But for different subject areas besides [mine], you know, they may not
be as readily available. I’m familiar with the ones in [my discipline]. If
they had a special interest in a certain area, it would be up to them to
find it.
Generally, when instructors prefer that students find data themselves, unfamiliarity with
the research process and how difficult it can be to find relevant datasets presents a
challenge to students. This, in turn, presents a challenge for instructors because they
must monitor, support, and encourage students to start the process long before any
deliverables come due. This is evidenced by comments from three instructors.
Students sometimes find HUGE, dense datasets that they then have to
carve out and try to figure out which parts are useful.
Other students have spent, you know, an entire semester just banging
their head against a dataset, a table, and just breaking that table into
something that can then be useful.
There is a really short window where they’ve got a time, “Ok, I need to
go out and get a dataset and figure out what kind of dataset…I need to
do,” and that has to be done early enough in the semester... So, in that
case, I often will be like, “If you want to do this kind of project where you
go out and generate data, I need to know about three weeks into the
semester to help you figure it out.
Ethical Challenges
Ethical challenges in teaching with data in the social sciences take many forms. Our
respondents touched on topics of bias, human subject protection, reproducibility and
transparency, technology outgrowing ethical thinking, and potentially trauma inducing
research. These issues are compounded by big data-mining efforts, in which data is
collected and analyzed and stripped of its context. These instructors expressed concern
that students often don’t know to look for many of these issues.
Maintaining objectivity and understanding data bias were identified as foundational
ethical issues that some of our instructors try to address in data-intensive classes.
Considerations include understanding who recorded the data and for what purpose,
erasure of the archives, and the categorization of individuals reflected in the data. Two
instructors elaborated on these ethical considerations.
The fact that a human being has been transformed in the Census
record into an age, sex, skin color, and nothing else is, in itself, an
ethical issue… and so even the data that we are dealing with, I think, is
just dripping with ethical considerations.
…but you know, I AM teaching how to use data and manipulate data to
answer very important social science questions, and I do forefront