Dams are ubiquitous environmental impacts that threaten aquatic ecosystems. The ability to compare across research studies is essential to conserve the native biodiversity that is impacted by the millions of low‐head dams that currently fragment streams and rivers. Here, we identify a previously unaddressed obstacle that impedes this generalization. Specifically, divergent spatial and taxonomic approaches that result from different conceptualizations of the dam‐biodiversity problem can produce conflicting science‐based conclusions about the same dam impact. In this research, using the same dammed and undammed sites, we evaluated the scientific generality of different conceptualizations of the dam‐biodiversity problem. We compared two different but commonly used spatial approaches—(1) above dam–below dam vs. (2) undammed–dammed comparisons—and 11 different, commonly used taxonomic approaches (three assemblage summaries, eight guilds). Sites above the dam structure had less diverse fish assemblages than sites below dams, whereas sites below the dam structure were similar to undammed sites. Thus, spatial approach 1 detected a large dam effect and spatial approach 2 detected a small dam effect. Similarly, some taxonomic responses (species richness, diversity, abundance, and number of guilds) detected large dam effects; other responses detected small (riffle specialist guild) or no dam effects (pool generalists). In summary, our results showed that how the problem was framed altered scientific conclusions and created different dam realities. The metaphor of how individual blind men disagree about the structure of an elephant, based on examinations of different body parts, reinforces the need for a coordinated, holistic perspective on dam research. Although no single approach is adequate for all problems, identifying the form, consequences of, and relationships among different research conceptualizations will set the stage for future syntheses of dam‐biodiversity research to advance science‐based conservation.