# Indep/Depvar list - All

Project: The Anthropology of Causality -- Cultural Consequences of Regionally Fluctuating Inequality -- Edu-Mod 2009-10: The Individual Studies

- SCCS R package
- White, White, Ren, and Oztan 2010 Causal Inference for Multilevel Networks of Early Ethnographically Well-Described Populations
**YOU MUST USE FIREFOX FOR THE SVG graphs!!!**

## Contents

## svg Causal graphs network Mar_15_2010

- pvar < .06. Showing Language and Distance effects for the dependent variables. The sizes of nodes show the extent of spatial clustering. Not shown are the negative versus positive or nonsignificant linguistic clustering. The colors show the layers of the directed asymetric graph (DAG).
- pvar < .06. Showing Language and Distance effects 2 for the dependent variables. Here the dependent variables begin at the two bottom rows, coded yellow for significant negative linguistic effects, green otherwise, and sizes of those nodes (which were dependent variables) showing relative significance of spatial clustering. This graph also minimizes the length of lines between levels. 10 dependent variables are at the lowest levels, 2 at the next level up, 1 at the third level, and only 3 independent variables at the top level. All features of the analysis are then shown in the causal graphs, except the regression coefficients as labels for the lines.

Each downward line in the graphs has a signed regression coefficient^{[1]} from two-stage OLS using the software of Eff, E. Anthon, and Malcolm Dow. 2009. *How to Deal with Missing Data and Galton's Problem in Cross-Cultural Survey Research : A Primer for R*, in * Structure and Dynamics: eJournal of Anthropological and Related Sciences* 3#3 art 1. Spatial and linguistic clustering effects are estimated in stage one and used to make a new estimated Galton-effects variable that is included in stage two along with other independent variables for which the causal effects regression coefficients are estimated. The statistical diagnostics test whether the error terms are now free of language and distance autocorrelation, heteroscedasticity (bunching of error terms), variables with significant effects that are not specified in the final model but which reside in the set of initial variables, or whether there are variables among those specified that would be more predictive when logged.

- ↑ The unstandardized regression coefficients for each of the arrows are shown only for the triangular graph #6 to show an example of calculation of both direct and indirect effects according to the
*causal graphs*methods of Judea Pearl. The magnitude of causal effects of the independent variables for each dependent variable can eventually be calculated when the fuller structure of these graphs, taking each independent variable as a dependent one, is estimated.

The independent variables with the most dependent variables effects are these:

- Moral gods (4 effects), 3 strongly significant
- Population density (3 effects), 2 strongly significant
- Food trade (3 effects), 2 strongly significant

## svg Causal graphs network Mar_9_2010 (updated)

http://bit.ly/98IC5J Causal graphs from cross-cultural research - subhead

**Viewable only in Mozilla firefox, these four causal graphs**have black solid lines for positive regression coefficients between independent vars and dependent vars (downward lines), and red dotted lines for negative coefficients.- The thickness of the lines reflects significance, starting at pvar <.06 (thin lines) up to pvar <.0003 (thick lines).

- 3. pvar < .06 - you can click the square boxes for URLs that provide background information. The four levels, which form a directed acyclic graph (DAG), represent, first, variables that appear only as predictors, then ones that are both dependent (from above) and independent (to the level below) in successive levels. Colors and sizes of nodes are those of graphs 1 and 2 above.

- 4. pvar < .06 bicomponent in 46: one transitive subgraph.

- 5. pvar < .06 bicomponent layers with one transitive subgraph - can you see it?

- 6. pvar < .06 showing how to compute causal effects that are both direct and indirect. This example shows how Causal relations are not functionally consistent, as in a graph where signs are balanced (positive products of signs in a closed circuit).

My perception is that everything in graphs 1-5 has a structure consistent with a DAG structure that is computable for causal graphs (for each dependent variable) according to Pearl. Is this correct? The regression coefficients are computed in the second stage of the 2SLS, with the endogenous variables (spatial and language clustering coefficients) estimated in the first stage for use as predictors in the second stage.

## svg network 3_3_10 all obsolete

- clickable svg of new network of findings by project Working69Copy2 seems like the left wing of results deal with
**sex, rape, interpersonal violence, police, control of dvellings have one them**then the right wing deals with**frat-int-groups, stress on resource, war/fighting, wealth/poor differences, (low fe-)male agriculture, and individual freedom to choose a spouse**. These intersect in the middle variable, money. What is needed now is to put red links for negative relations, and thicker links for greater significance. Because there are so many links,**its probably necessary in order to estimate causal effects to eliminate all but the most significan variables, e.g., pvar < .01**. - clickable svg of new network of findings by project WorkingCopy71 - obsolete
- clickable svg of new network of variables only ReducedCopy52
- clickable svg of cohesive network of variables ReducedBicomponent52
- clickable svg of cohesive variables only VarBi-comp30
- Causal SCCS network 3500 nodes

## Maps

- back to Day 17 - back to Day 18 - Final Models and Commentary - building the partial network of results. Click SQUARE nodes for map(s) of the dependent variables and Edu-Mod sites and Edu-Mod 2009: The Individual Studies.
- http://edumodgis.ss.uci.edu/SCCS1/default.aspx Fall 2009 maps
- http://edumodgis.ss.uci.edu/SCCS1_v6/default.aspx Winter 2010 maps
- Wikipedia:Standard cross-cultural sample#Cultures_in_the_standard_cross-cultural_sample data on individual societies
- GIS maps must be accessed in FIREFOX MOZILLA, GOOGLE CHROME ETC NOT I-EXPLORER. INSTRUCTIONS: Unclick |_| SCCS1. Or use http://edumodgis.ss.uci.edu/SCCS1_v6/default.aspx Winter 2010 maps Then click Queries and your variable number, and check the SCCS codebook for the category you want mapped. Then click that Query category for that variable to see stickpins for the societies in that category. To paste the map into Word, use Ctrl-PrintScreen, paste into word, click the image, then "Picture" then the "Crop" icon, use the siderails on the edges of the picture image to crop the image.

## Map requests

**If your variable is not in the Query list then add to one of the three lists below ACCORDING TO YOUR EDUMOD number:**

- Indep/Depvar list EduMods 1-15
- Indep/Depvar list EduMods 16-25
- Indep/Depvar list EduMods 26-all other

- New requests -- see also White-Veit SCCS Atlas and SCCS Maps in Spss (others available on request)
- Keep order of variables numeric!! Make a space before each variable number

v17 EM-10dv money v26 EM-23 bodycontact (bodily contact-early infancy) v33 EM M-6 pain infliction v54 EM M-6 father (role of) v72 EM-26iv exogamy (intercommunity marriage) v80 lrgfam - not yet mapped v143 - mapped but not used (laungering div of labor) v153 EM-18iv techspec (technological specialization) (corpun was an incorrect label)

- v453 EM-18iv corpun - you specified 153 by mistake. Thus, is not on the mapping site.

BIG CORRECTION Doug 13:53, 30 November 2009 (PST) to set the record straight: When you switched to depvarname<="wifebeating", Hiu Kwan. which is v453 you asked for v153 by mistake. To set the record straight since v153 was TECHNOLOGICAL SPECIALIZATION I made two changes: one to let polispec be v153 and then to let wifebeating be v453. Then I reran your results.

v156 EM-27iv etc popdens v157 POLITICAL INTEGRATION (not used by anyone) v167 EM-24dv pre_mar_sex v169 EM-20dv v169 EM-28dv extramaritalsex v203 EM-27 v203 EM-24iv etc gath v204 EM-18iv etc hunt v205 EM-24iv etc fish v208 EM-27iv brideprice (hi values: dowry) v227 EM-24iv exogamy v232 EM-24 v232 EM-26iv etc cultints (intensive cultivation) v233 EM-18 v233 EM-24 v233=6 EM-26 etc cereals v234 EM-20iv etc settype (settlement type) v236 EM-26iv etc localjh v237 EM-20iv etc superjh v238 EM-21iv etc moralgods v239 EM-17dv strategy (games) v242 agrlateboy - not yet mapped v243>1 EM-24iv etc plow v244 EM-26iv v244==2 EM-27iv etc pigs v244==4 EM-29iv etc bovines v245>1 EM-24iv milk v300 EM-6 v300 EM-26iv segadlboys v591 EM-29dv female control over dwelling v661 EM-11iv v661 EM-13iv v661 EM-41iv fempolpar v662 EM-13iv femsolgro - Sanday) v662 EM-41iv,femalesol - Sanday=femsolgro) v663 EM-26 fempower v664 EM-40 maletough ideomaletough v666 EM-28 v666 violence intervio v667 EM-11 rape 0v677 migr (no predictions) v678 EM-13dv foodstress (food stress or hunger) v678 EM-21dv stress (Food Stress or Hunger - Sanday) v679 warfight (Warfare or Fighting - Sanday) v693 War&Fighting (Frequency of Intercommunity Armed Conflict - Sanday) v740 EM-26iv fmargmar (marriage arrangements) v749 EM-18iv inheritance v754 EM-10dv wifebeating v754 EM-18iv (but changed depvar) v819 EM-10 v819 EM-24 v819 EM-26 foodtrade v838 EM-12iv v838 EM-29iv dateobs v857 EM-13 v857 EM-24 etc ecorich v857 EM-29iv v857 EM-10 v872 EM-26 v872 EM-20 v872 EM-17 iv nuclearfam v210=3 EM-23 v890 EM-29 femsubs v1260 EM-27 pathstress (pathogens) v1648 EM-6iv etc v1648 EM-29 war v1675 EM-12 homicide v1684 EM-13 pctFemPolyg v1685 EM-21 foodtrade v1710 EM-23 freintovio (freq or occasional interpersonal violence) v1721 EM-27dv wealth (wealthy) v1764 EM-16dv react2viol

Not mapped but an iv

v56 prinrelcaretakeryoung v270 stratif v1684 weatherpest v1734 mkt

SORTED BY EM

v1648 EM-6 v300 EM-6 v17 EM-10 v857 EM-10 v819 EM-10 v667 EM-11 v838 EM-12 v1675 EM-12 v857 EM-13 v678 EM-13 v1684 EM-13 v239 EM-17 v204 EM-18 v153 EM-18 v233 EM-18 v749 EM-18 v872 EM-20 v237 EM-20 v234 EM-20 v169 EM-20 v678 EM-21 v1685 EM-21 v238 EM-21 v26 EM-23 v1710 EM-23 v167 EM-24 v205 EM-24 v232 EM-24 v227 EM-24 v245 EM-24 v203 EM-24 v857 EM-24 v233 EM-24 v243 EM-24 v819 EM-24 v244 EM-26 v72 EM-26 v236 EM-26 v300 EM-26 v233 EM-26 v740 EM-26 v819 EM-26 v232 EM-26 v872 EM-26 v663 EM-26 v1721 EM-27 v1260 EM-27

v156 EM-27 v208 EM-27 v203 EM-27 v244 EM-27 v890 EM-29 v244 EM-29 v1648 EM-29 v591 EM-29dv v838 EM-29iv v857 EM-29iv v54 EM M-6 v33 EM M-6 v664 EM-40 v157 v661 v662 v666 v679 v693 v754 v857 v872 1764