Hi Roger,
Only one thing I don't quite follow...
The tab command will produce Kendall's tau-b and approximate SE. However,
Sommers' D produces tau-a only.
How do tau-a and tau-b differ?
Regards,
Deidra
On 6/12/06 1:10 AM, "Newson, Roger B" <[email protected]>
> A possible measure of ordinal trend is Somers' D, given (with confidence
> limits) by the -somersd- package, which you can download from SSC. For
> instance, if Deidra types
>
> somersd autism_dx kerr_breath_score, tr(z) tdist
>
> then -somersd- will produce a CI for Somers' D of breath score with respect to
> autism, which is the difference between two probabilities, namely the
> probabiliry that a randomly-chosen autistic person has a higher breath score
> than a randomly-chosen non-autistic person and the probability that a
> randomly-chosen nonautistic person has a higher breath score than a
> randomly-chosen autistic person.
>
> More about Somers' D and -somersd- can be downloaded from my website (see my
> signature below).
>
> I hope this helps.
>
> Best wishes
>
> Roger
>
>
> Roger Newson
> Lecturer in Medical Statistics
> Respiratory Epidemiology and Public Health Group
> National Heart and Lung Institute
> Imperial College London
> Royal Brompton campus
> Room 33, Emmanuel Kaye Building
> 1B Manresa Road
> London SW3 6LR
> UNITED KINGDOM
> Tel: +44 (0)20 7352 8121 ext 3381
> Fax: +44 (0)20 7351 8322
> Email: [email protected]
> www.imperial.ac.uk/nhli/r.newson/
>
> Opinions expressed are those of the author, not of the institution.
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Ron�n Conroy
> Sent: 05 December 2006 13:13
> To: [email protected]
> Subject: Re: st: probability and z-statistic
>
> On 5 Noll 2006, at 13:59, Deidra Young wrote:
>
>> I just tried Kit's suggestion. One variable has two categories and
>> the
>> other has three. My test is to determine if there is a difference
>> between
>> cases with and without Autism Diagnosis (2) using an ordered severity
>> measure with three levels (none, mild or severe). Using your
>> suggestion, I
>> obtained the following result. Does this look like the p value is
>> calculated correctly?
>>
>> Deidra
>>
>>
>> . tab autism_dx kerr_breath_score, all
>>
>> | Kerr Score for Disturbed Awake
>> Initially Diagnosed | Breathing Rhythm
>> with Autism | None Mild Severe | Total
>> --------------------+---------------------------------+----------
>> No Autism Diagnosis | 75 109 76 | 260
>> Autism Diagnosis | 11 29 15 | 55
>> --------------------+---------------------------------+----------
>> Total | 86 138 91 | 315
>
>
>
> The test for an ordered association is probably more in line with
> your hypothesis.
>
>
> . binreg Autism breathscore, rr
>
> Iteration 1: deviance = 400.4562
> Iteration 2: deviance = 300.4687
> Iteration 3: deviance = 291.5691
> Iteration 4: deviance = 291.3885
> Iteration 5: deviance = 291.3884
> Iteration 6: deviance = 291.3884
>
> Generalized linear models No. of obs
> = 315
> Optimization : MQL Fisher scoring Residual df
> = 313
> (IRLS EIM) Scale parameter
> = 1
> Deviance = 291.3884049 (1/df) Deviance
> = .9309534
> Pearson = 314.8318309 (1/df) Pearson =
> 1.005852
>
> Variance function: V(u) = u*(1-u) [Bernoulli]
> Link function : g(u) = ln(u) [Log]
>
> BIC =
> -1509.167
>
> ------------------------------------------------------------------------
> ------
> | EIM
> Autism | Risk Ratio Std. Err. z P>|z| [95% Conf.
> Interval]
> -------------
> +----------------------------------------------------------------
> breathscore | 1.103058 .1807885 0.60 0.550 .7999973
> 1.520926
> ------------------------------------------------------------------------
> ------
>
> The risk of autism goes up modestly for each 1-unit increase in
> breath score severity, and it's not significant.
>
> You can, of course, just to a trend test
>
> . jonter breath, by(Aut)
> Jonckheere-Terpstra Test
>
> J = 7498
> J* = 0.57
> Prob > J* = 0.28532
>
>
> or
>
>
> . nptrend breath, by(Aut)
>
> Autism score obs sum of ranks
> 0 0 260 40732
> 1 1 55 9038
>
> z = 0.61
> Prob > |z| = 0.544
>
>
>
> =========
> Ron�n Conroy
> Royal College of Surgeons in Ireland
> [email protected]
> +353 (0) 1 402 2431
> +353 (0) 87 799 97 95
> http://www.flickr.com/photos/ronanconroy
>
>
>
>
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
