A commonly used index to estimate the reliability of a building subjected to lateral loads is the drift ratio. Sophisticated computer programs such as STAAD-III have been developed to estimate the drift ratio based on variables such as beam stiffness, column stiffness, story height, moment of inertia, etc. Civil engineers at SUNY, Buffalo, and the University of Central Florida performed an experiment to compare drift ratio estimates using STAAD-III with the estimates produced by a new, simpler microcomputer program called DRIFT (Microcomputers in Civil Engineering, 1993). Data for a 21-story building were used as input to the programs. Two runs were made with STAAD-III: Run 1 considered axial deformation of the building columns, and run 2 neglected this information. The goal of the analysis is to compare the mean drift ratios (where drift is measured as lateral displacement) estimated by the three computer runs.

a. Identify the treatments in the experiment.

b. Because lateral displacement will vary greatly across building levels (floors), a randomized block design will be used to reduce the level-to-level variation in drift. Explain, diagrammatically, the setup of the design if all 21 levels are to be included in the study.

c. Write the linear model for the randomized block design.