Page 55 - Special Topic Session (STS) - Volume 3

STS515 Alison L. G. et al.
different teams each week, to promote the acquisition of strategies to
effectively work with others.
• Facilitated problem solving: Throughout our courses, students solved
regular practice problems in a facilitated manner. They were given
worksheets in the form of electronic notebook documents, which they
completed with the help of teaching staff and their peers. These
worksheets were designed as low-stakes formative assessments, geared
towards building skills and understanding concepts. This format is
particularly amenable to what-if type questions; e.g., it allowed us to use
simulation to explore important concepts such as p-hacking and
overfitting. An additional benefit was dynamic two-way feedback, with the
instructor experiencing first-hand where students had difficulties and
providing immediate help.
• Authentic assessment: Summative assessments for one course were
performed on computers, in the same software environment used
throughout the course. We tried to assess students in a way that was as
close as possible to how they would analyse data in real life. Students
were given a set of yet unseen data and asked to perform specific analytic
tasks, as well as answer conceptual, interpretation, and open-ended
questions. At the end of the exam, students submitted individual reports
combining their code, results, and answers in a reproducible manner. This
format afforded us the flexibility to make assessments that are aligned
with what we value.
More details on these strategies, together with material from our
respective courses, can be found at sta130.utstat.utoronto.ca and
utsc.utoronto.ca/~sdamouras/staa57. These courses have been designed to
start students on the trajectory to developing an adaptive statistical mindset.
Continued progression on the trajectory requires integration of novel
problems, learning that emphasizes understanding, and exploration and
discovery throughout our programs of study. For an overview of a complete
program of study designed with such considerations, see Gibbs (2018).

5. Conclusion
Statistics curricula have consistently ensured our graduates are highly
skilled and efficient at solving standard problems in familiar settings. The
requisite procedural knowledge and skills for doing this are sufficient in stable
environments. But our discipline has transformed rapidly with the advent of
Data Science. In light of this transformation, statistics educators have put a
great deal of thought and effort in updating guidelines, programs, and
courses. Along with teaching new knowledge and skills, we need to prepare
students to be able to respond to ongoing change. For this purpose, we have
proposed a focus on an adaptive statistical mindset, one characterised by

44 | I S I W S C 2 0 1 9