Assessment is necessary, but difficult, in statistics education. Multiple choice assessments are common, particularly for research purposes. Open-ended assessments may be more adept at revealing student understanding, but ensuring their validity can be difficult. The study presented here examines differences in student descriptions of histograms for different question prompts and scenarios in order to understand how best to ask such questions in research and teaching situations. The results show that different ways of asking students to describe histograms and different scenarios and shapes of histograms lead to systematic differences in student descriptions of histograms. The paper concludes with suggestions for phrasing questions, both for research and classroom assessment, and provides directions for future research based on our results.