Comments95

From InterSciWiki
Jump to: navigation, search

Offloads Comments95 October 13, 2013 Doug: Attached is the crosswalk between AWC and EA. There are 560 matches (there are 561 societies in the csv file, but the Swazi don't have an AWC code). If you have additional societies to code, you can add them to this crosswalk and send it back to me.

EA and EAfact in the workspace have been accordingly modified. To use the AWC 560 subsample of the EA: 
load(url("http://dl.dropbox.com/u/9256203/DEf01c.Rdata"),.GlobalEnv)
setDS("EA")
z<-which(!is.na(dx$awc))
dx<-dx[z,]
dxf<-dxf[z,]
wdd<-wdd[z,z]
wll<-wll[z,z]
wee<-wee[z,z]

October 11, 2013

  • Doug: I looked at it again, changed the double letters to single (which didn't affect the count, since I had included them before, with incorrect AWC code). After dropping all those with missing AWC codes, I come up with 561. There are two pairs of societies with the same AWC code (at the top of the attached csv file) so there are two more that need to be dropped (you need to choose). That will give a total of 559.
  • AWC - AWC_Letter_Codes New and old AWC Codes. The original version had 558 societies.

October 8, 2013

  • Sorry. I forgot to also adjust the weight matrices. The code you will need is as follows:
load(url("http://dl.dropbox.com/u/9256203/DEf01c.Rdata"),.GlobalEnv)
setDS("EA")
z<-which(!is.na(dx$awc))
dx<-dx[z,]
dxf<-dxf[z,]
wdd<-wdd[z,z]
wll<-wll[z,z]
wee<-wee[z,z]


October 6, 2013

I added a field called "awc" to EA and EAfact in the workspace DEF01c. To use the AWC subset of EA, do the following: from DEf01c only

load(url("http://dl.dropbox.com/u/9256203/DEf01c.Rdata"),.GlobalEnv)
setDS("EA")
dim(dx)
dx<-dx[which(!is.na(dx$awc)),]
dim(dx)
dim(dxf)
dxf<-dxf[which(!is.na(dxf$awc)),]
dim(dxf)

The number of rows in dx should fall from 1265 to 558.


Oct 9 2013 synthesis of the AWC from the EA

On 10/9/13 4:58 PM, Douglas White wrote: http://intersci.ss.uci.edu/wiki/index.php/AWC_Letter_Codes Four end in aa (twice) or bb (twice). If those are corrected to end in a or b they can go back in -- unless you can explain why they didnt match the EA. Four are simply missing from the list of AWC codes. These are:

A08e Hehe A10c B. Kavirondo A16e Temne C01c Arusi N08e Wishram Must Correct 4 A09aa* Chagga E19bb* Lakher N24aa* Aztec S15bb* Mundurucu total is your 558+8=566 which seems one too many so there should be at least one that should be thrown out.

Tolga and I are now able to do (1) the study of avoidance behaviors (25 dyadic variables, about a dozen which have sufficient frequency to include) (2) Tolga's thesis on foragers (3) our paper for the AAA in November, (4) a paper at ABQ.

Your help meant a great deal to us. If you can fix any of the remaining eight closures between the two samples (given the name variation I can easily verify identities of the societies from pinpointing sheets, EA listings, and the AWC books) then I can have a definitive list and code the avoidance data and several other features from Murdock's partly unpublished kinship coding sheets for which I have the only original (have sent to Mike Fischer, Dwight Read and anthropologists at AMNH). Point is that if not now, these data will never appear in the cross-cultural coded datasets, they are truly unique and highly valuable. And Chagga, Aztec, Mundurucu, Lakher, Arusi etc are classical ethnographies.

BTW, hearing from Fischer, I think my proposal has been accepted for HRAF to compile a searchable retrieval of text sample list for those of the SCCS that are archived for eHRAF, and also one of their 60 case "random" samples 2/3rd of which are in the SCCS and all but one of the other 1/3 have reasonable SCCS substitutes without altering the coverage of the sample in terms of types of societies.

-- many thanks

Douglas R. White http://intersci.ss.uci.edu/wiki/index.php/DW_home