}

> -----Original Message-----
> From: medstats@googlegroups.com
> [mailto:medstats@googlegroups.com] On Behalf Of Thomas Keller
> Sent: 10. november 2009 10:44
> To: MedStats
> Subject: {MEDSTATS} Bland-Altman-Plots in case of
> duplicates/replicates

> Dear community,

> I'm wondering if there are specials aspects to consider if
> Bland- Altman plots are drawn for experiments when
> measurements are performed in duplicates (or more replicates)
> per method.
> One could use means, of course, but in that case information
> about precision is lost, and differences could become smaller
> in average.
> But regarding single measurements, it is not possible to draw
> a point for each single replicate, because they could be
> considered to be random.

> I look forward to your helpful comments,

> Thank you Thomas

> *****************
> Th. Keller, www.acomed-statistik.de
> Leipzig, Germany

> This problem is treated in:

> @Article{Carstensen.2008b,
>   author =       {B Carstensen and J Simpson and LC Gurrin},
>   title =        {Statistical models for assessing agreement in method
>   comparison studies with replicate measurements},
>   journal =      {International Journal of Biostatistics},
>   year =         {2008},
>   volume =       {4},
>   number =       {1},
>   pages =        {Article 16}}

> freely available ashttp://www.bepress.com/cgi/viewcontent.cgi?article=1107&context=ijb

> best regards,
> Bendix Carstensen

> > -----Original Message-----
> > From: medstats@googlegroups.com
> > [mailto:medstats@googlegroups.com] On Behalf Of Thomas Keller
> > Sent: 10. november 2009 10:44
> > To: MedStats
> > Subject: {MEDSTATS} Bland-Altman-Plots in case of
> > duplicates/replicates

> > Dear community,

> > I'm wondering if there are specials aspects to consider if
> > Bland- Altman plots are drawn for experiments when
> > measurements are performed in duplicates (or more replicates)
> > per method.
> > One could use means, of course, but in that case information
> > about precision is lost, and differences could become smaller
> > in average.
> > But regarding single measurements, it is not possible to draw
> > a point for each single replicate, because they could be
> > considered to be random.

> > I look forward to your helpful comments,

> > Thank you Thomas

> > *****************
> > Th. Keller,www.acomed-statistik.de
> > Leipzig, Germany

> -----Original Message-----
> From: medstats@googlegroups.com
> [mailto:medstats@googlegroups.com] On Behalf Of Thomas Keller
> Sent: 10. november 2009 10:44
> To: MedStats
> Subject: {MEDSTATS} Bland-Altman-Plots in case of
> duplicates/replicates

> Dear community,

> I'm wondering if there are specials aspects to consider if
> Bland- Altman plots are drawn for experiments when
> measurements are performed in duplicates (or more replicates)
> per method.
> One could use means, of course, but in that case information
> about precision is lost, and differences could become smaller
> in average.
> But regarding single measurements, it is not possible to draw
> a point for each single replicate, because they could be
> considered to be random.

> I look forward to your helpful comments,

> Thank you Thomas

> *****************
> Th. Keller, www.acomed-statistik.de
> Leipzig, Germany