Combine radiocarbon dates from the same archaeological level

[Magdalena Gómez]

 Hi, maybe this is a very simple question...

How could I know which radiocarbon dates from the same level should I combine? What's the best option to find out that?

This is my case study:

 Plot()

 {

  R_Combine("")

  {

   R_Date("q GrA-27846",11240,60);

   R_Date("q GrA-27841",11640,60);

   R_Date("q GrA-27843",11780,60);

   R_Date("q GrA-22087",11980,80);

   R_Date("q GrA-24296",12060,60);

   R_Date("q GrA-22089",12500,90);

  };

 };

Thank you very much,

Magda

[Rayfo...@aol.com]

Hi Magda,

 

R_Combine is used to combine two (or more) radiocarbon determinations that are from the same source so that the merged material is from the same radiocarbon reservoir, for instance two bones from the same body.  The merging is carried out before the calibration.  A Chi Sq test is carried out to determine if the data can be merged, or if there is some problem with one or other of the determinations, possibly lab error or contamination.

Warning! X-Test fails at 5% - X2-Test: df=1 T=4.027(5% 3.8)

 

'T' must be less than the 5% figure to pass.

 

On the other hand, if you believe that the determinations are co-eval, but not from the same radiocarbon source, 'Combine' is a group function that combines any number of PDFs which all give independent information on a parameter.  It is carried out after calibration.  A Chi Sq test is carried out, however the test

parameter to examine is Acomb:

 

Warning! X-Test fails at 5% - X2-Test: df=1 T=15.921(5% 3.8)
Warning! Poor agreement - n=2 Acomb= 1.9%(An= 50.0%) 

 

'Acomb' must be greater than An to pass.

 

However, there is much more to it than simply combining likely data in a layer.  What is the material?. Is it charcoal? Is it intrusive to the layer.  Does it actually date the element of interest?

 

Why do you wish to merge the data?  It may reduce the error range, but leave you with a single date, when several dates in a Phase with boundaries can provide a more meaningful scenario for the layer. etc etc.

 

Best wishes

 

Ray (non expert)

[Erik]

Magda,

I have often run across this and been tempted to use combine. But as Ray pointed out, this is not usually appropriate.

If all of your dates are from the same level, my suggestion is to group them as a phase. You'll want to use boundaries, for example:

Sequence()

{

Boundary("Start");

Phase()

  {

   R_Date("q GrA-27846",11240,60);

   R_Date("q GrA-27841",11640,60);

   R_Date("q GrA-27843",11780,60);

   R_Date("q GrA-22087",11980,80);

   R_Date("q GrA-24296",12060,60);

   R_Date("q GrA-22089",12500,90);

  };

Boundary("End");

};