Background: In science education today, there is an emerging focus on what is happening in situ, making use of an array of analytical traditions. Common practice is to use one specific analytical framing within a research project, but there are projects that make use of multiple analytical framings to further the understanding of the same data, either in parallel or in sequence. Purpose: This methodological paper offers a description of using multiple theoretical lenses to address the question "What can be learned in groups discussing physics?" This paper aims to consider and discuss drawbacks and benefits of this design. Sources of evidence: In our earlier research project, different theories were purposefully applied in a series of stratified analyses on video data of university students solving physics problems. Level one used phenomenography and variation theory, level two used positioning theory, and level three used techniques from conversation analysis. Main argument: Each lens contributed new information about group work in physics. Partly due to the openness of our initial question and the character of our video data, every lens brought new relevant information to the picture of group work in physics. While the theoretical lenses did not reference the same data, they operated with data from the same social setting. We point out that although our analytical frameworks are not commensurable, our different results are: together they offer a better understanding for group work in physics. Conclusions: The main benefit was that every level of analysis provided new understandings to create a bigger picture about group work in physics. The order of the analyses was also crucial, since each analysis informed the framing of the next analysis. The biggest drawback was the amount of time and quality of work needed to conduct the analyses.