I have the opposite problem to this question. Often when I talk to someone from a 'hard' scientific field I am 'taught' how statistics work or how programming works. It is a bit like being mansplained but by a hard scientist (hardsplaining, would that be a word?). Once someone was surprised that I know what the Runge-Kutta method is, and I have lost the count of how many times I have been lectured about stuff like the normal distribution and inferential statistics.

Now, I work with quantitative social science, using statistical methods that many would consider quite advanced (e.g. exponential random graph models for social networks, panel econometrics, simultaneous equations) and had the necessary mathematical training for that.

I am also very familiar with several programming languages. I usually write a lot of code in R or Python, besides being proficient in statistical scripting languages like Stata. I also know a good deal of NetLogo and Java, which I used when I did agent-based modelling. I would write my stuff in Sweave and LaTex if most journals didn't ask for a word document.

My understanding is that precisely because of the observational nature of most social science data, those of us who are into quantitative methods are forced to learn very advanced techniques to deal with issues such as selection bias and unobserved heterogeneity. Furthermore, the intrinsic 'messiness' of social data means that we have to be quite good at data management, usually learning a programming language or two in order to clean our datasets. Moreover, the emergent nature of social phenomena has motivated many of us to use multi-agent simulations in our work, demanding us to learn how to program.

Yet, I get patronised by the person whose randomised experiment allow them to get away with a t-test. How to react when that happens without sounding too defensive?

I know that generations of armchair sociologists theorising about the social construction of this and that probably created this stereotype of the mathematically inept computationally illiterate social scientist, but I believe that this image doesn't reflect a great share of those in my field.

EDIT 1: Given the close vote, I'm adding this clarification to what I'm asking. I want to know of strategies, possibly by others in a similar situation, to assert their research credentials or skills in a friendly manner in a social situation where the phenomenon described happens. The type of prejudice that I describe sometimes leads my opinion or ideas to be disregarded because of judgements made based on an incomplete image of my field of study. I believe that there are others there in a similar situation (Economists, for example), who may have strategies to cope with it. There must be a nice way to convey one's competence past the initial impression caused by the stereotype.

  • 5
    I don't think this problem is really academic in nature. After all, it's hard to treat others as individuals, and applying stereotypes based on some simple traits is much easier. If there were a straightforward answer to this question, then things like racism or sexism would have become bygone issues of our society. You would think that scientists could do better than that...
    – Drecate
    Commented Dec 20, 2015 at 21:24
  • 5
    How about calling yourself a "quantitative sociologist"? Commented Dec 20, 2015 at 21:48
  • 2
    @Drecate: Sure, the problem of prejudices isn’t academic in nature, but this specific prejudice is and there may be specific strategies for dealing with it.
    – Wrzlprmft
    Commented Dec 20, 2015 at 22:02
  • 1
    @Kenji Both answers already there seem pretty good, get in there first. Imagine yourself on the other side, someone, say a Phd student you're talking to starts talking about problems you know exactly how to solve but doesn't mention any of the keywords. They might have spent every evening since they were 12 reading books on econometrics but unless they mention any of it you don't know. You could assume they're all secretly masters of the craft hitting no common problems and remain silent for fear of them getting offended but that would be kind of a dick thing to do.
    – Murphy
    Commented Dec 21, 2015 at 13:32
  • 1
    Roughly speaking that matches the job description of half my department and "point and click software to get a table with stars." isn't so far off what a reasonable fraction do so expecting people to infer that you're definitely in the other part may not be reliable. Thinking back, if people lecturing bothered me I'd be bothered a lot (I actually like when someone starts lecturing on things I'm interested in and don't know) but I think simply quickly waving off the start of lectures on things which I already know about saying I know about them makes the lectures go away before they bother me.
    – Murphy
    Commented Dec 21, 2015 at 14:44

5 Answers 5


I get patronised by the guy whose randomised experiment allows him to get away with a t-test. How to react when that happens without sounding too defensive?

I think this is the key aspect in your question. Once the "hardsplainer" starts patronizing you, you are on the defensive. And there is really no good way out of this situation once you react by starting to defend your methods.

  • Sometimes the hardsplainer will realize that his assumptions were erroneous and that you may indeed know more about stats than he.

  • Or he will in turn get defensive and start nitpicking your methods, probably getting in deeper and deeper water as he is discussing stuff he may not know much about. This is not a good conversation to have at social gatherings.

I'm afraid the second possibility will happen rather frequently, simply because people are not good at revising preconceived impressions.

So I'd recommend that you nip the problem in the bud, by not allowing your interlocutor to, in fact, preconceive the impression of "look, a social scientist, who probably doesn't know anything about statistics". Specifically, when you discuss your work, invest half a sentence to name-drop your analysis techniques.

I'm looking at how foo relates to bar. Because I only have observational data, not experiments, I use econometrical panel data models, and I find that...

If you hint that you use advanced models right before the hardsplainer can get the wrong impression that he can lecture you with fundamentals, he will be stopped cold. (Of course, you don't want to overdo it to come across as an arrogant know-it-all.)

Yes, it would be nice if this were unnecessary, because people didn't have the preconceived notion that social scientists are inept in terms of statistics. Unfortunately, this notion does have a basis in facts. I do statistics for psychologists, and I see that while they do get a solid grounding in statistics, they do frequently misapply models, or interpret them incorrectly, or don't understand why p-hacking is a Bad Thing. Then again, some hard scientists do suffer from the delusion that being an expert in some hard science means that they automatically also are experts in statistics.

  • The point to avoid "defensive-looking" is to treat the hardsplainer as if he were interested in the OPs methods. I personally found that giving these guys fine technical details is just the right level to get rid of the the disingenuously involved and keep the seriously interested on board for a serious scientific conversation - if the hardsplainer would find it arrogant, it says more about him than about the OP. Commented Dec 22, 2015 at 18:41
  • @CaptainEmacs: as to appearing arrogant, I'm more worried about innocent bystanders. After all, I'm advising on how I'd act before I even know who I'm talking to. I'm standing at a party with four people, none of who I know, and everybody introduces themselves. We need something that keeps a potential hardsplainer from hardsplaining, while not coming across as arrogant to the non-hardsplainers in the group. Commented Dec 22, 2015 at 19:40
  • @Kolassa That's exactly the point: you treat them as if they are interested in your methods. If they are, great! If they are not, and just are there for show-off, they'll drop out of the conversation. You will be the person, however, that treats them under the assumption of respect by default. Commented Dec 23, 2015 at 19:20

There is an important aspect of this dilemma which I did not see stated in your question: why is the opinion of the "hard-splainer" important to you?

Depending on the answer to this question, there are a number of different approaches that you might take. Although I am in a "hard" field myself, I face similar dilemmas in my interdisciplinary work, as I find that some researchers often dismiss or misunderstand computer science as "just data analysis," since their own experience with it has been largely limited to simple uses of Matlab, Excel, or specialized data analysis programs.

I have found that it is useful to develop a spectrum of responses, depending on my degree of investment in the interaction. From least to most investment, these are approximately:

  1. Nod and smile. When dealing with a random boor in an airplane or at a conference cocktail hour, I may simply choose to not engage. Why should I care what a fool believes when they cannot even bother to draw breath long enough for me to speak? So I nod, smile, say something politely vague, and then go to refill my drink / take care of some work / whatever.
  2. Turn the tables. If the person seems worth talking to and capable of listening, but is uninformed, then I'll turn it into a teaching moment:

    "Ah, it sounds like you're suffering from some common misconceptions about [subject]..."

    It's good to spend some time with philosophy of science and the history of your field, in order to understand why things are the way they are. I have some favorite examples and anecdotes, mostly having little to do with my own work, which help illustrate the base pop-science-level points I'd want to make. Don't be defensive: instead, have fun sharing your love of the field and its complexities, and your conversation partner is more likely to enjoy themselves too and actually learn something about your field.

  3. Face the problem head-on For people whose opinion really matters to me, such as potential collaborators, program managers, and decision-making committees, I often actually have pre-prepared slides or, in some cases, actual published papers. You don't have to listen to a lecture, but instead say something like, "It sounds like you're concerned about [issue]," and then head for your careful "101-level" explanation of how that issue is approached in your work. Preparing material of this type can actually be a surprisingly valuable exercise for yourself as well: I have found that many of the assumptions of my field, while valid, have a much more complex backstory than I commonly think about while working within those assumptions, and taking time to understand them has opened up new knowledge and opportunities for me.

Given the example you indicate, how about: "Yeah, a t-test is probably just fine if one has good data. But if you have really bad, realistic, data, you need much more sophisticated methods such as ..."

Ideally mention computational methods the techo (the "techno-macho", or hardsplainer in your language) is not likely to know or understand. If they had only sought to show off their superiority, that'll make them go away.

And if the person is seriously interested to learn, then, that's fine, too. In this case, go forth and explain.


Let's say you and a hard scientist have just struck up a conversation. Before saying anything about your work or position or department, make sure to find out a fair amount about what he does, and what statistical or computational methods get him excited, so you can tailor your remarks to his own provincial world view. (I'm saying he on purpose, because I suspect the behavior you described is more common among men.) If he peters out and pushes you for information before you've gotten enough from him, say something vague or even misleading, to throw him off the scent of social sciences, so you can get him to talk excitedly about himself a little longer.

When you're ready to start talking about yourself, say something that allows him to view you as a member of his own subtribe. Your remarks might be a little bit disloyal to your fellow social scientists, but hopefully you can live with that, since this is an informal one-on-one conversation.

For example: "My work is inter-disciplinary. I use such-and-so techniques that have been long established in the hard sciences, and apply them to interesting problems in the social sciences." Raise your eyebrows a little as you pronounce the phrase social sciences. It needs to have just a whiff of a skunky odor in the way you say it.

If you want to go a little further -- "It is remarkable how much can be accomplished by bringing rigorous statistical and computational methods to bear in previously unplowed fields."

(If the conversation eventually leads to criticisms of your area that are uncomfortable for you to hear, you may then safely establish your respect for your colleagues, for example: "You are thinking of the social sciences of 20 years ago. Actually, there are many researchers engaged in rigorous work in the social sciences currently. I am not the only researcher in my field applying sophisticated mathematical and statistical techniques to complex social science questions." Then you could rattle off some names. Not to convey meaning necessarily -- but to intimidate.)

Please do not take the arrogance and lack of humility of some hard scientists personally. It's really their problem, not yours.


I think this is why some people in biology started calling themselves mathematical/computational biologists. Then you also have bioinformatics people. Mathematical finance is a discipline. Etc. I'm pretty sure people changed their title for this very reason, and I think the quickest change is to change your informal title. If you just say you're a "mathematical/computational/statistical/quantitative __________" it likely sparks interesting conversation towards the direction of your research.

And if anyone is being a patronizing idiot, throw it right back at them. They're using a t-test? "Why wouldn't you be using M-estimators or a Bayesian model with t-likelihoods with the degrees of freedom as a hyperparameter? Your data seems like it would have outliers and so such basic statistics probably don't apply." If they don't get the point, they likely aren't worth talking to.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .