Multiple errors in SCCS variables.

SimonGreenhill commented 1 year ago

[x] 1. Marriage Arrangements [SCCS739]

D-Place Link: https://d-place.org/parameters/SCCS739#2/14.3/150.1

Note: Compare sample size and counts to source (see Broude and Greene, 1983, pp. 273-274)

[x] 2, Marriage Arrangements (Female) [SCCS740]

D-Place Link: https://d-place.org/parameters/SCCS740#2/14.3/150.1

Note: Compare sample size and counts to source (see Broude and Greene, 1983, p. 274)

[ ] 3. Inheritance of Property of Some Economic Value [SCCS590]

D-Place Link: https://d-place.org/parameters/SCCS590#2/22.2/145.5 Note: Compare sample size and counts to source (Whyte, 1978, p. 215).

[ ] 4. Inheritance distribution for real property (land) [Note, identical to EA075] [SCCS280]

D-Place Link: https://d-place.org/parameters/SCCS280#2/14.2/151.1

Note: Compare sample size and counts to source (appears in Divale, 2001, p. 46)

[ ] 5. Standard Polygamy Code (SCCS861).

D-Place Link: https://d-place.org/parameters/SCCS861#2/36.9/152.0

Note: Compare sample size and counts to source (appears in Divale, 2001, p. 128)

[ ] 6. Cousin marriages preferred [Note, identical to EA025] [SCCS228]

D-Place Link: https://d-place.org/parameters/SCCS228#2/16.6/150.1

Note: Compare sample size and counts to source (appears in Divale, 2001, p. 40)

xrotwang commented 9 months ago

@SimonGreenhill do you think we can fix this in the next couple of days? Or should we remove the questionable variables?

xrotwang commented 9 months ago

Re 1. and 2., i.e. the Broude and Greene data: Broude and Greene collected data for

186 societies and fifteen alternates

I checked the coding for SCCS740 in the source and all datapoints for the 186 societies are correct. What's missing in D-PLACE is the data for the 15 alternates. For SCCS739 the numbers of societies for each code are less or equal to the summary published in the paper. So I'd assume the issue is also the 15 alternates, inflating the counts in the paper.

xrotwang commented 9 months ago

Re 5.: The only discrepancy is two more societies listed under

3 Limited polygyny <20% of married males

We'll need to check the raw data.

xrotwang commented 9 months ago

Re 6.: In one edition of the SCCS codebook I found the following note:

228 HAS MAJOR DISCREPANCIES WITH PATRICK GRAY ETHNOATLAS RECODES

This might explain the discrepancies. (D-PLACE has 15 codes for SCCS228, whereas the codebooks I've seen only have 9.)

D-PLACE / dplace-dataset-sccs

Multiple errors in SCCS variables. #1