Relational model life tables (cont.)
A further explanation of alpha and beta
The graphs below show examples of this and also show how alpha and beta might vary. Remember that the plots show how different the observed data is - relative to the standard. The graph also shows how well the data fits the model you are constructing. If the data is grossly out of line it may be that a more appropriate standard needs to be selected. If deviations from a regression line are moderate the presumption would be that they are out of line because of poor quality data and that points on the line would be improvements. The "on line" points can easily be calculated from alpha and beta. The linear line also provides imputed points for ages not available in the observed data.