Charcoal 14C Dates from One Structure

[Eun Chae]

Hello everyone, I would like to ask for some feedback regarding the processing of several C14 dates in OxCal. I’ve only recently started reading papers on radiocarbon dating, so I may be lacking in some areas. I apologize in advance and thank you for your understanding.

I have reviewed the results of several charcoal samples from the same house structure.

I was not involved in the measurements myself, and have only been able to check the dates measured by institutions listed in the excavation report.
(Although the samples came from the same house structure, it is difficult to determine whether they originated from the same piece of wood. That said, it does seem that the leftover sample from the A institution’s radiocarbon dating was remeasured by institution B. Unfortunately, the report was unclear, so I could not confirm further details.)

As an example, the following uncalBP dates were obtained: 

House_01 3540 ± 50 / Institution A

House_02 3430 ± 50 / Institution A

House_03 2935 ± 20 / Institution B

House_04 2950 ± 20 / Institution B

House_05 3045 ± 20 / Institution B

House_06 2970 ± 20 / Institution B

All of these are charcoal samples, measured by different institutions.

The house structure in question is difficult to date using conventional archaeological typology.

Assuming that each sample came from the same house and from the same material, I tried applying R_Combine. As a result, three groups were generated.

The date from House_05 does not combine with either of the other two groups. 

House_Group_01 3486 ± 36 → House_01, 02 

House_Group_02 3045 ± 20 → House_05

House_Group_03 2952 ± 12 → House_03, 04, 06

In this case, is it appropriate to use Combine? I’m unsure whether Outlier is suitable, since the range of date differences is quite large. Also, because I didn’t measure the samples myself, I hesitate to exclude any of the dates arbitrarily, as they were published in the excavation report.

Furthermore, I’m uncertain whether I should use R_Combine / Combine, or instead apply a Phase using Bayesian statistics.

If I’m making a beginner’s mistake or misunderstanding anything, I would truly appreciate your advice.

Thank you.