Learning it 'Right' the First Time...

A.S. I realize that this is an incorrect use of 'right'. 

During my Ph.D. a few postdocs lamented that they felt as though they didn't have time to sit down and 'learn how to do things properly'. For instance, when you're analyzing data, there are often many ways to accomplish straightforward tasks such as removing redundant entries from a list or concatenating two large tables. Similarly, if you've ever had the opportunity to learn a bit of PERL scripting, you quickly find out that there are often many, many different ways to get the desired results. However, saying that there are many different ways to do something doesn't mean that all ways are equally as efficient. I had a pretty shocking realization of this a couple of years ago when changing a few lines of a script altered its run time from overnight to ~5 mins (I learned about hashes).

Changing the time it took to generate some data from 8-10 hours to 5 mins was a massive gain in efficiency. Instead of taking several days to get results, I could process my whole dataset in a morning. In fact, even if it took me a whole day of searching, reading, and trial-and-error to learn how to boost my efficiency, it would probably still have been a net benefit.

Yet, most of the time, I find myself in the same mindset as the postdocs referenced at the beginning of this post - it's easier to just trudge through these data (or this protocol) using the techniques that I know rather than sit down and 'waste time' to look up a more efficient method1. Despite realizing that learning how to do something correctly now will reap future efficiency rewards I nevertheless offset future gains due to present circumstances (always pressing myself to 'get more done').

I suppose that it's a small comfort that this phenomenon has been quite well-studied in both psychology and economics under the general framework of intertemporal choice. In general, people have a tendency to perform what is called 'delay discounting', or to discount (often sharply) the value of long term rewards once they pass some arbitrary time threshold (e.g., would you choose between receiving $100 today or $110 in a month?).

Such 'bias for the present' (see this press release for an example), is logical under many circumstances as the future is uncertain. It doesn't really make sense in my case because I've seen the work put into learning some new aspect of PERL, or R, or Linux (or any number of things) pay off time and time again.

I've resolved to be more judicious in my allocation of time towards improving my skills in areas where such improvement would benefit my work. One way to offset the present 'cost' of such self-instruction is to invest more of what would otherwise be 'free time' into it - this may be a good way to offset the 'guilt' of not feeling like I'm 'wasting time learning'.

I am in a field where learning is perpetual, so don't get the wrong idea. What's important to keep in mind is that it often pays off to learn how to do something properly (read efficiently) from the get-go.

1In my experience it's often obvious that there's a better way to do whatever it is I'm doing. When you're developing a new protocol or method, you're typically well aware that you're breaking new ground. Otherwise, there's nothing new under the sun.