Results based on a sample of Canadian households challenge the findings of most studies which show significant negative effects of schooling on the fertility of women under the age of 45. This is due to the application of methods to an optimization model which distinguish between those households which have completed their reproductive behaviour from those which have not. Completion status and the desired number of children are used to infer characteristics of the optimal programme which are then employed to derive a likelihood function. Traditional demographic methods have so far not fully utilized the distinction between incomplete and completed households in sample surveys. These methods also lead to the conclusion that completed fertility had increased from its all time low in the nineteen seventies.