ABSTRACT Applying scientific knowledge to confront societal challenges is a difficult task, an issue known as the science–practice gap. In Ecology and Conservation, scientific evidence has been seldom used directly to support decision‐making, despite calls for an increasing role of ecological science in developing solutions for a sustainable future. To date, multiple causes of the science–practice gap and diverse approaches to link science and practice in Ecology and Conservation have been proposed. To foster a transparent debate and broaden our understanding of the difficulties of using scientific knowledge, we reviewed the perceived causes of the science–practice gap, aiming to: (i) identify the perspectives of ecologists and conservation scientists on this problem, (ii) evaluate the predominance of these perspectives over time and across journals, and (iii) assess them in light of disciplines studying the role of science in decision‐making. We based our review on 1563 sentences describing causes of the science–practice gap extracted from 122 articles and on discussions with eight scientists on how to classify these sentences. The resulting process‐based framework describes three distinct perspectives on the relevant processes, knowledge and actors in the science–practice interface. The most common perspective assumes only scientific knowledge should support practice, perceiving a one‐way knowledge flow from science to practice and recognizing flaws in knowledge generation, communication, and/or use. The second assumes that both scientists and decision‐makers should contribute to support practice, perceiving a two‐way knowledge flow between science and practice through joint knowledge‐production/integration processes, which, for several reasons, are perceived to occur infrequently. The last perspective was very rare, and assumes scientists should put their results into practice, but they rarely do. Some causes (e.g. cultural differences between scientists and decision‐makers) are shared with other disciplines, while others seem specific to Ecology and Conservation (e.g. inadequate research scales). All identified causes require one of three general types of solutions, depending on whether the causal factor can (e.g. inadequate research questions) or cannot (e.g. scientific uncertainty) be changed, or if misconceptions (e.g. undervaluing knowledge) should be solved. The unchanged predominance of the one‐way perspective over time may be associated with the prestige of evidence‐based conservation and suggests that debates in Ecology and Conservation lag behind trends in other disciplines towards bidirectional views ascribing larger roles to decision‐makers. In turn, the two‐way perspective seems primarily restricted to research traditions historically isolated from mainstream conservation biology. All perspectives represented superficial views of decision‐making by not accounting for limits to human rationality, complexity of decision‐making contexts, fuzzy science–practice boundaries, ambiguity brought about by science, and different types of knowledge use. However, joint knowledge‐production processes from the two‐way perspective can potentially allow for democratic decision‐making processes, explicit discussions of values and multiple types of science use. To broaden our understanding of the interface and foster productive science–practice linkages, we argue for dialogue among different research traditions within Ecology and Conservation, joint knowledge‐production processes between scientists and decision‐makers and interdisciplinarity across Ecology, Conservation and Political Science in both research and education.