[BlindRUG] Measures of modality: understanding distribution shapes

[Godfrey, Jonathan]

Hi,

This is a great question.

Personally, I would be hoping the sample size is sufficient to allow me to increase the number of bins in a histogram to convince myself that the presence of bimodality is real. I can get the information from the VI() command on the histogram creation command. For example:
VI(hist(x))

I'm still trying to find a way of conveying a density curve to a blind audience. It's very easy to follow a tactile representation of a density, but conversion to text is proving challenging. I suspect this is one situation where sonification of the curve might also prove to be an answer.

Cheers,
Jonathan



-----Original Message-----
From: BlindRUG [mailto:blindrug-bounces at nfbnet.org] On Behalf Of Gjalt-Jorn Peters via BlindRUG
Sent: Thursday, 30 July 2015 6:29 p.m.
To: BlindRUG at nfbnet.org
Cc: Gjalt-Jorn Peters
Subject: [BlindRUG] Measures of modality: understanding distribution shapes

Dear BlindRUG readers,

I'm working on revising our first statistics course (in a psychology curriculum). We're updating it such that R can be used, to make it accessible to blind students.

I'm currently trying to figure out how to let blind students assess distribution 'shapes'. They can compute skewness and kurtosis, but it's also necessary to assess modality (how many 'mountains' there are). Is any of you familiar with any measures for this?

How do you normally assess the 'shape' of a distribution?

Thank you very much in advance, kind regards,

Gjalt-Jorn Peters
Dutch Open University