A Falsificationist Treatment of Auxiliary Hypotheses in Social and Behavioral Sciences

Systematic Replications Framework

Abstract

Auxiliary hypotheses AHs are indispensable in hypothesis-testing, because without them specification of testable predictions and consequently falsification is impossible. However, as AHs enter the test along with the main hypothesis, non-corroborative findings are ambiguous. Due to this ambiguity, AHs may also be employed to deflect falsification by providing “alternative explanations” of findings. This is not fatal to the extent that AHs are independently validated and safely relegated to background knowledge. But this is not always possible, especially in the so-called “softer” sciences where often theories are loosely organized, measurements are noisy, and constructs are vague. The Systematic Replications Framework (SRF) provides a methodological solution by disentangling the implications of the findings for the main hypothesis and the AHs through pre-planned series of systematically interlinked close and conceptual replications. SRF facilitates testing alternative explanations associated with different AHs and thereby increases test severity across a battery of tests. In this way, SRF assesses whether the corroboration of a hypothesis is conditional on particular AHs, and thus allows for a more objective evaluation of its empirical support and whether post hoc modifications to the theory are progressive or degenerative in the Lakatosian sense. Finally, SRF has several advantages over randomization-based systematic replication proposals, which generally assume a problematic neo-operationalist approach that prescribes exploration-oriented strategies in confirmatory contexts.