Best-fit values to the yearly word frequencies from the Google 1-gram database.

The listed values yield the 95% confidence interval, i.e., from to for the nonlinear least-squares estimates of each parameter in the previous column (assuming normally distributed errors).

The start date indicates the first year of the time series, which was estimated to be the start of the Bass curve. Errors on the parameters were calculated except where “n.d.” indicates that, in the fitting process, the model is insensitive to this parameter.