[R-sig-teaching] R for introductory epidemiology for MPH students--base or with a package or two?

[Timothy Bates]

I'd say stick to base R wherever possible.

We added tidyverse in post graduate lectures this year.In practice, it was
just a user-preference generating a lot more non-statistics to learn,
distracting from the actual statistics that R does so well, and that is
what stats courses are for, after all :-)

Based on that, we're dropping all the tidyverse this year. Just saying to
students that there are various tool chains they might adopt as they
progress beyond the course. Tidyverse also made marking harder (tutors who
don't know it), and makes debugging harder, as it encourages copy-paste of
long chains of piped functions, that can't easily be probed for
intermediate results. May as well just wrap the whole lot into a function
if that's the case.

The exceptions, I think, are packages where there isn't support in base R.
But that's really restricted to things like structural equation modelling.

Some visualization and table formatting tools can also be helpful: like
visreg.

Others will no doubt agree and disagree :-)
T

On Tue, Jul 3, 2018, 6:16 AM K Imran M <drki.musa using gmail.com> wrote:

> Chris,
>
> I introduced R to our DrPH students like 3 or 4 years ago. All of our
> DrPH students come from MD background with very little or no
> programming background. I started teaching them with base R and
> gradually increase the use 'tidyverse' packages. I can't help but
> noticing that 'tidyverse' fastens the learning process especially for
> data wrangling (using dplyr) and exploratory data analysis (using
> ggplot2). The help documents (vignettes) for tidyverse are very
> useful, the books are available (R graphics Cookbook, R for Data
> Science) and the websites (ggplot2 website, Cookbook for R websites)
> are excellent. I also teach them Rmarkdown which they are using very
> happily to create pdf and MS Word (especially) for their assignments.
> So as for now, tidyverse constitutes for almost 95% of my teaching and
> my workshop when it comes to data wrangling and data exploration.
>
> Regards,
> Kamarul
> -------------------------------------
> Kamarul Imran Musa
> MD, MCommunityMed, PhD (Statistics & Epidemiology)
> Associate Professor (Epidemiology & Biostatistics) and
> Public Health Medicine Specialist,
>
> Department of Community Medicine,
> School of Medical Sciences,
> Universiti Sains Malaysia,
> 16150 Kbg Kerian Kelantan
> Malaysia
> Tel +6097676621
>
>
> Malaysian Medical Council Registration: 34450
> ResearcherID: http://www.researcherid.com/rid/N-3198-2015
> Google-scholar: 'Kamarul Imran Musa'  at https://goo.gl/D3o3y6
> ORCID ID: orcid.org/0000-0002-3708-0628
> ScopusID: 57194536466
> On Tue, Jul 3, 2018 at 4:30 AM Christopher W. Ryan <cryan using binghamton.edu>
> wrote:
> >
> > I'll be teaching intro epidemiology in a new MPH program, starting this
> > fall. Weekly sessions, each 3 hours long. Expecting 12-20 students. I
> > plan to try to make it fairly interactive, with a "computer lab" as part
> > of almost every class session. Using R. I'll do an initial "needs
> > assessment" prior to or on first day of class; for now I assume none of
> > the students are at all familiar with R. My first thought was to limit
> > my efforts to base R, rather than try to use any installable packages.
> > Any opinions about that? I specifically wonder whether Hadley Wickham's
> > tidyverse way of doing things is become so commonplace (and rightly so!)
> > that I should introduce this. It certainly makes data wrangling much
> > easier, and that is a lot of what epidemiologists do, since we are so
> > often given existing data that were not recorded with future analyses in
> > mind.
> >
> > Thoughts on any of the above?  Thanks.
> >
> > --Chris Ryan, MD, MS
> > Binghamton University,
> > SUNY Upstate Medical University,
> > and
> > Broome County Health Department, NY, US
> >
> > _______________________________________________
> > R-sig-teaching using r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-sig-teaching
>
> _______________________________________________
> R-sig-teaching using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-teaching
>

	[[alternative HTML version deleted]]