Using deep learning to predict ideology from facial photographs: expressions, beauty, and extra-facial information.

Deep learning techniques can use public data such as facial photographs to predict sensitive personal information, but little is known about what information contributes to the predictive success of these techniques. This lack of knowledge limits both the public's ability to protect against revealing unintended information as well as the scientific utility of deep learning results. We combine convolutional neural networks, heat maps, facial expression coding, and classification of identifiable features such as masculinity and attractiveness in our study of political ideology in 3323 Danes. Predictive accuracy from the neural network was 61% in each gender. Model-predicted ideology correlated with aspects of both facial expressions (happiness vs neutrality) and morphology (specifically, attractiveness in females). Heat maps highlighted the informativeness of areas both on and off the face, pointing to methodological refinements and the need for future research to better understand the significance of certain facial areas.

Cognitive ability is a powerful predictor of political tolerance.

OBJECTIVES: Despite the broad appeal of abstract notions of political tolerance, people vary in the degree to which they support the political rights of groups they dislike. Prior research highlighted the relevance of individual differences in the cognitive domain, claiming the application of general tolerance ideals to specific situations is a cognitively demanding task. Curiously, this work has overwhelmingly focused on differences in cognitive style, largely neglecting differences in cognitive ability, despite compelling conceptual linkages. We remedy this shortcoming. METHODS: We explore diverse predictors of tolerance using survey data in two large samples from Denmark (N = 805) and the United States (N = 1,603). RESULTS: Cognitive ability was the single strongest predictor of political tolerance, with larger effects than education, openness to experience, ideology, and threat. The cognitively demanding nature of tolerance judgments was further supported by results showing cognitive ability predicted tolerance best when extending such tolerance was hardest. Additional small-sample panel results demonstrated substantial 4-year stability of political tolerance, informing future work on the origins of political tolerance. CONCLUSIONS: Our observation of a potent role for cognitive ability in tolerance supports cognitively oriented accounts of tolerance judgments and highlights the need for further exploration of cognitive ability within the political domain.

A Major Limitation of the Direction of Causation Model: Non-Shared Environmental Confounding.

Determining (1) the direction of causation and (2) the size of causal effects between two constructs is a central challenge of the scientific study of humans. In the early 1990s, researchers in behavioral genetics invented what was termed the direction of causation (DoC) model to address exactly these two concerns. The model claims that for any two traits whose mode of inheritance is sufficiently different, the direction of causation can be ascertained using a sufficiently large genetically informative sample. Using a series of simulation studies, we demonstrate a major challenge to the DoC model, namely that it is extremely sensitive to even tiny amounts of non-shared confounding. Even under ideal conditions for the DoC model (a large sample, N = 10,000), a large causal relationship (e.g., a causal correlation of .50) with very different modes of inheritance between the two traits (e.g., a pure AE model for one trait and a pure CE model for another trait) and a modest degree (correlation of .10) of non-shared confounding between the two traits results in the choice of the wrong causal models and estimating the wrong causal effects.

A nationwide registry-based cohort study of incidence of tonsillectomy in Denmark, 1991-2012.

OBJECTIVE: To update tonsillectomy incidence rates in Denmark and identify whether the incidence rates vary between geographical areas in the country during the period 1991-2012. DESIGN: This was a retrospective nationwide cohort study using data from the comprehensive Danish patient registries. Data included surgical code, gender and date of birth, date and place of surgery. SETTING: Hospitals and private oto-, rhino-, laryngology (ORL) offices. PARTICIPANTS: Danes who underwent tonsillectomy in the period 1991-2012. MAIN OUTCOME MEASURE: Tonsillectomy. RESULTS: In the 22-year period, 177 211 tonsillectomies were conducted, and the overall incidence of tonsillectomy decreased significantly over time. The overall annual incidence of tonsillectomies decreased from 155.7 per 100 000 inhabitants in 1991 to 129.4 per 100 000 inhabitants in 2012. In 1991, 5.5% of tonsillectomies were performed in office settings, while in 2012, it had increased to 26.6% (P < .01). The highest incidence of tonsillectomy was found in children between 3 and 4 years of age (predominance of males, P < .01) and in 15- to 19-year-olds (predominance of females, P < .01). Each year, 0.44% of children underwent tonsillectomy, and by the age of 20, 7.7% of the Danish people has had a tonsillectomy. There were significant regional differences in the number of tonsillectomies. CONCLUSION: The incidence rate of tonsillectomies in Denmark decreased significantly in the study period, but with great regional variance.