Friday, July 20, 2012

Friday Fun: Science is Slow

Picture by Scientific American blogger Jason Goldman
When I was in my first year as a graduate student, one sage Emeritus professor said to me, probably in a reaction to some of my bubbling enthusiasm about psychology, "Michael, you know that science is slow, don't you?"

In response I must have said something like "Oh yeah!" or "Of course I do!" but in reality, I didn't. Now, having several battle-tested years under my belt (battling reviewers, puzzling data, my own malaise) I know a bit more about what that professor was saying to me all those years ago. In today's Friday Fun here on PYM, I'll show you some data revealing how long it takes me to publish empirical papers. You'll see after you read this exactly what is meant by the phrase: Science is slow.



One quick aside: I consider psychology to be a science in that we use scientific methods. Others will disagree, and to them, I don't really know what to say. Maybe read what Tim Wilson said on the matter and come back to me if you have questions? ;)

Why is science so slow? You might ask that question and in truth until I really had a few studies under my belt I didn't realize exactly how slow. It's slow because of one or more of the following occurring on a regular basis:

(1) The results of your study don't fit your hypothesis and you are left puzzled.


(2) The data confirms your hypothesis, but unfortunately, the hypothesis was not a good one, and so you have more questions than answers.


(3) The data show nothing that others don't already know. You're crushed.


(4) The study was designed with tragic flaws which render it's interpretation impossible.


(5) The data say the opposite of what you expected and so you are not certain whether (a) the data represent a real phenomena or (b) human error caused this unexpected effect. (The human is you in this case).


(6) The study you designed requires an empirical technique (e.g., scoring minute-to-minute physiological responses, coding facial expressions) and it's going to take a while for you to learn how to do that.


Etc...

Anyway, you get the idea. We all have what seem like brilliant theories about how people think in social contexts and unfortunately, not all of these ideas pan out successfully when you get right down to the actual empirical work. At times the uncertainty is frustrating and soul-crushing. At other times, it represents the fun in psychological science. No matter the emotions, what these little speed bumps mean is that the formation of a published research paper takes a little longer than you'd expect.

So how long does it take? To answer this question, I sampled 12 of my empirical research papers for which I am relatively certain of the earliest time the project started (e.g., when data collection began, or when ideas for the project were hatched). I then counted the months from that starting point until the project was completed. The raw data are reproduced below:
Click to enlarge!

On average, it has taken me about 56.3 months to bring a paper from inception to publication (nearly 5 years!!!). You might wonder if much of that time is spent waiting for journals to copy edit and publish the paper, but actually, that time lag isn't all that long according to the figure (about 5.6 months on average for journals to publish papers). Some worry a great deal about publication delays, suggesting that it hurts impact factors (Yu, Wang, & Yu, 2006), but psychology doesn't seem to have too much of a problem there. The majority of the time is spent collecting data, analyzing data, being discouraged, collecting more data, analyzing more data, etc...

Before I end, I want to draw your attention to a couple of interesting points on the graph. Paper #9 -- the shortest transit time between inception and acceptance for publication at 14 months -- went so rapidly because it is a single-study paper using participants collected online. You can see how that would speed the whole process up dramatically. In contrast, my two longest transit papers, #12 and #2 have some fun stories associated with them. One of the studies from Paper #2 (84 months) was my undergraduate honors thesis, and so there was a bit of lag time for this publication where all the regular research speed bumps happen, in addition to me learning the ropes of research. Paper #12 (111 months from start to finish!) interestingly enough has data collected before I even became a graduate student. The data were collected for an entirely different set of research hypotheses, but we were able to use it to test some new hypotheses several years later!

After seeing these informal data, what do you think about the idea that science is slow? Is that a problem for psychology?



Yu, G., Wang, X., & Yu, D. (2006). The influence of publication delays on impact factors Scientometrics DOI: 10.1007/s11192-005-0249-4

3 comments:

  1. Publication is definitely slow, but the detail available in science also makes things dreadfully slow. When we have any unanswered nuance in an idea or data, it can be explored so deeply that it becomes its own research discipline. This is really attractive because it allows scientists to distinguish themselves, e.g. graduate students (me) trying desperately to find something new to study. I think this makes science a terrible way to actually learn things that could be useful right now.

    ReplyDelete
    Replies
    1. Science can be conservative at times, which adds to the delay. That is sometimes a necessary circumstance given that researchers should be trying to get things right, at least to the best of their ability. Sometimes fame, recognition, or prestige can get in the way though...

      Thanks for reading!

      Delete
  2. This comment has been removed by the author.

    ReplyDelete