Experimental planning

Back to RDM Home Page 
Back to Data collection

1. Clarify the need for an experiment

Before collecting any data, ask:

• What is the purpose of this experiment?
• Is the information already available in the literature, existing datasets, or models?
• Can theoretical calculations or simulations provide sufficient insight?
• Do you have the resources—time, budget, equipment, and expertise—to conduct the experiment?
• Are there ethical, safety, or technical limitations?

Clearly defining the why helps ensure the experiment is necessary and feasible.

2. Define the research objectives

Determine what you aim to discover, test, or demonstrate.

• List your main research questions or hypotheses.
• Rank them by importance and feasibility.
• Decide whether the study should be exploratory, confirmatory, or comparative.
• Avoid focusing too narrowly on details at the expense of broader goals.

3. Identify variables

Experiments involve inputs (independent variables) and outcomes (dependent variables):

• Independent variables: what you change (e.g., temperature, time, dosage).
• Dependent variables: what you measure in response (e.g., growth, signal strength, error rate).

Before beginning:

• Decide which variables are critical.
• Determine the appropriate range or levels to test.
• Ensure tools or techniques are available and properly calibrated to measure them.
• Evaluate if the expected precision is sufficient to answer your research question.

4. Design the experiment

To get the most information with the least effort:

• Choose a study design (e.g., factorial, randomized block, response surface).
• Limit the number of variables tested simultaneously unless using a multivariable design.
• Consider replicates to account for variability and improve confidence.
• Plan how to document and organize results to avoid data loss or misinterpretation.

Three key questions:

• What types of measurement errors should you minimize?
• What is the minimum number of experiments needed?
• When and how often should trials be repeated?

References & Suggested Reading

You can replace general advice with references grounded in good practice. Here are solid, discipline-neutral sources:

• ELIXIR Belgium RDM Guide – Experimental Design section

Back to RDM Home Page 
Back to Data collection