To be or not to be good science

Guest blog post by Audrey Richard.

Recently, a friend of mine claimed to me that: "if you want to increase your chances to attract a man and eventually seduce him, go for red. It's a scientific fact. Clothes, lips, accessories: the whole package." Like this:

Why he felt he had to give me such a piece of advice, I won't tell. But red is indeed known to evoke many physiological and psychological effects in humans, including women being more desirable to men when they are associated with red clothes or objects. To indicate their sexual receptiveness and ovulation, females from several primate species show enlarged, red swellings that are preferred by males compared with less brightly coloured skins. Based on these observations, it has been proposed, under the name of the "sexually salient hypothesis", that men might themselves have a biological predisposition to interpret the colour red as a signal sent by women to announce their own sexual proceptivity. And shortly after my friend had explained all of this to me, I came across a PLoS ONE paper in which anthropologists conducted a study to assess whether the "sexually salient hypothesis" could be, at least in part, validated by a tendency of men to be more attracted to reddish than pink genitalia. So, basically, young and heterosexual male participants were asked about their preferences for female skin colouring. The answer is given in the title of the study: "Red is not a proxy signal for female genitalia in humans". Incidentally, men actually tended to find the redder ones significantly less attractive. At least according to the conditions and methods chosen by the authors of this particular study to collect and analyze the data.

I wasn't disappointed by the conclusion, nor was my mind blown: cross‑species analogies have their limits. Fair enough. I wasn't expecting anything special actually. But at this point I got curious about the generalist coverage this paper could have had and dug a little deeper. I ended up on the Slate website, read the article and most importantly the comment thread. Finally, among the more or less serious comments about the story (that were far from being the majority of comments), I found these:

"Don't pretend this mumbo jumbo is even close to being real science".

"[...] If I wanted to read trash, I'd turn on Fox"

"[...] Maybe we are running out of things with substance to study and even report?"

[Disclaimer: In the name of making my point, I will further pretend that a reading of less‑than‑50 words comments has given me a clear insight into the mind, thinking and true intentions of people. Don't do this at home.]

I'm not judging the persons to whom these opinions belong for the sake of great virtues including the respect every human being deserves and freedom of speech and [fill in the blank with any notion of the Declaration of the Rights of Man and of the Citizen you feel might be appropriate]. However, one of these fine and delicate people actually made a comment under the alias "Monkeybutt". Very a propos in the context indeed, but assuming he or she (depending on red swellings, remember?) has free will, I might have judged this one a little bit. But what can I say? I do have opinions myself after all. For instance, I believe that someone who reads the words "genitalia" or "vagina" and instantly makes a link with "trash" despises women, perhaps as a result of mother issues. But that's not the point. The point is that I fear these people may have labeled the whole thing "bad science" based on the study's topic alone. Don't get me wrong, I'm not desperately standing up for the study. I'm just wondering: are there any valid reasons to brandish the red flag? Blaming the topic is irrelevant. Because you know, sometimes... A monk... Thousands of hybridization experiments with peas... And BAM: you've got modern genetics. I mean, seriously, it was PEAS. So let's keep an open mind about that, shall we?

In the end, the question actually is this: what defines something as "good" science? I suppose the answer to that question implies that the very nature of science has already been elucidated, but I won't even pretend I'm bold enough to go down that road, especially not in a foreign language: I'm French, not crazy (I can hear you, you sarcastic). Anyway it doesn't mean we can't TRY to understand together some of the principles that underlie quality science as defined by the scientific community. Although this exercise may be quite challenging, I suppose it would be a hopeless endeavour for anyone who was not familiar with - wait for it - the scientific method. What would you do without me? Indeed.

The pyramid of stages to be passed to eventually achieve technology starting from common sense shows the scientific method as both the result of philosophy of science and the basis of scientific specialties. Well, I couldn't have made this from scratch but intuitively it really makes sense. A common scientific method implies that no matter which scientific field you are particularly in love with, you should be able to understand the way another scientist conducts his or her research and other scientists should be able to understand yours even though you are worlds apart at first sight. Incidentally, it probably explains why, as a virologist, I can still see the point of research conducted by anthropologists or neuroscientists, even though such research makes my eyes and ears bleed. As to what exactly the scientific method is made up of, I guess we've all been given the same lesson at some point. But what does the theory really mean? If you intend to commit to science, this might interest you as a practicing scientist‑to‑be and therefore huge reader of papers.

Principals of "Good Science" as they Appear in an Scientific Article:

1. Hypothesis formulation, aka "What is the question in my topic that hasn't been answered yet?" or "What do I want to demonstrate?" (or "What am I forced to work on???" Hi PhD students!)

Very first step, but not trivial whatsoever. Assuming you have a hypothesis to formulate, it means that basically you have made an interesting observation, checked the literature to place it in the context of the field and concluded it is worth testing. That's something. In a paper, it's typically the Introduction part. But BEWARE! According to Karl Popper, philosopher of science, your hypothesis should be falsifiable, meaning that a methodology should exist, not only to prove it but also to contradict it (Lord Popper, forgive me for describing your whole theory in such a scandalously simplistic way)

2. Hypothesis testing, aka "How am I going to (dis)prove my hypothesis to be true?"

Tricky part since this is the moment when you have to choose the methods most appropriate to answer your question. You'll find hypothesis testing in the Materials & Methods part and all along the Results part of a study, since the testing process generates the data using the chosen methods. The possibilities here are almost infinite depending on the topic, but also, let's be honest, money. Sometimes great cellular and animal models or perfect equipment do exist but your lab just can't afford them. However, objections like "I have the materials but no protocol" or "I have the materials and protocol but I've never done this before" are not an option.

