The Science Citation Index, Journal Citation Reports (JCR), published by the Institute for Scientific Information (ISI) and designed to rank, evaluate, categorize and compare journals, is used in a wide scientific context as a tool for evaluating researchers and research work, through the use of just one of its indicators, the impact factor. With the aim of obtaining an overall and synthetic perspective of impact factor values, we studied the frequency distributions of this indicator using the box-plot method. Using this method we divided the journals listed in the JCR into five groups (low, lower central, upper central, high and extreme). These groups position the journal in relation to its competitors. Thus, the group designated as extreme contains the journals with high impact factors which are deemed to be prestigious by the scientific community. We used the JCR data from 1996 to determine these groups, firstly for all subject categories combined (all 4779 journals) and then for each of the 183 ISI subject categories. We then substituted the indicator value for each journal by the name of the group in which it was classified. The journal group may differ from one subject category to another. In this article, we present a guide for evaluating journals constructed as described above. It provides a comprehensive and synthetic view of two of the most used sections of the JCR. It makes it possible to make more accurate and complete judgements on and through the journals, and avoids an oversimplified view of the complex reality of the world of journals. It immediately reveals the scientific subject category where the journal is best positioned. Also, whereas it used to be difficult to make intra- and interdisciplinary comparisons, this is now possible without having to consult the different sections of the JCR. We construct this guide each year using indicators published in the JCR by the ISI.