CS235 Chapter 6: Warrants for Research Arguments
Document Summary
Warrants the standards used to evaluate whether particular evidence represents an effective way to support a claim. Warrants address issues of agreement and values of precision and accuracy. Also value precision because they use numerical data. 3 types of evidence needed to demonstrate causality: time order, covariation, control over rival hypothesis. 2 conditions must be met in order to establish causal explanations: necessity necessary conditions must be present for effect to occur, sufficiency cause is best possible explanation for effect under study. Evaluate merit of causal arguments by 2 standards of evidence: reliability, validity (discovery paradigm warrants cont. ) Random errors in measurement sometimes called noise. 3 sources of random error contribute to inconsistent measurement: random individual differences. Every participant, text, or situation is unique: lack of instrument clarity. Questions poorly worded or instructions too ambiguous: errors in data processing. When data are translated from on form to another.