Painting with a broad brush, abductive reasoning1 occurs when we seize upon the simplest plausible explanation for (at least) the important observations. In effect, it works backwards from observation of consequences to explain them by way of the minimum possible number of degrees of freedom2. See also reasoning. Contrast with inductive reasoning and deductive reasoning.
We use abductive reasoning every time we construct multiple trouble-shooting hypotheses3 and check them in order of increasing number of participating variables and interactions between them. See, for example, Atkinson and Donev4.
Use of abductive reasoning can lead directly to serious biases unless the pattern of thought is detected and explicitly verified. See Verification by Wishful Thinking and Verification by Neglect for additional insight. This pattern of reasoning, when coupled with a proper validation effort, can also lead to significant projected cost and schedule impacts when applied to complicated systems. See also Verification by Schedule Variance Analysis and Verification by Cost Variance Analysis.
Footnotes