Abstract
<p>Data wrangling is an essential skill for reproducible research and graduate employability, but research methods curricula typically emphasise learning a decision tree of statistical tests over preparing raw data for visualisation and analysis. Our exploratory mixed methods study evaluated psychology students’ ability to analyse a processed versus unprocessed dataset and explored their experiences of the analysis process. Forty-two participants used either a processed dataset requiring minimal preparation (n = 25) or an unprocessed dataset requiring multiple wrangling steps (n = 17) to address the same research question. We collected quantitative data on the results that participants could produce and qualitative data on their experiences at each stage of the analysis process. 92% of participants using the processed data and 76% using the unprocessed data successfully produced the correct result, with attrition notably higher for those given the unprocessed data. A content analysis highlighted that students found the problem solving element of data wrangling rewarding but reported challenges in identifying the appropriate wrangling steps and choice of analysis without prescriptive instructions. The key implication of our findings is that psychology research methods curricula should provide scaffolded training in data wrangling and reproducible workflows prior to their independent research project.</p>