The diversity of substantive expertise and analytical techniques spanned by learning analytics can be a challenge for understanding and interrogating methodological choices. But active pursuit of methodological clarity is essential to productive multivocality. This was the focus of the LAK ‘17 Methodology in Learning Analytics workshop from which the idea for this special section arose. Discussions were framed using a building metaphor of developing structurally sound bridges between disciplinary methods. Strong arguments are buttressed by supporting evidence and warrants but also engage with potential alternative explanations. Papers in this special section further the discussion and encompass a range of relevant approaches including sensitivity analysis, model evaluation, model fit, causal inference, and visualization as a methodology. All advance the case of structural integrity in this bridge discipline and highlight the importance of adequate descriptions of methodological considerations to aid our common understanding.