Selecting Age-Dependent Concrete Strength Comparisons for Offline Materials Instruction

Abstract

Materials instruction must distinguish a measured strength ordering from a claim that remains valid at another specimen age. This investigation develops an offline activity that records a prediction before revealing later-age compressive strengths and evaluates the selection of comparisons without repeating formulations. The principal analysis uses 115 formulations measured at both seven and twenty-eight days. Requiring a strength separation of at least 3 MPa at each age leaves 5,141 admissible comparisons, including 185 reversed orderings. Reversal-prioritized selection supplies ten examples involving twenty different formulations. Among a thousand randomly ordered ten-case sessions, 732 contain no reversal, whereas selection by the largest early-strength difference exposes none. Integer optimization establishes that the principal comparison pool can support thirty formulation-disjoint reversals. Selection-order losses become apparent at a stricter seven-to-ninety-day comparison: greedy selection delivers six reversals, although nine are attainable. Alternative age intervals, contrast margins and treatments of repeated measurements distinguish numerical sensitivity from limitations in case supply. Seven graphical analyses characterize observation coverage, shared-formulation dependence, attainable selection, random delivery, contrast magnitude, rank movement and simultaneous composition changes. The instructional contribution is a precise connection between measurement conditions, comparison selection and staged evidence disclosure.