Do Article Influence scores overestimate the citation impact of social science journals in subfields that are related to higher-impact natural science disciplines?