Limits...
How multiple causes combine: independence constraints on causal inference.

Liljeholm M - Front Psychol (2015)

Bottom Line: Participants were presented with pairs of candidate causes that were either confounded or not, and that either interacted or exerted their influences independently.Consistent with the causal power view, uncertainty about the existence and strength of causal relationships was greater when causes were confounded or interacted than when unconfounded and acting independently.An elemental Bayesian causal model captured differences in uncertainty due to confounding but not those due to an interaction.

View Article: PubMed Central - PubMed

Affiliation: Department of Cognitive Sciences, University of California Irvine, CA, USA.

ABSTRACT
According to the causal power view, two core constraints-that causes occur independently (i.e., no confounding) and influence their effects independently-serve as boundary conditions for causal induction. This study investigated how violations of these constraints modulate uncertainty about the existence and strength of a causal relationship. Participants were presented with pairs of candidate causes that were either confounded or not, and that either interacted or exerted their influences independently. Consistent with the causal power view, uncertainty about the existence and strength of causal relationships was greater when causes were confounded or interacted than when unconfounded and acting independently. An elemental Bayesian causal model captured differences in uncertainty due to confounding but not those due to an interaction. Implications of distinct sources of uncertainty for the selection of contingency information and causal generalization are discussed.

No MeSH data available.


Related in: MedlinePlus

Model predictions and behavioral results for the confounded target cause. Predictions of the Bayesian causal model are shown in the top row, and mean human judgments in the bottom row, for structure judgments (left), strength estimates (middle) and uncertainty in strength estimates (right). Structure judgments were derived by coding structure choices as “No influence” = −1, “Can't tell” = 0, “Generative influence” = 1. Labels “Weak” and “Strong” indicate the model-derived strength of the target cause (C) at the end of the 2nd phase. Error bars = SEM.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4530320&req=5

Figure 3: Model predictions and behavioral results for the confounded target cause. Predictions of the Bayesian causal model are shown in the top row, and mean human judgments in the bottom row, for structure judgments (left), strength estimates (middle) and uncertainty in strength estimates (right). Structure judgments were derived by coding structure choices as “No influence” = −1, “Can't tell” = 0, “Generative influence” = 1. Labels “Weak” and “Strong” indicate the model-derived strength of the target cause (C) at the end of the 2nd phase. Error bars = SEM.

Mentions: Human data for structure, strength and uncertainty about strength are shown in the bottom rows of Figure 3 (for the confounded target cause) and Figure 4 (for the interacting target cause), with corresponding plots of model prediction shown in the top rows. A 2 (Phase) × 2 (Group) mixed analysis of variance (ANOVA) was performed on each type of rating (strength and uncertainty about strength) and for each type of causal power violation (confounding and interaction). The results of these analyses are reported in relevant subsections. Cohen's dz (hereafter dz) is reported for all pairwise comparisons. Throughout the results, group-labels “Strong” and “Weak” refer to the model-derived strength of the relevant target cause at the end of Phase 2.


How multiple causes combine: independence constraints on causal inference.

Liljeholm M - Front Psychol (2015)

Model predictions and behavioral results for the confounded target cause. Predictions of the Bayesian causal model are shown in the top row, and mean human judgments in the bottom row, for structure judgments (left), strength estimates (middle) and uncertainty in strength estimates (right). Structure judgments were derived by coding structure choices as “No influence” = −1, “Can't tell” = 0, “Generative influence” = 1. Labels “Weak” and “Strong” indicate the model-derived strength of the target cause (C) at the end of the 2nd phase. Error bars = SEM.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4530320&req=5

Figure 3: Model predictions and behavioral results for the confounded target cause. Predictions of the Bayesian causal model are shown in the top row, and mean human judgments in the bottom row, for structure judgments (left), strength estimates (middle) and uncertainty in strength estimates (right). Structure judgments were derived by coding structure choices as “No influence” = −1, “Can't tell” = 0, “Generative influence” = 1. Labels “Weak” and “Strong” indicate the model-derived strength of the target cause (C) at the end of the 2nd phase. Error bars = SEM.
Mentions: Human data for structure, strength and uncertainty about strength are shown in the bottom rows of Figure 3 (for the confounded target cause) and Figure 4 (for the interacting target cause), with corresponding plots of model prediction shown in the top rows. A 2 (Phase) × 2 (Group) mixed analysis of variance (ANOVA) was performed on each type of rating (strength and uncertainty about strength) and for each type of causal power violation (confounding and interaction). The results of these analyses are reported in relevant subsections. Cohen's dz (hereafter dz) is reported for all pairwise comparisons. Throughout the results, group-labels “Strong” and “Weak” refer to the model-derived strength of the relevant target cause at the end of Phase 2.

Bottom Line: Participants were presented with pairs of candidate causes that were either confounded or not, and that either interacted or exerted their influences independently.Consistent with the causal power view, uncertainty about the existence and strength of causal relationships was greater when causes were confounded or interacted than when unconfounded and acting independently.An elemental Bayesian causal model captured differences in uncertainty due to confounding but not those due to an interaction.

View Article: PubMed Central - PubMed

Affiliation: Department of Cognitive Sciences, University of California Irvine, CA, USA.

ABSTRACT
According to the causal power view, two core constraints-that causes occur independently (i.e., no confounding) and influence their effects independently-serve as boundary conditions for causal induction. This study investigated how violations of these constraints modulate uncertainty about the existence and strength of a causal relationship. Participants were presented with pairs of candidate causes that were either confounded or not, and that either interacted or exerted their influences independently. Consistent with the causal power view, uncertainty about the existence and strength of causal relationships was greater when causes were confounded or interacted than when unconfounded and acting independently. An elemental Bayesian causal model captured differences in uncertainty due to confounding but not those due to an interaction. Implications of distinct sources of uncertainty for the selection of contingency information and causal generalization are discussed.

No MeSH data available.


Related in: MedlinePlus