An experiment isn't a set of steps — it's an argument. Before you plate a single cell or weigh a single reagent, answer these five questions in writing.
1. What is the hypothesis? State it as a testable claim, not a curiosity. "Compound X inhibits enzyme Y" is testable. "I want to see what happens" is not.
2. What is the null? If your treatment does nothing, what would you expect to see? If you can't answer this, you can't design a control.
3. What are the controls? Positive, negative, and vehicle. All three. Every time.
4. How many replicates? Biological ≠ technical. Three of each is the polite minimum for anything you plan to publish.
5. What analysis will you run — before you look at the data? Choosing your statistical test after seeing the results is a well-known way to publish work that doesn't replicate.
Answer these five and your experiment is doing scientific work. Skip them and you're just labelling tubes.
— The SciMantra Team
