Primers for Modeling in SASci 2014 session and Wiley Companion

From InterSciWiki
Jump to: navigation, search

Using Dow-Eff software at http://socscicompute.ss.uci.edu

INSTRUCTIONS FOR WILEY COMPANION AND SASci ABQ CONFERENCE MODELING FOR SCCS (Murdock-White data codebook 1969-2014) using the CoSSci online Gateway. Douglas R. White (talk) 09:38, 7 January 2014 (PST) -- these instructions are an update of orientations for a fall 2013 class taught by Assistant Professor Ren Feng at Xiamen University, previously a visiting Predoctoral fellow at UCI who TAed, along with Tolga Oztan, the fall 2011 course Human Social Complexity - Cultural Consequences taught by DRW using the Standard Cross-Cultural Sample database with the 2011 version of [software for R gui]. These instructions are for 2014 CoSSci DEf software. Any instructor is welcome to use this site for a class, please contact drwhite@uci.edu so you get help with setup.

To use the current version of Dow-Eff functions (DEf), viewers, researchers, students or instructors need to sign-in and remember a user name and password to be used in following the instructional video How-tos and guides: YouTube How to share histories and how to access histories shared by other users at the home site http://socscicompute.ss.uci.edu. It is from one or more of these histories that anyone can learn to use the CoSSci web site shown at the http://intersci.ss.uci.edu/ Main Page. For a preview of what a finished model might look like, the [Visual Manual] (VM) at that site lets students and visitors see what a finished network-of-variables example looks like. Clicking the VM1-VM2-VM3-VM4 examples helps to see four interlinked modeling results discussed -- on pages 18-21 -- in a still early version of Chapter 5 of a Wiley Companion to Cross-Cultural Research that can be downloaded or viewed at http://tinyurl.com/WileyCh5CCRNetsofVars. Click here to see Background to Chapter 5, which uses http://socscicompute.ss.uci.edu, with SCCS, the Standard Cross-Cultural Sample, for the modeling. Other databases for modeling include EA (Ethnographic Atlas), LRB (Lewis R. Binford's forager data), WNAI (from Jorgensen's Western Indians book), and XC (371 societies in the Atlas that merge all data from SCCS, EA, Binford, WNAI datasets). See the four codebooks for EA SCCS LRB WNAI.

Screen Shot 2014-01-08 at 11.00.19 AM.png
Screen Shot 2014-01-08 at 11.22.27 AM.png
Screen Shot 2014-01-08 at 11.28.30 AM.png
For understanding outcomes of DEf modeling, students can click the VM1-VM2-VM3-VM4 pages, copy to their PC/Mac the data-entry variable numbers, and enter them in the http://socscicompute.ss.uci.edu SCCS option windows. You can ask for maps for one or more of your variables by naming them in the maps window below the variable windows at the right of this paragraph. The model runs in 2 minutes, and may yield bad input variables that turn the screen red, or may save successful outputs 1) for maps, if any, 2) names of societies, and 3) a *.csv file with model results where the dependent variables are listed along with independent variable predictors. You then have to locate where the *.csv Results are saved to your computer, open the *.csv file, save as an *.xls file that you can give names that identify the DepVar and IndepVar or you can identify them by their Galaxy output numbers. To Save Histories alongside the work of other students is part of class assignments and allows others to learn modeling strategies. The user needs to have signed in, rename the history file by clicking and renaming the title at the upper right and then use the upper menu bar to save a continuous series of resultant models in a permanent history file. From these outputs, students, instructors or visitors can print their maps, study and edit their *.xls files and discuss them in class with the instructor.

For the 13 students in Ren Feng's fall 2013 class, results are saved in (click) Published_Histories, a permanent online site. By following the steps taken from the low-numbered Galaxy files to the last model, clicking the option for output SCCS h, a newcomer to DEf can trace the decisions make by a given student or researcher. DRW's favorite student model at the is by jiangchan-v741 (go to model 179: SCCS h and click to open the Galaxy *.csv), where v741 is the SCCS widow remarriage variable. She and four others received As in the class eight had Bs and 1 had a C.

Running iterations of a model allows an easy path, in the *csv output, to viewing ToTry and DidWell to improve or stabilize the model. In need to explain where to find these lists of variables in the Galaxy *.csv output. Here is what that part looks like in the csv:

totry <---- this is a ToTry variable. There may be one, several or many. (Mostly not in the current model). x 1 AnimXbwealth:v204 2 EvoGodSq 3 sq149 4 sq149Sq 5 v149Sq didwell <---- DidWell variables (those in the current model) x 1 AnimXbwealth 2 bio.8 3 SuperjhWriting 4 v149 5 v204

Past Commentaies

Comments65Ren Feng's course fall 2013 Xiamen
Comments77
Comments104 
Other Offloads
Comments34 Ignore preliminary notes on using R gui

Possibilities for AICc R software for Networks-of-Variables

Comments129