# [Rasch] Fwd: Reliability, separation and strata

Fabio La Porta fabiolaporta at mail.com
Wed Mar 1 20:35:29 AEDT 2017

```Hi Mike,
thanks for your explanation, I have understood now.

Just the last question: the Ben Wright's method can work also on a
distribution of persons approaching normality?

I am asking this because I like very much the possibility to define
graphically the ranges of the statistically different levels of performance
along the measurement continuum.

On the other hand, I have always find it difficult to draw strata on the
measurement continuum, as they have, almost always, decimal values, making
the whole process rather inaccurate.

Thanks
Fabio

2017-03-01 10:15 GMT+01:00 Mike Linacre <mike at winsteps.com>:

> Fabio:
>
> G and H are described at http://www.rasch.org/rmt/rmt94n.htm
> and http://www.rasch.org/rmt/rmt163f.htm
> - they both depend on the sample of persons, and expect the sample to have
> a normal distribution.
>
> Ben Wright's method at http://www.rasch.org/rmt/rmt144k.htm
> is independent of the sample of persons. It is not G or H.
> Ben Wright suggests that if you need G or H for a skewed sample, then
> resample a normal distribution of persons from the skewed sample. See
> "Later Note" at the bottom of http://www.rasch.org/rmt/rmt144k.htm
>
> OK?
>
> Mike L.
>
>
> On 3/1/2017 6:07 PM, Fabio La Porta wrote:
>
>> Hello Mike,
>>
>> I understand that 1) is the formula for the separation ratio (G), but I
>> do not understand the formulas for 2) and 3).
>>
>> Are these related to the calculation of the reliability index starting
>> from H [(4G+1)/3 ?] and, respectively, the method suggested by Ben Wright
>> for skewed distributions in http://www.rasch.org/rmt/rmt144k.htm
>> ?
>>
>> Also, the above method for skewed distributions allows to calculate G or
>> H?
>>
>
> ________________________________________
> Rasch mailing list
> email: Rasch at acer.edu.au
> web: https://mailinglist.acer.edu.au/mailman/options/rasch/fabiol
> aporta%40mail.com
>