To take development and budgeting decisions for research activi- ties the officials in charge need to constantly evaluate and assess the quality of research. Over the years a handful of scoring methods for academic journals have been proposed. Discussing the most prominent methods (de facto standards) we show that they cannot distinguish quality from quantity at article level and that they are inherently biased against journals publishing more articles. If we consider the length of a journal by the number of pages or characters, then all methods are biased against lengthier journals. The systematic bias we find is analytically tractable and implies that the methods are manipulable. We show that the strategies for successful manipulation are relatively easy to infer and implement. The implications of our findings extend beyond the evaluation of academic research, to related settings like the ranking of web domains. Non-manipulable methods for measuring intellectual influence exist.