© 2015 The Author. Research quality is an increasingly important metric for determining funding allocations, promotion and tenure, and professional prestige. A key metric often used as a proxy for research quality is the ranking of the journal in which a manuscript appears. While citation-based measures of journal quality are commonly used, less is known about other dimensions of journal quality and prestige. We report results from an international study using Best-Worst Scaling to investigate researchers' journal preferences. Respondents used two criteria to assess journals: the impact a paper in the journal would have on career progression, and the impact beyond academia of papers in the journal. Among the sample of journals studied, the American Journal of Agricultural Economics is ranked at the top for career progression for the aggregate sample, while Science was rated at the top for broader impact. We find no significant correlation between the journal scores based on the two criteria, nor between them and the journals' impact factors. These results suggest that impact beyond academia is poorly aligned with career incentives and that citation measures reflect poorly, if at all, peers' esteem of journals. Heteroscedastic scale-adjusted latent class models reveal marked heterogeneity in journal preferences related to researchers' institutional affiliation and geographic region. We find significant differences in error variance over people and choices: people were less consistent when choosing their least, as opposed to their most, preferred journal. This finding has broader implications given the burgeoning use of best-worst surveys.