Question

Experimental Design and Statistics Question

If our dataset included 100 possible explanatory variables, we would expect to see statistical significance for 5 comparisons, by chance alone. (That is 5%, after all, and we have set our alpha value at 5%!) Given this, explain why it is important to clearly state your null hypothesis, plan your study design, and draw your predicted results before you conduct your experiment.

Answer 1

It is important to clearly state your null hypothesis(H0), plan your study design, and draw your predicted results before you conduct your experiment because how can you conduct your experiment without knowing the likely outcomes of the experiment?; how do you know whether to reject the null hypothesis without stating the null hypothesis in the first place?; how do you proceed with your experiment without planning the study design?

To conduct your experiment smoothly and to draw the appropriate conclusion from the results of the experiment, you must state the null hypothesis(H0), plan your study design and draw your predicted results.

Predicted results mean either to reject H0 or not to reject H0 - these are the likely results of the experiment but to find out what the exact result is and how strong the evidence is for the result obtained, you will conduct the experiment by properly planning such as what experiment design is suitable for the given problem, how to go about it, what sample size is required, how many treatment and control groups are required, time period of the experiment, post hoc tests, etc,. - all these need to be planned before conducting the experiment so that results and conclusions drawn will be reliable. Moreover, the process of the experiment goes well if there is a proper plan.