Coding Afters

[Charles Lindsay]

I have a question about using the After command.

In the following extract from a model:

 Sequence("House F")
     {
      After("Construction")
      {
       R_Date("T-531", 950, 50);
       R_Date("S-1102", 1130, 135);
       R_Date("S-1104", 1630, 105);
      };
      Boundary("S: House F");

does OxCal treat all three dates as "afters" or just the first date (T-531)?

[Christopher Ramsey]

Charles

It should make all of those dates to be TPQs for the Boundary.

The only problem with doing this is that the dates are then assumed to be entirely independent from each other - which means the more of them there are the later it will push the Boundary.

For this reason it would be better to model this as either:

Sequence("House F")
{
Boundary("Start construction");
Phase"Construction")

{
R_Date("T-531", 950, 50);
R_Date("S-1102", 1130, 135);
R_Date("S-1104", 1630, 105);
};
Boundary("S: House F");

or:

Sequence("House F")
{
Tau_Boundary("Start construction");
Phase("Pre-Construction")

{
R_Date("T-531", 950, 50);
R_Date("S-1102", 1130, 135);
R_Date("S-1104", 1630, 105);
};
Boundary("S: House F");

The first model would be appropriate if the dates were all part of a coherent phase. The second is better if the distribution of dates is in the form of a long tail of residual material.

It is very unlikely that the second model here will be significantly different to the one you gave - but if the number of dates in the After command was high - and if they were tightly packed up against the end date it would make a difference and the above approach would likely be better.

Best wishes

Christopher

--
You received this message because you are subscribed to the Google Groups "OxCal" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oxcal+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oxcal/0fde42c2-c1b2-477c-85f3-440dae763dfdn%40googlegroups.com.

[Charles Lindsay]

Christopher:

Thank you so much for the information and  the advice. I will run the model again with both the suggested modelings and see what, if any, differences arise in the outputs.

Cheers

Charles