Design your experiments

design your experiments

Experimental Design

Experimental Design Assistant

The following article aims at giving you an overview over useful strategies and tools for designing your experiments.

If you have additional questions or wish to book a personal meeting with our staff to discuss the design of your experiments, feel free to write an email to UUBF.

If you are planning to conduct experiments in UUBF's facilities, please make sure to carefully read through the important information section on this website.

Experimental Design

Before you start your experiments, it is advisable, that you spend a sufficient amount of time designing your experiments. A good study design will help to reduce errors as well as being easier to replicate and yielding more consistent results. In addition to this, a well-designed experiment may reduce the number of animals needed to achieve meaningful results. For more information on how to optimize the use of animals in your experiments, please consult our page regarding the 3Rs.

The information provided on this website is not meant to be a comprehensive summary of all steps required for planning and designing experiments, but is rather aiming at helping you consider some vital points of a study during the design process. Please refer to the PREPARE guidelines for planning animal research and testing when designing and planning experiments.

1. Clearly define your hypothesis

A clear hypothesis will help you in the following steps of your study design and make it clearer, what experiments you will need to conduct.

2. Choose which model organism to use for your experiments

The choice of your model organism will greatly influence the design of your study. If you are unsure about the choice, feel free to write an email to UUBF for a consultation.

3. Choose an appropriate test for your hypothesis

Once you have decided on a (or several different) model organisms to work with for your study, it is important to choose the right test(s) to answer your research question. UUBF offers a variety of different tests for mice, rats and fish. You can check which types of tests are available at UUBF in our lists of available tests. If you cannot find the test you were looking for in these descriptions, feel free to write an email to UUBF and describe what type(s) of test you are looking for.

4. Determine the groups needed for the experiment

In this step, you will have to decide what type of groups you will need. This choice depends largely on your hypothesis. Keep in mind, however, that the number of groups will directly effect your following choices such as statistical tests, sample size and data analysis routines. It might be a good idea to return to this step if you realize in the following steps that for example your required sample size will be to large.

5. Make a draft for data analysis and statistical tests

If you followed the previous steps, you will likely already have a good grasp on the type of data you will gain from your experiments, as well as the parameters you want to analyse. Based on this information, the needed steps of data preprocessing as well as statistical analysis will be apparent. By drafting a plan for data analysis and statistical testing already before starting your experiments, you can make sure that you will
a) gather exactly the data you will need to perform appropriate testing
b) streamline your data analysis and
c) avoid compromising your statistical testing, as you are only performing one, appropriate test.

6. Determine the sample size needed to achieve meaningful results

There are several ways to plan the sample size you will need for your experiments. We recommend you to perform a power analysis based on the groups you have determined as well as the proposed statistical test for your hypothesis. You can find a list of online tests as well as the corresponding R code to implement the test yourself in on the following webpage

Power and sample size tests

If the power analysis results in a very large sample size, you now have the opportunity to adjust your experimental setup or groups.

It can be preferential to conduct a pilot experiment in order to determine your needed sample size (please notice that you will need to have ethical approval also for conducting pilot experiments, see paragraph 8).

For more information regarding power and sample size, please consult the NC3R article on the topic.

7. Contact veterinarian and facility staff

Once you have a clearly designed plan for your experiments, it is important to discuss the planned procedures with the responsible veterinarian as well as the staff of the facility in which you will be conducting your research. This gives you the opportunity to improve your design, find pitfalls and prepare for your ethical application

8. Write up experimental plan and send in ethical application

Once you have discussed and, if necessary, revised your experimental plan, it is time to write up the final version of your planned experiments as well as sending in an ethical application for your planned experiments.

Feel free to write an email to UUBF at any stage in your design process for feedback and discussions.

Designing experiments using the Experimental Design Assistant

The British National Centre for 3R (NC3R) provides a valuable tool, the Experimental Design Assistant (EDA), which can be used to design your experimental procedure. The tool allows you to create an experimental scheme for your experiments including the creation of groups and consideration of interfering variables (such as the time of day, sequence in which treatments are performed etc.). You can furthermore enter your independent variables and measured values as well as data transformations. Based on this information, you can then use the EDA to get suggestions on possible statistical tests suitable for your experimental design.

After entering all appropriate information in EDA, you can generate a report in which suggestions for improvements of your experimental design will be given.

When starting to use EDA, you will have the opportunity of either using it as a simple sketching tool for experimental processes or more in-depth designing. Getting used to some of the more in-depth aspects of the tool will, however, require some time. In addition to this, the quality of the suggestions you will get will largely depend on the amount of information you provide in the tool.


The backbone of all Experiments

A well-designed experiment is more likely to yield meaningful results while minimising the risk of unforeseen errors.

From design to Result

Put your plans in practice using UUBF's facilities, equipment and expertise.

gain insights

Get them most out of your data by optimising your data analysis.

Report your findings

Make sure you provide the most value to the scientific community by publishing your results in a clear and concice way.

Last modified: 2023-10-23