Can fractal analysis enable us to classify a lava flow to a morphologic type (e.g., ‘a‘ā vs. pāhoehoe) solely by examining the geometry of the flow’s margin? If so, these classifications would provide insights into the rheology and dynamics of the flow when it was emplaced. Furthermore, the potential to classify lava flows from remotely-sensed data would particularly benefit the analysis of flows for which field access is not feasible. The technique’s current framework depends on three assumptions: (1) measured lava margin fractality is scale-invariant, (2) different morphologic types are consistently distinguishable based on their measured fractality, and (3) any modification of margin fractality by substrate slope or topographic confinement would be minimal or have a recognizable signature. In this study, we critically evaluate each of these assumptions at meter scales using 15 field-collected margin intervals from a wide variety of morphologic types in Hawaiʻi, Iceland, and Idaho. Among the 12 margin intervals that satisfy the current framework’s expectations, 5 exhibit notably scale-dependent fractality and all 5 from transitional lava types would be classified as ‘a‘ā or pāhoehoe at some scales. Additionally, an ‘a‘ā flow on a 15° slope (Mauna Ulu, Hawaiʻi) and a spiny pāhoehoe flow confined by a stream bank (Holuhraun, Iceland) exhibit significantly depressed fractalities but lack distinctive signatures for these modifications. We therefore conclude that all three assumptions are invalid at meter scales. Although fractal analysis of lava margins can provide some constraints on morphologic type, unique classification is not robust at these scales.