to do Poisson regression, click on Generalized Linear Model, and make a custom model with an identity link function. Note: had you used the Counts/Poisson loglinear or log link function, the link function would be logarithm so the mean response rate is exp(a+bX) rather then a+bX. This may be preferable because response rates should not be negative; but it makes the model more difficult to interpret. You could use the custom model with identity link.