Hacker News Reader

Comment by opus132 | Hacker News Reader

> Couldn't tell an average from a median

Without meaning to attack you personally (especially in the context of the rest of your comment), a comment like this annoys me a bit.

I presume by "average" you mean arithmetic mean, but the median is also an average, and depending on the context the median might be a far more useful statistic than the mean. Confusing the mean and the median is one thing, and perhaps you actually used this terminology in the interview; "confusing" "the average" and the median isn't really worthy of comment, and sounds more like a breakdown in communication between interviewer and candidate rather than a lack of technical knowledge. It just seems vaguely hypocritical to me to be expecting a certain level of ability from the candidate and then using imprecise/informal terminology.

View on HN

I know nothing about GP, but note that it is possible that English is not his first language, and the subtleties of the story and vocabulary might be lost in translation.

In my native language (French), as you describe it, there is no word for "average" (the technical and the common terms "moyenne" are exactly the same). Until your comment, in English, I assumed you could use "average" and "mean" interchangeably.

My native language is English and I'm fluent and well educated.

Although I do know that "average" covers mean, median and mode, when I was young I was not taught that way. First I was just taught "average" with a formula, and many years later "median" and "mode". I don't think I noticed the term "mean" until I learned about median and mode.

You can see this reflected in scientific calculators used at school, where the button for calculating mean is labelled "Avg".

And in Microsoft Excel or LibreOffice Calc, the function to calculate mean is called "AVERAGE".

So nobody should be hard on themselves for not knowing the difference. You probably weren't taught the difference, and common tools work against it.

It looks to me that the common word for statistical mean is "average" in English, but it's not the correct technical language.

> So far as I can tell, the common word for statistical mean is "average" in English

You're quite probably right about common parlance..

> but it's not the correct technical language.

..but as you point out it's not really appropriate in a technical context, indeed I don't think I've come across the term "average" being used in this way in technical literature; I've normally either seen expectation (for probability theoretic cases) or mean (for statistical cases).

> So nobody should be hard on themselves for not knowing the difference.

Absolutely, which is why I think being hard on someone for not knowing the difference between "the average" and the median is similarly an issue.

You realize that an interview is not the same as writing a technical article, right? It should be a conversation between two potential coworkers, and, in that context, I think saying "average" instead of "mean" is totally fine.

Incidentally, "mean" doesn't quite cut it, if you want to be a stickler about it. There are several types of means: arithmetic, geometric, harmonic, etc. By that standard, you just failed the interview.

> "mean" doesn't quite cut it, if you want to be a stickler about it.

I'm going to assume, then, that you didn't read my original comment which started off this line of discussion? By that standard, you just failed the interview. ;)

Sorry, I dropped this! :P

Touche.

French and English are strictly equivalent.

moyenne is mean or average. The formal name is "arithmetic mean" and "moyenne arithmétique".

median is médiane. people frequently use any word but median in both languages when they want to talk about the median.

https://en.wikipedia.org/wiki/Arithmetic_mean

https://fr.wikipedia.org/wiki/Moyenne_arithm%C3%A9tique

Almost correct.

The mean, median, and mode are three types of averages. None is 'the average'; though if you asked someone random on the street they'd probably give you the mean average.

This really depends on context. When evaluating a likelihood (prediction), "average" is generally colloquial for "expected median."

On an average day?

Mode: On a typical day <<-- most common meaning for average in day to day conversation

Mean: On a day that is in the calculated middle (for some variance) <<-- most common meaning when we think the question is about math

Median: On a day that is the middle for all the possible days <<-- nobody thinks of this, but they think this is what the mean is.

In most practical situations, your initial impression is correct. As a native English speaker, I've never heard seen or heard someone use "average" to refer to a statistical summary other than the mean. If it doesn't refer to the mean, it doesn't mean anything formal/technical at all. I read GP as "median is sort of an average in the broad sense", but this is honestly kind of a stretch.

Math definitions only..

I see it quite commonly, from technical discussions to newspaper article. When people say "the average income in the US" they are almost never talking about the mean.

I think they are. I doubt any newspaper articles use it to refer to the median, at least I've never seen them use it like that. Do you have any examples?

Almost every discussion about average salary or iq uses median. Even finding results for mean salary is difficult. The first page of results for mean us salary gives a list of results for average, many of which specify they are actually the median.

https://www.google.com/search?q=mean+us+salary

Well, https://en.m.wikipedia.org/wiki/List_of_countries_by_average... as a counterpoint. Almost all the Wikipedia pages about average wage use mean.

All you're pointing out here is that a candidate should be able to differentiate between average, mean, median, and mode, and should be able to swiftly correct "what's the difference between average and median" to "do you mean mean and median? because a median is a type of average".

Given that, showing more charity to the comment you're responding to would have been the correct course of action.

Yeah but will embarrassing the interviewer work well as a strategy to get hired?

Depends on the interviewer.

Do you want to work at a place where you cannot event define the words used before solving a problem?

...I always learned in grade school math class that "average" = "mean", very specifically. I'm pretty sure there would have been test questions that required this knowledge.

It's very possible my schooling was wrong, but presumably I'm not the only one.

I have a graduate education in probability and statistics; in the technical literature, an "average" is an estimator which, subject to some bias, predicts the value a sequence tends to. My professor was fond of saying, "the function f(x) = 15 is an average, but since it's a constant function it will almost always be a terrible one" to drill this into our minds.

The arithmetic mean is colloquially called the average and differentiated from the median and mode, but technically speaking these are all averages. They are all estimators of central tendency. It would be not be out of the ordinary if a statistician asked, "which average do you mean" for clarity. Which definition of average is best depends on what you're trying to measure.

I dislike this interview question because it's an area where it's 1) more trivia than practical knowledge, and 2) easy to not recognize someone who knows more about it than you do. It's like if I tried to test your knowledge of the difference between "affect" and "effect", and then you correctly used "effect" as a verb and I thought you were wrong.

If I repeated everything I just said about estimators in an interview, the interviewer might think I'm too ivory tower to realize that "of course he's just talking about the mean!" But then I could also ask why I'm being asked a question like this in a software engineering interview.

It's a language game where no one wins (somewhat in the sense of Wittgenstein).

> "the function f(x) = 15 is an average, but since it's a constant function it will almost always be a terrible one"

It has minimum variance, how can it be a terrible estimator?.. ;)

Hah, that's a good rebuttal. Well it would be an excellent estimator if it was unbiased (or at least had very little bias). But if f is an estimator for the true average of a function F where E[F] = 100 (for example), it's very clearly not unbiased.

It is a bit tongue in cheek.

> It is a bit tongue in cheek.

As was my reply ;)

It is a very precise function, I'll give it that.

Thank you for articulating this. This issue has many manifestations in the interview process. It can be so frustrating.

Reply on news.ycombinator.com