Figure 1From: Error and bias in under-5 mortality estimates derived from birth histories with small sample sizes Demonstration of weighting scheme used for moving window variants of complete birth history methods. Each panel shows the weights assigned to child-months in each month between 1990–2010 when predicting for a given year (2000 in the top row and 2005 in the bottom row) using a particular window length (five, 10, or 20 years, shown in columns). The color of the line indicates which type of weights are used.Back to article page