An analysis is performed which indicates that beta particle backscattering measurements are highly sensitive to source-scatterer separation distances. It is shown that the primary betas emitted by the source strike the scatterer according to a Cauchy statistical distribution. Then, making the assumption that the primary betas are adsorbed on the scatterer and isotropically reemitted, an effective counting geometry can be obtained. A comparison of this effective geometry with the source geometry will then give an indication of the expected backscatter signal sensitivity. It is shown that a 50-mil separation distance can result in a backscatter measurement error of 25%. Zumwalt's empirical relationship for saturation backscattering is used to analytically predict the expected normalized (source signal equal to one) signal as a function of source-scatterer separation distance and scatterer atomic number. Finally, aluminum, nickel, niobium, palladium, and tantalum scatterers are employed using thallium-204 (204Tl) and phosphorus-32 (32P) beta sources in conjunction with a thin-window halogen-quenched G-M tube to compare experimental and analytical results. This experiment shows that Zumwalt's equation provides an excellent fit to the experimental results in all instances except when employing the low atomic number scatterer, aluminum.