Power Play

Friday, December 13, 2019

UCSF Parnassus Campus, somewhere deep in the animal tower – Statistics is a key element in learning to plan good biological experiments that will hopefully yield results that are both biologically and statistically significant.

To break it down, please get a P-value of <0.05 or repeat your experiment. However, with great statistical significance, comes the need for great statistical power.

For that, you will need a sufficient number of biological replicates. Animal studies unfortunately, can be expensive, making it difficult to power such experiments.

Enter grad student Chary Pinchpenny’s new method of powering a study.
As part of her thesis project, Pinchpenny had to do some animal experiments which required special housing and diet.

The unique arrangements required a lot of money for fancy equipment leaving her with just enough budget to get sufficient animals.

However, as grad student stories always go, tragedy struck one day when she lost some of her mice quite literally, when someone left the cage unsecured.

The lab has since been able to round up the four-legged escapees, but the mice could no longer be used, due to exposure to undesirable environmental factors like bacteria, viruses, and fake news.

Undeterred, Pinchpenny decided to use her remaining mice in a creative way.

“I realized that there is more than one way to look at the word power. Traditionally we think of great power as big numbers of stuff. But what if we look at power in other ways, like energy?” Pinchpenny says with great gusto.

With that brain wave, Pinchpenny has swopped out the remaining mice’s drinking water with Red Bull, a popular energy drink.

For all of you who are wondering, of course it was given before its expiry date. In this set up, the mice are more energized and powered to get through the whole experiment.

To further maximize her resources, Pinchpenny plans to do all her analyses on a MacBook Pro hooked up to two chargers. The statistical test will be the student’s t-test of course, because Pinchpenny is a graduate student.

There are questions that Pinchpenny herself admits still need to be addressed.

Pinchpenny wonders out loud, “With this novel experimental set up, should I be looking at P-adjusted instead of P-value, since my experimental setup has been modified?”