Our quantitative training, much like that of many scholars, was largely procedural — mechanical, repetitive, and lacking in opportunities for deeper reasoning or sense-making. We were taught to ‘turn the crank’: to follow a prescribed set of procedures guaranteed to yield results, without encouraging exploration of the underlying assumptions or implications. This approach left little room for recognizing the substantial array of choices — researcher degrees of freedom — that shape research outcomes, yet often remain unacknowledged.

It was not until we began to examine the impact of using single versus multi-level models [1], along with the treatment of missing data [2], that the impact of these methodological choices came to light. This realization led us to start exploring the choices available to us, grounding these decisions in theoretical frameworks. Specifically, we expanded our inquiry to see how these methodological decisions could be leveraged to focus on equity in research. By scrutinizing model specifications [3] and employing Multi-level Analysis of Individual Heterogeneity and Discriminant Analysis (MAIHDA) [4], we began to see how our choices could directly influence and enhance equity-focused outcomes. This shift toward a more reflective approach to research methodology emphasized the integration of critical theory, setting the stage for a deeper examination of quantitative methods as tools for creating a more just society [5] (Figure 1).

With the ascent of Critical Quantitative (CritQuant) research [6], new perspectives on using quantitative methods to challenge systemic inequities have emerged. CritQuant draws from a diverse array of perspectives [7], including well-known theories like Intersectionality [8] and critical theory [9], as well as lesser-known theories such as Strategic Positivism and Afro-futurism [10]. This diversity enriches its approach, broadening the scope of its applications. Quantitative critical (QuantCrit) research, grounded specifically in Critical Race Theory, has become a particularly influential component of CritQuant. QuantCrit’s tenets emphasize that numbers are not neutral and the importance of speaking for data to confront racial and systemic injustices. While there is some methodological guidance on CritQuant research 11••, 12••, the discourse frequently leans more toward theoretical discussions rather than offering practical application strategies. In our experience, this gap in the literature leaves many researchers in a quandary about how to effectively embody critical theories in their quantitative analyses. As a result, they may either abandon these critical approaches or, worse, unwittingly perpetuate oppressive research practices. In response, our review of articles aims to provide clear guidance on aligning quantitative methods with critical theory on four topics that we have found to be transformational in our research practices (Figure 2). Specifically, we address specification of intersectional models, interpreting model uncertainty, addressing missing data, and framing of inequities. We focus our review on recent articles, but also include some older foundational references.

In this article, we examine some of the central methodological choices that CritQuant researchers must navigate. Recent research shows growing interest in these questions, as evidenced by examinations of latent class analysis [13], model specification [3], interaction terms [14], and student prior preparation [15] as tools for critical investigations. Specific guidance, however, remains underdeveloped in many areas [11]. We begin to address this gap by exploring critical methodological considerations with a focus on regression modeling. Specifically, we address the specification of intersectional models, interpreting model uncertainty, addressing missing data, and framing of inequities. Many of these ideas, however, apply beyond regression modeling.