An ethical canvas for online experiments

Data

Ownership & Rights:

Transparency about data is necessary.
The owner of the data must have approval to use the data.

Privacy & Security

The privacy and anonymity of users must be guaranteed.
No more data may be collected than necessary.

Transparency

People must be able to easily identify the purposes for which their data is being collected.
People must have the opportunity to be excluded from online experiments.

Limitations

It must be clear what the characteristics and shortcomings of the data are.
The characteristics and shortcomings of the data must be known to everyone who works with it and must be included in reliability calculations.

Experiments

Tools & Statistics

The experimentation tools and statistics must guarantee reliable results.
A quality program must be in place that clarifies, analyzes and improves possible limitations.

Experimentation skilIs

To guarantee reliable analysis and conclusions. The people who perform the experiments must have the right skills and knowledge of experimental design, data analysis and statistics.

Cost-benefit

A go/no go decision must be made prior to each experiment by weighing the potential benefits against the potential disadvantages.

Welfare measurements

An ‘evaluation criteria’ for human well-being must be defined.
The experiments must be able to be analyzed on the basis of their impact on these evaluation criteria.

Application

Behavioral knowledge

The people conducting the experiments must have enough experience and knowledge of behavioral sciences to analyze and draw conclusions about the possible pros and cons for human well-being.

Extensive analysis

The possible effects on human well-being must always be thoroughly analyzed and evaluated.

Minimizing disadvantages

The evaluation criteria for the well-being of people must outweigh the criteria aimed at the well-being of the company.

Communication

People must be able to rely on or submit a request to make changes to the company’s service.