in case of a study with 20 rats divided into 4 groups with 5 rats in each group, due to lack of animal at a time we planned to do first two groups in one time and again within short duration we will do another two groups. first group consists of control group and remaining three groups consist of three modes of experiments. when first two groups will be enrolled, the animals for remaining two groups will be mature enough and then we planned to do experiment in remaining two group.
You can expect more variation in your results as not all groups were treated at the same time. It depends on your experiment whether this will be of importance. In reporting you should be (very) clear on how you did the study.
keep in mind that any difference in ambient condition between two successive groups may alter the animal response and then the conclusion you may draw. it would probably be better to repeat the same 4 treatments on the available animals for ex:
1rst period 10 animals: 2 control, 3 T1, 3 T2 and 2 T3
2d period 10 animals: 3 control, 2 T1, 2 T2 and 3 T3
so that at the end you minimise the confounding effect of the period (i.e uncontrolled ambient conditions) over the treatments
dear Maurice Mahiey, thank you sir for your suggestion. sir, is there any documents mentioning what you are suggesting as a better option. if you have any link or document i want to request you to provide that to me.
Sir, C A (kees) Kan,
could you please send me article reporting such procedures? i would be grateful to you.
All methodology has its pros and cons but I suggest a crossover design where each treatment can be alternatively applied to each group on rotation basis with all its limitations
dear Laxman Khanal,
"All other things being equal" is the base of experimental design, meaning that only the factors (treatments) you study must vary. That is why Taye T Ejeta suggests a crossover design, in order to eliminate the effects of individual rat differences. If the experiment did not require to kill the animals you can look at a "latin square" design, for example. If you are using R for your statistical analyses (free and open source), you can found many packages and tutorials on the web to build experiment plans and go further
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology