To delineate further the cause of acetabular loosening in a group of 238 total hip arthroplasties performed between 1980 and 1982, the authors subjected those total hip arthroplasties to a multifactorial analysis utilizing the log normal model. Data collected on each patient consisted of age, sex, weight at the time of surgery, diagnosis, cement thickness between zones 1 and 2 and between zones 2 and 3, polyethylene thickness, lateral tilt, and metal backing of the acetabular components. The acetabulum was considered a failure by radiographic loosening with migration greater than 5 mm, revision in patients younger than 60 years of age (P = 0.154), or when the acetabular cup was metal-backed (P = .001).