Selective influence through conditional independence

Abstract

selective influence, conditional independence, marginal selectivity, processing architectures, Thurstonian modeling,

Violation of contextual generalization of the Leggett–Garg inequality for recognition of ambiguous figures

“…Violation of these equalities is interpreted as exhibition of contextuality. In psychology and cognitive science the equalities (1)- (3) represent the special case of so-called marginal selectivity [37]- [39]. It is clear that if at least one of these equalities is violated then one cannot assume existence of context independent joint probability distribution.…”

The statistical properties of the Survivor Interaction Contrast

“…Informally, an experimental manipulation selectively influences a subprocess if that manipulation affects that sub-process and none of other sub-processes being measured. For a discussion of more formal definitions of selective influence, see (Dzhafarov, 2003;Ashby and Townsend, 1980). Under certain constraints on the distributions, direct tests for selective influence exist (Kujala and Dzhafarov, 2008), but in general we can only test that a set of empirical consequences of selective influence are not violated in the data.…”

A Qualified Kolmogorovian Account of Probabilistic Contextuality

“…Its intuitive meaning can be different in different applications. The only constraint imposed on this relation, (complete) marginal selectivity, is as follows (Dzhafarov, 2003). Let index subsets I ⊂ K and J ⊂ L be such that if A ℓ ← α k for some ℓ ∈ J then k ∈ I.…”

“…The existence of such a reduced coupling for a given A is the central theme of the theory of selective influences (Dzhafarov, 2003 (20) and to the ensuing existence of a reduced coupling. This can be done by allowing C k x in (18) to be different from an identity coupling.…”