A split-plot statistical design was used to identify and systematically study the effects of synthesis variables on the pore properties of silica-doped alumina (SDA) from data obtained for 96 samples. Seven preparation variables (mixing method, time, environment; calcination ramp rate; drying temperature, environment; and alcohol concentration) were found to alter surface area, pore volume, and pore diameter over wide ranges, i.e. factors of 1.8, 1.4, and 3.3, respectively. Large pore diameters (> 40 nm) were obtained by addition of excess alcohol. Large pore volumes (> 2 cm 3 /g) were obtained by drying before calcination at 100C for 24 h. High surface area (> 400 m 2 /g) was obtained when no alcohol was used. The gamma phase of all SDA samples was thermally stable to 1200C. Using split-plot statistical analysis of the experimental data, models were developed which predict quantitatively the relationship between surface properties (surface area, pore volume and pore diameters) and synthesis parameters. Based on these models, optimal conditions to produce SDA samples with large (40-60 nm) or medium (16-19 nm) pore diameters, high surface area (>250 m 2 /g) and large pore volume (>1 cm 3 /g) are quantitatively predicted. Model predictions of optimal conditions were accurately