Two metamodels of causal effects

Two metamodels, termed Model S and Model V, are proposed for definition, measurement, and generalization of quantitative causal effects. The effect is defined as a part change in score in Model S and as a part change in variance in Model V. Two additional changes, total and remainder change, are defined. The latter is due to all other factors or variables than the cause, while total change is the sum of remainder and effect change. Furthermore, it is shown how contrafactual concepts, which imply that some parts of the study situation are supposed to be otherwise, enter into the metamodels. Casual effects are defined and measured in terms of non-contrafactual concepts, except that statistical induction includes contrafactual as well as non-contrafactual inferences. Non-statistical generalization involves both kinds of inferences. Contrafactual definitions are considered inadequate, and a contrafactual interpretation of statistical adjustment is unnecessary and should be replaced by a non-contrafactual one.     

Children's Ideas About What Can Really Happen: The Impact of Age and Religious Background

Five- to 11-year-old U.S. children, from either a religious or secular background, judged whether story events could really happen. There were four different types of stories: magical stories violating ordinary causal regularities; religious stories also violating ordinary causal regularities but via a divine agent; unusual stories not violating ordinary causal regularities but with an improbable event; and realistic stories not violating ordinary causal regularities and with no improbable event. Overall, children were less likely to judge that religious and magical stories could really happen than unusual and realistic stories although religious children were more likely than secular children to judge that religious stories could really happen. Irrespective of background, children frequently invoked causal regularities in justifying their judgments. Thus, in justifying their conclusion that a story could really happen, children often invoked a causal regularity, whereas in justifying their conclusion that a story could not really happen, they often pointed to the violation of causal regularity. Overall, the findings show that children appraise the likelihood of story events actually happening in light of their beliefs about causal regularities. A religious upbringing does not impact the frequency with which children invoke causal regularities in judging what can happen, even if it does impact the type of causal factors that children endorse.     

Motivated Reasoning in an Explore-Exploit Task

The current research investigates how prior preferences affect causal learning. Participants were tasked with repeatedly choosing policies (e.g., increase vs. decrease border security funding) in order to maximize the economic output of an imaginary country and inferred the influence of the policies on the economy. The task was challenging and ambiguous, allowing participants to interpret the relations between the policies and the economy in multiple ways. In three studies, we found evidence of motivated reasoning despite financial incentives for accuracy. For example, participants who believed that border security funding should be increased were more likely to conclude that increasing border security funding actually caused a better economy in the task. In Study 2, we hypothesized that having neutral preferences (e.g., preferring neither increased nor decreased spending on border security) would lead to more accurate assessments overall, compared to having a strong initial preference; however, we did not find evidence for such an effect. In Study 3, we tested whether providing participants with possible functional forms of the policies (e.g., the policy takes some time to work or initially has a negative influence but eventually a positive influence) would lead to a smaller influence of motivated reasoning but found little evidence for this effect. This research advances the field of causal learning by studying the role of prior preferences, and in doing so, integrates the fields of causal learning and motivated reasoning using a novel explore-exploit task.     

The heterogeneous processes of cheating: Attention evidence from two eye tracking experiments

Dishonesty erodes society. Although much is known about dishonesty, the process leading up to the decision of whether to be honest or dishonest is often assumed to be homogenous and is not well understood. In this paper, we take a more nuanced approach and explore more closely the process of deciding whether to cheat or be honest when an opportunity to cheat arises. We do this in two laboratory eye tracking experiments. In our first experiment (n = 193), we identify heterogeneity in the decision to cheat. Some decisions on whether to cheat or be honest exhibit relatively little variation and appear to be oriented towards cheating (or honesty) with apparently little consideration of the alternative, whereas other decisions seem to be characterized by a higher degree of consideration for both decision alternatives. Our second experiment (n = 299) demonstrates that a gaze dependent intervention in the choices process is able to affect the behavioral outcomes.     

The Role of Independence and Stationarity in Probabilistic Models of Binary Choice

After more then 50 years of probabilistic choice modeling in economics, marketing, political science, psychology, and related disciplines, theoretical and computational advances give scholars access to a sophisticated array of modeling and inference resources. We review some important, but perhaps often overlooked, properties of major classes of probabilistic choice models. For within‐respondent applications, we discuss which models require repeated choices by an individual to be independent and response probabilities to be stationary. We show how some model classes, but not others, are invariant over variable preferences, variable utilities, or variable choice probabilities. These models, but not others, accommodate pooling of responses or averaging of choice proportions within participant when underlying parameters vary across observations. These, but not others, permit pooling/averaging across respondents in the presence of individual differences. We also review the role of independence and stationarity in statistical inference, including for probabilistic choice models that, themselves, do not require those properties. Copyright © 2017 John Wiley & Sons, Ltd.     

