Extended Data Fig. 4: Tie social identity determinants of respondent accuracy.
From: Cognitive representations of social networks in isolated villages
We find that characteristics related to the social identity of a pair of individuals (i and j) affects how well that tie is conceived of by individuals k. (a-d) LHS, marginal effects on accuracy in ROC-space. Grey shading represents the 95% bootstrapped confidence ellipse of the predictions from the two models. RHS, marginal effect of each individual accuracy measure: the true positive and false positive rates and the summary measure, Youden’s J. Intervals represent 95% confidence levels, calculated via normal approximation for the two rates, and bootstrapped for J. (a) Religion combination of tie members. (b) Indigenous status of the pair. Parameters are fit from separate models of each rate, conditional on tie verity in the reference network. (c) Absolute difference in wealth between the tie members. (d) Average wealth of the tie members. (e) Interaction between the average wealth of a pair and the cognizer’s wealth on the (LHS) TPR and (RHS) FPR. (f) Interaction between the average wealth of a pair and the cognizer’s wealth on the summary measure, J. See Methods for details of model specification. Results are from n = 9,998 survey respondents in both panels, corresponding to 177,928 individual responses for the TPR estimates, and 477,393 responses for the FPR estimates.
