On Fitting a Model to a Population Time Series With Missing Values