Using published data : errors and remedies /

The apparent authenticity of published data can be as dangerous as it is inviting. This guide points out the main dangers (sampling errors, measurement errors, and invalid or unreliable procedures) and analyses the various ways in which these problems arise, giving numerous examples.