# Exploratory causal analysis for networks of ethnographically well-studied populations

From InterSciWiki

- 2010 for Submission. Scott D. White, Douglas R. White, Tolga Oztan, Ren Feng. Exploratory causal analysis for networks of ethnographically well-studied populations
- Cultural Consequences of Regionally Fluctuating Inequality (Evil eye) (Moralizing gods)

- Christian Brown and Tony Eff, 2010. The State and the Supernatural: Support for Prosocial Behavior, comments on Roes and Raymond's article Belief in Moralizing Gods, showing how ordinary methods of cross-cultural analysis lead to spurious results. They use instead the all-important approach (further refined in *Rccs*) 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. Structure and Dynamics: eJournal of Anthropological and Related Sciences 3#3 art 1.
- See also 2009 The Open Social Science Journal, 2009, 2, 70-73, Moralizing Gods and the Arms-Race Hypothesis of Human Society Growth, by Frans L. Roes, Lauriergracht 127-II, 1016 RK Amsterdam, The Netherlands
- Belief in Moralizing Gods

Scott D. White, Douglas R. White, Tolga Ozta, and Ren Feng: Article outline and grant proposals

## Contents

- 1 Introduction to estimation methods for societal databases
- 1.1 Identification problems in individual and societal level data
- 1.2 Two-stage estimation methods for societal databases
- 1.3 Ordinal probit regression
- 1.4 Application and choice of dependent variables
- 1.5 Finding variables for estimating effects in causal graphs, both direct and indirect
- 1.6 What is needed to estimate effects in causal graphs

- 2 Effects of observational nonindependence (endogenous “social effects”) and corrections for endogeneity
- 3 The regression model example: General Polygyny
- 4 Evaluation of Alternative Models
- 4.1 Evaluation of Paige and Paige’s theory of resource and reproductive politics
- 4.2 Evaluation of Whiting and Whiting’s (1975) theory for general polygyny
- 4.3 Evaluation of Low, Ember and Ember debility/mortality replacement theories
- 4.4 Evaluation of Boehm’s (1993, 1997a,b, 1999) theory for egalitarian resistance to local authority
- 4.5 Evaluation of Binford’s survey of foraging societies and polygyny

- 5 Evaluation of the causal graphs approach
- 5.1 Advantages of 2SLS regression, using Instrumental variables (IVs)
- 5.2 Advantages of multiple imputation (MI).
- 5.3 Extended Instrumental variables (XIV)
- 5.4 Regression Findings for General and Sororal Polygyny without MI and IVs
- 5.5 Diagnostics of fit
- 5.6 Avoidance of overfitting
- 5.7 Replication
- 5.8 Specification and Robustness
- 5.9 Evaluation of ordinal probit options
- 5.10 Limitations and corrections for incommensurate dependent variable sample size
- 5.11 Proposed refinements

- 6 Conclusions: Causal graphs
- 7 Appendix: Extended Instrumental variables (XIV) for Causal Graphs, by Halbert White
- 8 Grant proposal outlines

## Introduction to estimation methods for societal databases

### Identification problems in individual and societal level data

### Two-stage estimation methods for societal databases

### Ordinal probit regression

### Application and choice of dependent variables

### Finding variables for estimating effects in causal graphs, both direct and indirect

### What is needed to estimate effects in causal graphs

#### Ordinal probit and o^t

#### Needed for R release is package documentation: Scott: How do I do this?

## Effects of observational nonindependence (endogenous “social effects”) and corrections for endogeneity

##### Scott: For equations see Dow's article cited in that section, url in references

## The regression model example: General Polygyny

### Spatial and historical nonindependence in general polygyny

### 2SLS Findings for General Polygyny

### A validation test: is General Polygyny wealth-generating?

## Evaluation of Alternative Models

### Evaluation of Paige and Paige’s theory of resource and reproductive politics

### Evaluation of Whiting and Whiting’s (1975) theory for general polygyny

### Evaluation of Low, Ember and Ember debility/mortality replacement theories

### Evaluation of Boehm’s (1993, 1997a,b, 1999) theory for egalitarian resistance to local authority

### Evaluation of Binford’s survey of foraging societies and polygyny

## Evaluation of the causal graphs approach

### Advantages of 2SLS regression, using Instrumental variables (IVs)

### Advantages of multiple imputation (MI).

### Extended Instrumental variables (XIV)

### Regression Findings for General and Sororal Polygyny without MI and IVs

### Diagnostics of fit

### Avoidance of overfitting

### Replication

### Specification and Robustness

### Evaluation of ordinal probit options

### Limitations and corrections for incommensurate dependent variable sample size

### Proposed refinements

## Conclusions: Causal graphs

#### Pdf of the article (for Sociological Methodology and Research)

EduMod-72 Exploratory causal analysis for networks of ethnographically well-studied populations

## Appendix: Extended Instrumental variables (XIV) for Causal Graphs, by Halbert White

## Grant proposal outlines

### Sociology NSF outline 2010

- Why causality?
- Generality of the approach
- Educational value
- Reverse endogeneity
- Personnel

D White: SRAS, IMPRESS, SCCS, eJournals, vHumboldt, ASA economics, ASA networks, Eur.Physics S White T Oztan K Chalak Consultant: J Pearl Consultant: H White

### Anthropology NSF outline 2010

Title: Mathematical social science applied to anthropological data

- Roles of the mathematical social scientist
- Publically available anthropological data
- Ethnographies as a focus of study
- Ethnographic and related databases

SCCS, Atlas, foragers Archaeology Cross-polity

- Critiques

- Personnel

D White: SRAS, IMPRESS, SCCS, eJournals, vHumboldt, ASA economics, ASA networks, Eur.Physics S White J Pearl H White T Oztan