Also, I am not wedded to a GUI, I'd be happy to just use R but I want to make sure the code doesn't get interfere with learning the methods and concepts, which I think at their level is more important.
Chris On Tue, Dec 3, 2019 at 12:07 PM Christopher David Desjardins < [email protected]> wrote: > Hi Randall, > > Thanks for the email. > > I was indeed quite happy to see your PDF on the Lock website when I began > teaching the course. The last time I looked at your PDF, it it looked like > it was mostly code to do replicate their examples in R, which while > helpful, isn't quite what I'm looking for. I'd love for them to learn R in > that manner, but I think it's slightly beyond their skill set. > > Ideally here's what I'd like. > > I would have the StatKey interface for bootstrapping and randomization > tests written in Shiny (or a similar toolkit with R as the backend) such > that there is a main plot that shows the bootstrap or randomization > distributions (with an option to set the seed) and two side plots to show > the original sample and the current bootstrap/randomization sample. If > possible, I'd like the "current" sample plot to be updated based on the > statistic that is hovered over in the main plot (probably this could be > done with ggplot2 and an additional package or d3.js). As far as I can tell > nothing like this exists for R or Shiny (I'd love to be wrong, though). So, > basically a one-to-one port of StatKey to R. > > It would be great to see this functionality packaged in a Shiny app under > a tab called "Inference via Simulation". Another tab might be called > "Theory-based Inference" (or some better name), that provided a drop down > menu to select a test (binomial, t-test, regression, ANOVA, chi-square), an > option to calculate an arbitrary CI, a plot of the data for assumption > checking, model output printed, and a plot showing the p-value all within > that tab. This could be broken up to make it a little less busy. Finally > there might be another tab for just "Exploring the Data" both numerically > or graphically (essentially a tab for EDA). Additionally, there might be > the option to show/hide the R code. I would like students to be able to > upload the data as a CSV to the app. > > I've got quite a bit of experience making Shiny apps, but don't want to > create one if someone has already made something similar to this that I > could either fork or just use. > > Regarding JASP and jamovi, I found their language to be weird and creating > confidence intervals cumbersome. E.g., creating a confidence interval for a > mean, as far as I could tell, requires running a one sample t-test, making > sure H_0 the Test Value(?) is set to 0 otherwise the CI won't make sense, > clicking location parameter, and then setting the confidence interval and > hoping the students don't select effect size instead, which they do. In > JASP, you can't edit data directly but must edit it in Excel. and you can > edit directly in Rcmdr. JASP's focus on Bayesian is also confusing for my > non-Bayesian course as students select the wrong tests very easily. I do > love that they have the ability to export to LaTeX tables by default, very > nice. Also, in JASP you can't get dot plots, which I think are very helpful > for teaching. Neither software is capable of bootstrapping and > randomization tests as far as I know. > > Chris > > > > > > > On Tue, Dec 3, 2019 at 9:53 AM Randall Pruim <[email protected]> wrote: > >> Chris, >> >> The Lock5withR package includes all the data sets (first edition at this >> point) with more traditional R encoding than the package provided by the >> authors. >> >> Soon (I hope — it’s just pending a final review and polish), there will >> also be a companion volume as a bookdown document using R, mosaic, and >> ggformula to recreate all the examples. An older version is available as a >> PDF here: >> >> http://www.lock5stat.com/other/Lock5withR.pdf >> >> >> For "StatCrunch in shiny", perhaps you can say a bit more about what you >> are looking for and how jamovi and JASP fall short. >> >> I hope you find a solutions that works well for you. >> >> —rjp >> >> >> >> >> On Dec 3, 2019, at 9:25 AM, Manuel Spínola <[email protected]> wrote: >> >> Dear Chris, >> >> How about Radiant? >> >> Manuel >> >> El mar., 3 dic. 2019 a las 7:53, Christopher David Desjardins (< >> [email protected]>) escribió: >> >> Hi, >> >> I teach an introductory statistics course for non-stat/math majors that >> are >> primarily coming from the social sciences or business. I am using the Lock >> textbook, >> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.lock5stat.com_&d=DwIFaQ&c=4rZ6NPIETe-LE5i2KBR4rw&r=S6U-baLhvGcJ7iUQX_KZ6K2om1TTOeUI_-mjRpTrm00&m=rvRMe6UbaU3kDRIQHpLP_3zbje0eW1UBt3Fp6o9KSec&s=qnNW5anVawrMQVNo45I1_4M8Tw3ljsmBYw4qusBVUxs&e= >> , and their software StatKey, >> >> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.lock5stat.com_StatKey_&d=DwIFaQ&c=4rZ6NPIETe-LE5i2KBR4rw&r=S6U-baLhvGcJ7iUQX_KZ6K2om1TTOeUI_-mjRpTrm00&m=rvRMe6UbaU3kDRIQHpLP_3zbje0eW1UBt3Fp6o9KSec&s=N3nV9BtqXRK3zJlRPrEQoqs1wssIPU_MFhAB46Kgf8E&e= >> . I really like the way that StatKey does >> randomization tests and bootstrapping, however, I don't like that it's not >> possible to perform theory-based tests on their website (e.g., if I wanted >> my students to perform an independent samples t-test not by hand or run a >> simple linear regression). >> >> Ideally what I was hoping for was to use R in my introductory classes but >> to have some way to make it less intimidating. I want my students to be >> able to run descriptives and create basic graphics (including dot plots), >> randomization tests, bootstrapping, and run t-tests, chi-square, ANOVA, >> and >> regression. >> >> I tried using RStudio, but it is still overkill for what I want my >> students >> to be able to do. They don't need an IDE. Randall Pruim has kindly made a >> PDF for using R with Lock5, but that is too much for what I want my >> students need. Programming isn't a principal outcome of my course. >> >> I have looked in JASP and jamovi as well, but they don't fit my needs. >> >> The closest thing I have found to what I'm looking for is Rcmdr, but it >> freezes on my Mac periodically regardless of if I use it from the >> Terminal, >> the R GUI, or RStudio. Has anyone else encountered that issue? Rcmdr is >> great because it's GUI driven but also pastes the R code, which is nice >> for >> the more advanced students in the class. >> >> What I am wondering is. >> >> 1. Is anyone using R with Lock5 and how do you use R to do it? >> 2. Does anyone know of a Shiny app that does what StatKey does AND >> includes >> some options for descriptives and inference similar to Rcmdr? I am >> basically looking for a Shiny StatCrunch. >> >> I have thought about creating a Shiny app to do all of this, but if >> someone >> already has a wonderful solution, I don't want to reinvent the wheel. I >> created something very, very basic[1] for my students just to explore the >> Lock5 data and I just might need to expand on it to get what I want. >> >> Thanks for reading my long winded email, >> Chris >> >> [1]: >> https://urldefense.proofpoint.com/v2/url?u=https-3A__cddesja.shinyapps.io_lock5explorer_&d=DwIFaQ&c=4rZ6NPIETe-LE5i2KBR4rw&r=S6U-baLhvGcJ7iUQX_KZ6K2om1TTOeUI_-mjRpTrm00&m=rvRMe6UbaU3kDRIQHpLP_3zbje0eW1UBt3Fp6o9KSec&s=ngdgop1Zp6IG6qxQPENeSz7lk7jGaKc413s_nBuJ7C4&e= >> >> [[alternative HTML version deleted]] >> >> _______________________________________________ >> [email protected] mailing list >> >> https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dsig-2Dteaching&d=DwIFaQ&c=4rZ6NPIETe-LE5i2KBR4rw&r=S6U-baLhvGcJ7iUQX_KZ6K2om1TTOeUI_-mjRpTrm00&m=rvRMe6UbaU3kDRIQHpLP_3zbje0eW1UBt3Fp6o9KSec&s=FBH1JXCQ8weBt8iLEG2ak-fyZrbrbhQIUcUNrVdVYP8&e= >> >> >> >> -- >> *Manuel Spínola, Ph.D.* >> Instituto Internacional en Conservación y Manejo de Vida Silvestre >> Universidad Nacional >> Apartado 1350-3000 >> Heredia >> COSTA RICA >> [email protected] <[email protected]> >> [email protected] >> Teléfono: (506) 8706 - 4662 >> Personal website: Lobito de río < >> https://urldefense.proofpoint.com/v2/url?u=https-3A__sites.google.com_site_lobitoderio_&d=DwIFaQ&c=4rZ6NPIETe-LE5i2KBR4rw&r=S6U-baLhvGcJ7iUQX_KZ6K2om1TTOeUI_-mjRpTrm00&m=rvRMe6UbaU3kDRIQHpLP_3zbje0eW1UBt3Fp6o9KSec&s=VkPHGc5PzRnq4g6kLOx9L_P-9OFvx_cw-CY3jWCceuM&e= >> > >> Institutional website: ICOMVIS < >> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.icomvis.una.ac.cr_&d=DwIFaQ&c=4rZ6NPIETe-LE5i2KBR4rw&r=S6U-baLhvGcJ7iUQX_KZ6K2om1TTOeUI_-mjRpTrm00&m=rvRMe6UbaU3kDRIQHpLP_3zbje0eW1UBt3Fp6o9KSec&s=pc08BWmjY5nHD_C3vT6RKjEKnpX5-BOMoK22JYTX9AA&e= >> > >> >> [[alternative HTML version deleted]] >> >> _______________________________________________ >> [email protected] mailing list >> >> https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dsig-2Dteaching&d=DwIFaQ&c=4rZ6NPIETe-LE5i2KBR4rw&r=S6U-baLhvGcJ7iUQX_KZ6K2om1TTOeUI_-mjRpTrm00&m=rvRMe6UbaU3kDRIQHpLP_3zbje0eW1UBt3Fp6o9KSec&s=FBH1JXCQ8weBt8iLEG2ak-fyZrbrbhQIUcUNrVdVYP8&e= >> >> >> [[alternative HTML version deleted]] _______________________________________________ [email protected] mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-teaching
