Skip to main content
Fig. 2 | BMC Medical Research Methodology

Fig. 2

From: Aiming for a representative sample: Simulating random versus purposive strategies for hospital selection

Fig. 2

Prediction error at hospital and national level for n(academic) = 4, n(teaching) = 4 and n(general) = 4. Median prediction errors of red blood cell (RBC), plasma (FFP) and platelet (PLT) use, for different sampling strategies. 95 % centiles are provided for the strategies involving a random element. Number of simulations = 1000. LARG = largest hospitals, MAXVAR = maximum variation in number of RBCs, RAND = random, REGVAR = regional variation, 2REG = two regions

Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us