Skip to main content

Tomczuk | Reconciling Inductive Reasoning, Deductive Reasoning, and Empirical Findings

February 27, 2017

Sara Tomczuk
Sara Tomczuk explains how the deductive approach to writing her dissertation proposal helped her prepare for the field.

Revising Your Research Question in the Field and Analyzing the Resulting Data

Sara Tomczuk encountered a dichotomy of inductive and deductive reasoning through her dissertation research and shared her experience with the tension between these two approaches at a recent QUAL Speaker Series talk. She argued that the planning and prospectus writing stage of her project was very deductive in nature, but once she was in the field collecting data, she faced a lot of surprises. She embraced the inductive early learning from the interviews and observations in the field and adjusted her research plan. Yet, by the time she was writing her dissertation draft, she found herself once again presenting a lot of her findings deductively.

The audience during Sara Tomczuk's QUAL presentation

Sara Tomczuk started her talk by asking the audience to share their qualitative research experiences with each other.

Tomczuk, a PhD Candidate in Sociology, argues that qualitative researchers ought to be more forthcoming with the dichotomy in their finished, published work. In her field, Tomczuk said, most researchers present their work as deductive (from literature review, to research questions, to data collection as if it was a test on existing hypotheses) even if that is not how the process really happened.

Using her own research on conflict between the majority and the Roma minority populations in Czech Republic and Slovakia, she walked the audience through the three stages: before, during, and after data collection.

Before: Research planning and prospectus writing

When writing her dissertation proposal, Tomczuk said, she never expected the project to be purely deductive. After all, things could go wrong in the field, new events could take place that could challenge the theories she was using.

“The world doesn’t stop, unfortunately, for you to do research,” she said.

But even though she was aware of these potential disruptors, Tomczuk wrote and defended her prospectus as a deductive project, centered on the research question: When does ethnic conflict erupt into violence in Central Europe? This approach was helpful in planning her field work, especially which cases/locations she would study, which key informants she would interview, and what instruments she would use (interview questions scripts).

During: Collecting data in the field

Despite the preparation, once in the field, some unexpected disruptors did emerge. Tomczuk spent a year in the field studying majority/Roma relations. Early on, she found she was no longer confident in her measures of violent and non-violent conflict because hate crimes weren’t recorded or definitions of them were often changed, skewing the reported data. Anti-Roma demonstrations, on the other hand, were easier to track. So, Tomczuk changed her measure of conflict to reflect what she could reliably collect information on – demonstrations.

She also discovered a new category of sources – church organizations – she had not previously planned to interview.

[pullquote]The world doesn’t stop, unfortunately, for you to do research.[/pullquote]Given these changes, some of the sites she had chosen to study still worked, in Czech Republic, but the ones she had identified in Slovakia were no longer a good fit. She expanded her interviews to 10 different towns in Slovakia, from the original two locations planned.

Tomczuk was anxious to make these changes to her research design while in the field. She conferred with her committee members back at the University of Washington, with her key informants, and others to make sure this all made sense. She struggled to stick to the deductive design, while still being responsive to new and unexpected discoveries her sources revealed.

She also experienced a strange phenomenon: what came up as inductive quickly became deductive. After coming across a new or unexpected insight, she soon incorporated it into her protocol for future interviews and even attempted, when possible, to go back to earlier interviewees and ask these additional questions.

After: Analyzing the data

The phenomenon of inductive morphing into deductive continued during the analysis stage of Tomczuk’s dissertation. She used ATLAS.ti, a qualitative data analysis software program, to code her nearly 70 interviews, recorded in three languages. In this process, inductive themes from the field become deductive codes. However, to her surprise, inductive coding still happened and some new ideas emerged only at this stage, revealing a new wave of discovery.

Lessons for qualitative researchers

All our work is at least part inductive and deductive, Tomczuk argued. In her Q&A discussion with faculty and grad students at the talk, two themes come across:

  1. Adjusting your design in the field is not uncommon. It is also not unique to qualitative work; quantitative researchers also often have to adjust their instruments in the field – changing the interview questions and/or survey questions based on what early findings indicate.
  2. The reporting stage should be more open and honest in terms of unexplained variance and admitting when that happens.

To see Sara Tomczuk’s presentation from this QUAL Speaker Series talk, click here.