[en] Current practices for the measurement of skin dose are reviewed and found to be inadequate. The current INEL dosimeter was examined for systematic and random error. Systematic (i.e., variation with energy) error was found to range over a factor of 10 while the random error was reasonably small for the larger β/γ ratios. Other designs with thicker windows as is more common would show even larger systematic errors. Various methods for improving beta dosimetry were reviewed. A new dosimeter design utilizing three chips, each having a window of difference thickness, was proposed. According to the calculations, this dosimeter should markedly reduce systematic error but will introduce somewhat more random error. Preliminary measurements were carried out related to angular dependence and charged particle equilibrium. The proposed dosimeter design was tested with betas from a uranium slab. The average of the seven results was in excellent agreement with the known dose rate and the standard deviation of the result was 16%