3. Deductive and inductive logic, aka "What do these data say to me at this point? What should I do consequently? What is missing so that I can eventually prove (or disprove) my hypothesis?"

With deductive and inductive logic, scientists make things seem so easy and obvious and well‑interlaced. As if everything's worked so perfectly and they've never wanted to set their bench on fire. Impressive. The way the Results part is structured should show you these logical links.

4. Controls, replication and reproducibility, aka "What can I do to make sure that my data are not biased and can be interpreted as objectively as possible?"

REQUIREMENTS‑FOR‑GOOD‑SCIENCE ALERT. The saying goes that many scientists don't know the difference between replication and reproducibility. A sad era we are living in. This is REALLY concerning because "I observed this mind‑blowing result in 3 independent experiments performed in triplicates and I can see you are trying to hide that you are hating me a little right now" is so not the same as "My 9 replicates gave me this mind‑blowing result in one experiment among 163 others I've unsuccessfully performed in the exact same conditions and made me cry every night for the past 6 months, so could you try to hide that you are willing to smile, thank you very much". If someone just tells you "My mind‑blowing result is reproducible", you are stuck because it's too embarrassing to inquire into whether he or she has used the word "reproducible" properly. As a reminder, reproducibility implies that a whole experiment can be reproduced independently several times by the same person, or even better (but rare), by different people. By contrast, replication concerns the repetition of each experimental condition (replicates) in the same experiment and therefore assesses the inner variability of the method. If you are skeptical about how reproducibility can be roughed up, please read the features in Nature and Reuters that are linked below and you'll understand why my keyboard is wet with tears. I could mention how statistics are insanely misused but I can't afford a shrink.

5. Interactions between data and history, limits to science's domain, aka "What can I conclude? Has the question been answered? How do the biases I couldn't get rid of impact my interpretations? Am I in agreement or in contradiction with previous studies? What is my contribution to the topic? What's next?" or "Let's write science history" (Regards, PIs!) or "I REALLY need to get out of here. Is there some kind of voodoo ritual that makes articles write themselves?" (Hi again PhD students!) or "Has a 9^th author (on 8 listed) ever won the Nobel Prize?" (Ah undergrads, so cute...)

You've surely recognized here the crucial Discussion part. Please let me just add: DO NOT over-interpret. Don't speculate so that your work looks sexier than it is. Because it won't. But you, you'll probably look like you are trying to fool people. Or simply delusional.

Of course these are VERY GENERAL principles. Some of them have to be adjusted depending on the scientific specialty, like falsifiability in paleontology since (dis)proving hypotheses related to such a field requires a DeLorean DMC‑12. Many other factors are also to take into account, and two that should be more particularly mentioned are ethics, for reasons as various as obvious, and last but not least, peer‑review, which undoubtedly provides credibility to science. However, I personally keep an eye on not letting the "high impact factor mojo" cast its spell too much on me because it makes skepticism lazy (I love you though Nature). It's a tricky balance to achieve actually.

And what about you? Are you committed to science or intending to? If you are in a lab already, do you feel that maybe sometimes the urge to publish overwhelms the respect of scientific method? Feel more than free to share your observations and experiences. But please, do not sign in as "Monkeybutt".

Audrey recently graduated with a PhD in molecular and cell biology. She doesn't know exactly what her career is going to be made up of, but she does know that there’s going to be science in it. Viruses make her heart beat a little faster - she has a big crush on them, actually. And she is French, so she plays the “I don’t understand what you are saying” card if you are not kind enough with her (she thought you should know about that.)

---

References:

1) Source references:

Original paper testing the sexually salient hypothesis http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0034669 (Figure 1 = graphic content)

Generalist coverage on Slate website http://www.slate.com/articles/health_and_science/science/2012/04/red_genitalia_study_testing_the_sexually_salient_hypothesis.html

2) Photo credits

a. http://en.wikipedia.org/wiki/File:Womaninredposter.jpg#file

b. Adaptation from http://fr.wikipedia.org/wiki/Fichier:Gregor_Mendel.png

c. Adaptation from http://books.google.fr/books?id=iVkugqNG9dAC&printsec=frontcover&hl=fr&source=gbs_ge_summary_r&cad=0#v=onepage&q&f=false

3) Further readings on the effects of red colour on humans:

a. http://www.psych.rochester.edu/faculty/elliot/documents/ElliotNiesta_RomanticRed.2008.pdf

b. http://www.nature.com/nature/journal/v435/n7040/abs/435293a.html

c. http://www.sciencedaily.com/releases/2008/10/081028074323.htm

4) Readings on philosophy of science and scientific method:

a. http://plato.stanford.edu/entries/pseudo-science/#AltDemCri

b. http://books.google.fr/books?id=iVkugqNG9dAC&printsec=frontcover&hl=fr&source=gbs_ge_summary_r&cad=0#v=onepage&q&f=false

5) Readings on reproducibility issues (RECENT):

Nature Editorial http://www.nature.com/nature/journal/v483/n7391/full/483509a.html

In Reuters http://www.reuters.com/article/2012/03/28/us-science-cancer-idUSBRE82R12P20120328?feedType=RSS&feedName=scienceNews&utm_source=dlvr.it&utm_medium=twitter&dlvrit=309301