Research ethics and A/B Testing
A/B testing may provide a limited sense of truth, does not consider unintended consequences, and may create a false assumption of correctness.
An example of this procedure can be found in experiments conducted by Experian. As acknowledged by its former CEO, some researchers/designers A/B tested the effect of images while customers were waiting between a query and the results. They found out that aggressive images were triggering higher rates of purchasing. Senior management first considered the approach unethical, as customers were not informed of the experiment, while the researchers found it very appealing. However, it was fortunate that it was reverted as the company did not have any policy or framework in place. As the designers were not aware that this approach was not a good practice, this example opens a debate on what kind of education, methods, and corporate frameworks need to be put in place to avoid this kind of behaviour and insert a sense of accountability and responsibility when conducting research to develop products and features. Within the specific case of A/B testing, this example calls for a reconceptualisation of this research method to insert mechanisms of control and impact awareness.
Keywords:
Research Methods, Digital Technology
Submitted by
Dr Fernando Galdon
Royal College of Art
United Kingdom
Submitted on
May 8, 2023
This was one example of unethical design.