Folkbiology meets microbiology: a study of conceptual and behavioral change

Health education can offer a valuable window onto conceptual and behavioral change. In Study 1, we mapped out 3rd-grade Chinese children’s beliefs about causes of colds and flu and ways they can be prevented. We also explored older adults’ beliefs as a possible source of the children’s ideas. In Study 2, we gave 3rd- and 4th-grade Chinese children either a conventional cold/flu education program or an experimental “Think Biology” program that focused on a biological causal mechanism for cold/flu transmission. The “Think Biology” program led children to reason about cold/flu causation and prevention more scientifically than the conventional program, and their reasoning abilities dovetailed with their mastery of the causal mechanism. Study 3, a modified replication of Study 2, found useful behavioral change as well as conceptual change among children who received the “Think Biology” program and documented coherence among knowledge enrichment, conceptual change, and behavioral change.     

Causation and contrast classes

I argue that causation is a contrastive relation: c-rather-than-C* causes e-rather-than-E*, where C* and E* are contrast classes associated respectively with actual events c and e. I explain why this is an improvement on the traditional binary view, and develop a detailed definition. It turns out that causation is only well defined in ‘uniform’ cases, where either all or none of the members of C* are related appropriately to members of E*.

Error probabilities for inference of causal directions

A main message from the causal modelling literature in the last several decades is that under some plausible assumptions, there can be statistically consistent procedures for inferring (features of) the causal structure of a set of random variables from observational data. But whether we can control the error probabilities with a finite sample size depends on the kind of consistency the procedures can achieve. It has been shown that in general, under the standard causal Markov and Faithfulness assumptions, the procedures can only be pointwise but not uniformly consistent without substantial background knowledge. This implies the impossibility of choosing a finite sample size to control the worst case error probabilities. In this paper, I consider the simpler task of inferring causal directions when the skeleton of the causal structure is known, and establish a similarly negative result concerning the possibility of controlling error probabilities. Although the result is negative in form, it has an interesting positive implication for causal discovery methods.     

Addressing confounding errors when using non-experimental, observational data to make causal claims

In their recent book, Is Inequality Bad for Our Health?, Daniels, Kennedy, and Kawachi claim that to “act justly in health policy, we must have knowledge about the causal pathways through which socioeconomic (and other) inequalities work to produce differential health outcomes.” One of the central problems with this approach is its dependency on “knowledge about the causal pathways.” A widely held belief is that the randomized clinical trial (RCT) is, and ought to be the “gold standard” of evaluating the causal efficacy of interventions. However, often the only data available are non-experimental, observational data. For such data, the necessary randomization is missing. Because the randomization is missing, it seems to follow that it is not possible to make epistemically warranted claims about the causal pathways. Although we are not sanguine about the difficulty in using observational data to make warranted causal claims, we are not as pessimistic as those who believe that the only warranted causal claims are claims based on data from (idealized) RCTs. We argue that careful, thoughtful study design, informed by expert knowledge, that incorporates propensity score matching methods in conjunction with instrumental variable analyses, provides the possibility of warranted causal claims using observational data.     

Simmering in the Soviet pot: language heterogeneity in early Soviet socio-linguistics

At the beginning of the ’30s—the period of lively debates on the relation between language and society—one of the main issues in linguistics was language heterogeneity. On the example of the texts by Boris Larin, Georgij Danilov and Lev Jakubinskij we shall compare two attitudes about unity and division of a language. If the studies by Larin and Danilov in various ways establish divisions in society and language at the end of the ’20s, in the ’30s there is a marked tendency to recognize language unity and the cohesiveness of the proletarian society, as seen in socio-linguistic analyses by Jakubinskij. The conclusion, suggested at the end of this exposition, claims that the idea of one national language grows in importance in the discourse of the Soviet linguistics at the beginning of 1930s. Disappearance of the contemporary language heterogeneity in the discourse of Soviet linguists of the period corroborates how linguistics adapts to the political conceptions of society.