The Wikipedia Adventure

From CommunityData
(Redirected from TWA)
Jump to: navigation, search

Project page for The Wikipedia Adventure paper co-authored by Sneha Narayan, Jake Orlowitz, Jonathan Morgan, Mako Hill, and Aaron Shaw.

Current Status[edit]

We are in the process of revising this for submission to CSCW 2017. The timeline and milestones for submission are living on this page (below).


  • We're writing on sharelatex. The project lives here.
  • Bibliography is currently stored in a shared zotero directory. Aaron and Mako can share access if needed.

Current TO DO list[edit]

  • introduction
  • Background and prior work
    • Continue to improve gamification subsection (mako)
  • System design
    • Use Webstrates paper (UIST 2015) as a model (mako)
  • Study 1:
  • Study 2:
  • Discussion & Conclusion
    • make sure connections to framing are clear
  • Throughout
    • Full read-throughs for tone, consistency, clear contributions.
    • Proofreading

Recent changes log[edit]

These have been addressed, but will likely still benefit from further attention so they're here for archival purposes

  • introduction
  • compress the introduction. reduce redundancy (as)
    • focus on two study framing
    • Change framing to show that two studies were always part of the plan
    • Incorporate idea that large scale field testing is important
    • emphasize hypotheses
      • challenges of deploying new systems in a massive community with passionate, experienced members
      • large-scale user survey and invitation-based field experiment under realistic conditions provide more useful evidence (than hallway testing or small-scale usability studies).
  • Study 1:
    • replace bar charts w new visualizations (sn)
  • Study 2:
    • emphasize hypotheses
    • add details of random assignment (sn)
    • add more descriptive information about the sample and experiment
    • incorporate randomization check (compare days in study across treatment & control)
    • restrict main analysis to last 100 days of sampling (sn)
      • revise all reported statistics to reflect this
    • include boxplots (w scatter) of dependent variables for all treatment and control units.
    • Include Mann-Whitney tests for distribution shift
  • Discussion & Conclusion
    • Make sure to address big contributions.
    • Elaborate explanations of null effect.
    • Note that most studies on gamified systems don't do quantitative impact assessment, highlight this as a contribution of our
  • Throughout
  • re-built refs-processed and eliminated missing citation errors (as)
  • change language around Teahouse (jm)
    • Add citations currently in text
    • Remove excess uses of passive voice
    • Add keywords and ACM categories.

Timeline for CSCW 2017 submission[edit]

  • do multiple passes through the whole paper, editing language, grammar and everything else (sn, as, whoever else wants to)
  • share this with folks peer reviewing CSCW paper (sn)
  • read it a hundred times, make sure the title works

Friday, May 27th[edit]

  • submit final version! (though earlier is better)

Post-submission TO-DO list[edit]

  • Reproducibility
    • Create RData files for all datasets, models, results, etc.
    • Convert paper to .rnw (knitr).
    • Create communitydata (wiki?) page to host reproducibility content
    • Make sure git repository is up to date and all files have READMEs
  • Study 2 analysis
    • Report full models w all model fit stats etc. (presumably we cut this down for the CHI note. No need for CSCW).
    • Collect additional covariates and include them in the models
      • user data via API (gender, edit count pre-inclusion in the study, etc.)
      • added measures should enhance precision of estimates (even smaller SEs!)
    • Convert boxplots to a faceted grid in ggplot2.