Mixed methods research has long provided strategies for weaving together qualitative and quantitative evidence. But according to a new arXiv preprint, far less guidance exists for a different kind of integration: combining two computational representations of the same qualitative corpus. The paper proposes a procedure called Parallel-Then-Integrated (PTI) to fill that gap.
The PTI procedure is described as a methodological contribution, meaning it offers researchers a structured way to run two computational analyses side by side and then integrate their results. The abstract does not specify which computational methods are involved, but the framing suggests it is intended for text analysis workflows where different algorithmic representations of the same data need to be reconciled.
The paper is an arXiv preprint, so it has not yet undergone peer review. Its main significance is in naming and formalizing a problem that many computational text researchers likely encounter: how to move from separate computational outputs to a coherent, integrated evidence base. The authors position PTI as a response to a specific methodological blind spot rather than a general-purpose mixed methods framework.