< back to specials
Mixed Methods in the Making
Dirk Hovy, Moderator
- Allow us to tackle large data sets
- Can enhance reproducibility/significance
- Show connections/patterns we would miss otherwise
Rely on large data sets, don't work on small
Can give a false sense of significance
Can hide interesting facts in a barrage of numbers
Caroline Sporleder
- Analysis of large amounts of data (text) that would take too long to analyse manually
- Birds-eye perspective: trends and interdependencies over longer time periods or over various domains/sources
- Quantifiable effects and reproducible results
Data can be misleading (e.g. sample bias, digitisation noise)
Data analysis can only show correlations not causality
Data are useless without a good research question and a good strategy for finding answers in the data
Michael Piotrowski
- Quantitative methods can yield quantitative insights and can provide important input into qualitative methods, but there is no single right way
They cannot answer qualitative research questions, and that’s what most research questions in the humanities ultimately are
We still need to know where data comes from and what the results mean: “garbage in, garbage out”
They tend to create an illusion of factuality
They certainly do not obviate the need for theories
Nils Reiter
- Quantitative → formal
- Intersubjective application / detection of categories
- Allow diachronic analysis on large data sets, e.g., shift from telling to showing
- Good scientific practice: quantitative methods account for reproducible research
- Enforces conceptual clarity through operationalization attempts
- Side effects: More understanding of digitalization of society by citizens/researchers/policy makers
Difficult to bridge that gap – if you want to bridge it
Only rarely answer “why” questions
Focus on small, heterogeneous data sets
No interpretable quantitative results
Side aspect: Maintaining “compatibility” with qualitative humanities research beneficial for the career of young researchers