Title: Statistical Methods for Evaluating Core Strength Results

Author(s): Ronald L. Dilly and Woodward L. Vogt

Publication: Symposium Paper

Volume: 141

Issue:

Appears on pages(s): 65-102

Keywords: coefficient of variation; compressive strength; cores; quality control; samples; standard deviation; statistical analysis; strength; tests; tolerances (mechanics); General

DOI: 10.14359/3990

Date: 12/1/1993

Abstract:
Strength interpretation problems are created when standard practices and procedures for sampling and testing concrete mixtures were not followed during construction. Cylinder and core compressive strength records are reported for a project that required extensive coring due to low-standard 28-day cylinder strengths. Records are reported for 4000- and 6000-psi concrete mixtures. Over 80 core strengths correspond to the 6000-psi mixture that was used to cast columns, grade beams, pedestals, and shear walls. The described statistical methods were useful for analyzing the quality of core strength data and interpreting the significance of the results. Core strength results were analyzed by mixture, placements, and type of structural member. For the analysis, "Stem and Leaf" and "Box and Whisker" plots were used to identify outliers. "Analysis of Variance" was used to test for equality of mean strengths. "Fisher's" and "Tukey's" procedures were used in identifying significantly different mean strengths. The Chi-square test was applied to evaluate normality of distributions. The characteristic in-place strength was determined by using the tolerance factor. The analysis shows the importance of obtaining representative core strength samples when determining code compliance.