TY - JOUR
T1 - Two-sample test for ambivalent subset relationship in fuzzy set qualitative comparative analysis
AU - Veri, Francesco
N1 - Funding Information:
I would like to express my gratitude to the two anonymous reviewers for their diligent and thorough work. It is evident from their comments that they are not only experts in QCA, but also deeply passionate about their roles as reviewers. Additionally, I would like to thank Prof. Alessia Damonte for her feedback on an earlier version of this manuscript, which was presented at the 2022 Italian Political Science Association Annual Conference in Rome.
Publisher Copyright:
© The Author(s) 2023.
PY - 2024/4
Y1 - 2024/4
N2 - In fuzzy set qualitative comparative analysis (fsQCA), ambivalent subset relationships (ASR), occur when solution term X is in subset relation with the outcome Y and its absence ~ Y, leading to false-positive results. While ASR can be empirically detected in small-N and medium-N cases through in-depth case knowledge, it is challenging to identify them in large-N case designs. QCA parameters such as proportion reduction inconsistency (PRI) and consistency are commonly used to identify simultaneous subset relationships (SSR), but they are not specifically designed to detect ASR. To address this issue, this article introduces the DTS test, a new test based on two-sample statistics. The DTS test identifies distributional convergence between a solution term’s empirical cumulative distribution function (eCDF) and an eCDF of solution formulas with asymptotic ASR characteristics. By comparing empirical solutions’ patterns with spurious artificially built solutions' patterns, the DTS test reduces the risk of causal fallacies in interpreting the empirical results. Overall, the DTS test provides a valuable tool for identifying and addressing potential ASR bias in fsQCA, particularly in large-N case designs.
AB - In fuzzy set qualitative comparative analysis (fsQCA), ambivalent subset relationships (ASR), occur when solution term X is in subset relation with the outcome Y and its absence ~ Y, leading to false-positive results. While ASR can be empirically detected in small-N and medium-N cases through in-depth case knowledge, it is challenging to identify them in large-N case designs. QCA parameters such as proportion reduction inconsistency (PRI) and consistency are commonly used to identify simultaneous subset relationships (SSR), but they are not specifically designed to detect ASR. To address this issue, this article introduces the DTS test, a new test based on two-sample statistics. The DTS test identifies distributional convergence between a solution term’s empirical cumulative distribution function (eCDF) and an eCDF of solution formulas with asymptotic ASR characteristics. By comparing empirical solutions’ patterns with spurious artificially built solutions' patterns, the DTS test reduces the risk of causal fallacies in interpreting the empirical results. Overall, the DTS test provides a valuable tool for identifying and addressing potential ASR bias in fsQCA, particularly in large-N case designs.
KW - Ambivalent subsets relationship
KW - False positive
KW - Parameters of fit
KW - Qca
KW - Set theory
KW - Spuriousness
KW - Two sample test
UR - http://www.scopus.com/inward/record.url?scp=85160648357&partnerID=8YFLogxK
U2 - 10.1007/s11135-023-01687-8
DO - 10.1007/s11135-023-01687-8
M3 - Article
AN - SCOPUS:85160648357
SN - 0033-5177
VL - 58
SP - 1235
EP - 1253
JO - Quality and Quantity
JF - Quality and Quantity
IS - 2
ER -