Data and Error

As part of working in a lab, a large part of every project is patience. Results aren't a form of instant gratification but the product of continued investment. In the same way that you might reflect on a chemistry lab that's been left to sit for 24 hours, I've had plenty of time to reflect on what I've learned in between the active parts of my project.

Data is the end result of experimentation. Every step in investigation is to eventually be able to procure data to respond to the initial hypothesis. On the front door to the lab technician's office there is a comic in which a postdoc is approached by a criminal who tries to rob him of his money. In response, the postdoc informs the criminal of his profession and exclaims that he has no data. At the end, the criminal ends up giving the postdoc money to get him back on his feet. While there's a bit of hyperbole involved, the value of data is evident. In many labs, they will hire two people to compete for a single position and at the end whoever has the results (data) is kept. Despite the presence of luck in obtaining results, it's evident that data is the ultimate end goal of every project.

While I have yet to get any data (patience is key) having just started the preparations for my project, error has already emerged as a key concept that also ties into the importance of data. Error is essentially how wrong a result can be and can exist from various sources both obvious and discreet. In measuring with a ruler with ticks at each centimeter, it becomes impossible to accurately gauge how many millimeters something is down to the first digit and thus your results could always be off by almost a centimeter. Error is everywhere in science and in the classroom setting, only really appreciated in chemistry due to the relative importance of labs in chemistry. For those who took AP Chemistry, the silver nitrate lab stands out. Despite multiple tests all aimed at the same goal, there was variation both in repeated trials of a single test and even larger variation between tests. Despite all being conceptually sound, they all returned different results that often varied immensely. This discrepancy can be attributed to error.

In context of my own project, results are to be collected in 24 hour intervals where the cells are able to grow. However, to collect them all within exactly 24 hours is an impossibility and thus some amount of error results from the time cells are allowed to sit. To account for this, we keep it within 10% of the original time but in doing so we acknowledge the error. Even further than timing, pippeting isn't the most precise of techniques with residual liquid leftover between every attempt. Counting cells turns out a different result every time as counting is all up to the user's discretion. Beyond that, there are the most minuscule errors in every action and the errors that we don't realize. Yet despite all these possible uncertainties, science has produced undeniable results. I think what makes science beautiful to me is that they're able to draw so much out of imperfection. Newtonian mechanics despite failing when taken near the speed of light are adequate for common use, and in the same way the imprecise results we obtain in science can be applicable to everyone despite the various errors. Despite error, scientists are still able to procure meaningful data and make discoveries that have improved the quality of life.