Shape of distribution of histogram

Question

If we have the following histogram:

enter image description here

And, want to describe it based on one of the following options, which one do you think is correct? Why?

positively skewed
symmetric
none of the above
negatively skewed

My initial suggestion would be "3. none of the above".

What do you think?

Thanks.

It would be good to write down your reasoning for choosing 3. — nico, Apr 09 '12 at 12:37
@nico: (+1) I think you are right and we all want to see some motivation and thought put into homework questions. It usually is helpful to point the OP to the FAQ in these instances, too, I find. — cardinal, Apr 09 '12 at 14:39

Peter Ellis · Answer 1 · 2012-04-09T00:53:56.447

5

I think you should ask yourself, if it isn't symmetric, then why not? Is it really likely that something that is not symmetric is not skewed in either the negative or positive direction?

What is the feature about this histogram that makes you think it isn't symmetric? Then, how would you describe that feature?

edited Apr 09 '12 at 00:53

answered Apr 08 '12 at 23:51

Peter Ellis

17,650

(+1) Good leading questions. But yes, it is possible for an asymmetric histogram not to be skewed: http://stats.stackexchange.com/questions/24853. – whuber Apr 09 '12 at 00:39
1

Yes, I thought of that as I wrote the answer. But in this case... – Peter Ellis Apr 09 '12 at 00:53
This should really be a comment, as it does not answer the question. – nico Apr 09 '12 at 12:36
@nico: it's common, and I think better, for homework questions to receive leading questions rather than solutions. – Neil G Apr 09 '12 at 12:58
@Neil G: I agree with you 100%, but an answer... well it should provide an answer. If you just want to ask questions to the OP (which is very good) then you should do that in a comment. I guess the main problem is not your answer, but rather the poor quality of the question. – nico Apr 09 '12 at 13:09
2

@nico: It's not my answer, but if you think leading questions don't make good answers, maybe you should ask about that in meta? – Neil G Apr 09 '12 at 13:17
@Neil G: I cannot vote to close, but I flagged and downvoted the question (sorry did not notice it was not your answer, and by the way, there is already something in meta: http://meta.stats.stackexchange.com/questions/12/how-should-we-deal-with-obvious-homework-questions). – nico Apr 09 '12 at 13:25
6

@nico: My feeling is that Peter has responded to the question exactly as expected and intended given this is a homework question. – cardinal Apr 09 '12 at 14:36

NaN · Answer 2 · 2012-05-28T10:43:39.917

Check out the wikipedia article http://en.wikipedia.org/wiki/Skewness.

I see that the mean and standard deviation are provided, not clear what is the statistic below that. If you are able to infer the median from this histogram, you could also calculate the Karl Pearson coefficient of skewness. This is defined as 3(mean - median) /Standard Deviation

If this statistic is greater than zero, the distribution is positively skewed, if negative then distribution is negatively skewed (this test might not work for bi-modal distributions though)

score -4 · Answer 3 · answered Apr 09 '12 at 12:28

-4

The answer here is "1. positively skewed".

answered Apr 09 '12 at 12:28

Simplicity

545

7

You should really explain why – nico Apr 09 '12 at 12:36

Shape of distribution of histogram

3 Answers3