratory project should prove very helpful.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics