Experimental design: Recommendations
Give advice on experimental design to the user in person (recommended) and as written guidance. The users should receive training how to design the experiment properly. (Online training can be also found at NIH [1] [2] [3], Johns Hopkins University [4] [5], Harvard University [6] [7] [8] or at the Center for Open Science [9] [10] [11] [12]).
The CF/PI/specialist should check the experimental design before proceeding further to ensure the results will be meaningful and to avoid a waste of money and resources. Many CFs find it important of being involved in the project from the very beginning. Study design flaws are a key source of irreproducibility [13].
Encourage the users to develop and adhere to a rigorous study plan before starting experiments, including the following recommendations:
- Hypothesis is defined before starting experiment
- Methods are appropriate
- Experimental conditions are defined
- Samples size is defined
- Relevant controls are included
- Experiment is replicated
- Samples are randomized and blinded
- Data inclusion/exclusion criteria are pre-defined
- Data analysis and statistical analysis of the results are pre-defined.
