I am in my master's thesis and I am seeking to look at stress levels on individuals who consume red and black rice. Is it enough to observe 10 individuals? What statistical method should I use?
Your group should comprise more individuals to be considered normally distributed (e.g. n = 60). Start with testing your individuals' baseline of stress (stress test online: Holmes and Rahe)
15 persons could first eat red rice (2 weeks) and then 2 weeks black rice. Then then an other group eats first black rice and after that red rice, Then you find out if the order the rice impacts their stress level. The you have one group (15 persons) that eats the rice they usually eat for 4 weeks and another group (15 persons) that eats no rice at all.
I would suggest estimating the effect size you might expect (probably small) and looking at a set of power analysis tables to tell you what size groups you need to have a chance to get a significant effect. I am afraid my guess would be that you would want hundreds of people in each group. Certainly not 15. The mechanism by which rice might impact stress is also important - if the mechanism is slow you need long time periods. You also need careful matching of the groups. Whilst the Japanese diet for example shows a possible impact on stress we don't know if this is independent of other factors within the Japanese society and peoples. This is a difficult question to address without significant resourcing of the research.
You should try to increase your sample size(>.40 n ) to get results than have even a modest probability of replicating.If you suspect that your independent variable has very HIGH EFFICACY you can use nonparametric statistical tests (Siegel,1956) in a pilot exploratory study, to determine the empirical boundaries of your research domain.
- Badges
- Science topic
- Similar topics
- Biological Science
- Physiology
- Physiological Processes
- Stress