I have some issues with the idea that the h5-index and h5-median as reported by Google Scholar Metrics are superior to the impact factor as a statistic to capture ‘journal quality’. First of all a correction: you talk about what it means if the h5-median is lower than the h5-index, but this is impossible by design, as this is the median of a set of numbers all equal or higher than the h5-index. The big problem with using h5-index (or h5-median) as a quality indicator for a journal is the fact that it does not correct for the number of papers a journal publishes. The impact factor does this. So a journal that publishes many papers (think Plos One) will tend to have a higher h5-index. If two journals would merge, their joint h5-index will typically be higher than their individual h5-indices. Clearly the number of papers a journal publishes is not a good measure of its quality.