Difference between SPSS and r standard errors in glm Probit analysis?

Hello all,

My name is Edwin R. Burgess, I am a Ph.D. candidate in biology at Northern Illinois University. My dissertation is being done on the effects of commonly used filth fly pesticides on parasitic wasp species used for biological control. Right now, I am doing binary mortality response with a single explanatory variable (dose) on 5 concentrations of one pesticide. Our department is very big on using SPSS, and I have learned to conduct the basic 'Analyze'/'Regression'/'Probit' models with it. Recently, I've started to learn r, and through a combination of the 'glm' and 'dose.p' functions, I have learned to obtain the same slope and intercept, as well as LD50 calculations. However, the standard errors, Z-scores, and p-values calculated through the Probit model in SPSS comes out VERY differently in r. Additionally, the 95% confidence intervals for the LD50 come out very differently between SPSS and r. I am completely puzzled how I am getting the same slopes, intercepts and LD50's, but totally different SE, Z, p-values, and 95% CI. Does anyone have some insight on this?