Method for Forensic Analysis of Residential Floor-Elevation Data

An analysis of the levelness of three residential floor slabs constructed on-grade in the Phoenix, Arizona, metropolitan area is presented. The slabs were built using typical methods by a trade subcontractor for a large production builder in different subdivisions. Measurements of the elevation of the floor slabs were made at 0.3 m (1 ft) spacing across the entire surface of all three slabs within the first week after concrete placement. The resulting elevation data were then used as input in several different analytical techniques, including the maximum elevation difference, maximum slope, $F_L$ local, and $F_L\text{.}$ Special consideration was given to the sample size effects and the repeatability of the results for different samples. All of these methods were found to exhibit dependency on sample size or location, leading to undesirable characteristics. An analysis based on the three-standard-deviation range of the elevation measured from a sample of about 80 points widely spread across the surface of the slab is shown to provide an effective balance for the minimization of both sampling and nonsampling error terms.