PI Quotes – 9 – “The most important experiments are the ones you don’t do”.

Some of the people I worked with have been a great inspiration for me. And overall, I really like to take a few steps back to think about how science works in general. The other day, trying to decide between which experiments to do first, I thought about what a great supervisor once told me: “The most important experiments are the ones you don’t do”.

In the same lines, Louis Armstrong once said that the most important notes were the ones he didn’t play; and Rodin said that the sculpting process was about removing the stone that was not part of the sculpture.


It is unsettling a first, but let me explain. In the case of scientific research, we all know that we have only limited amounts of time to discover new things, and make a point of it. Very often, I find myself in front of an ocean of possibilities, curiosities, and things that a want to try and test. One option to tackle this is to get students and let them help you. The other is to make choices, more or less rationally. Choosing, not what is the most important, and what is a bit less important, and leave it on the side for later. Only so that you uncover the main path, or the critical experiment. By the way, people now call this the “money experiment” or the “money figure”. Not sure I approve that appellation, but you get the idea.


Being in the second-half of a PhD means doing this on a weekly basis, and I sometime regrets early days of innocent and random wanderings trough experiments. I am a least grateful that my PI let me do that for a while, if only it could last a bit longer …

1 thought on “PI Quotes – 9 – “The most important experiments are the ones you don’t do”.

  1. So true. Interestingly, that is also what is know to be very important in the industry: as a product manager, you have to say “no, we will not implement that feature”. One good trick to know what to do and what not to do, is to apply the “why” question: for industry, ask the users why they want a new feature (and then look at the best way to respond to this appeal); for science: why do I want to do that experiment, what is the exact question asked. In addition, I always find it useful to draw the possible outcome of one experiment and see what could be interpreted from each possible results. It helps in control group design, but also in checking that the experiment do make sense (sometimes, two different outcomes can come, and you realise that both outcomes could be explained with each of your two hypotheses: the experiment does not help answering the question).
    It is always good to remember that doing science is a job, too.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s