1 \documentclass[xcolor=dvipsnames]{beamer}
3 % set up the file to create notes in the output PDFs
7 \renewcommand{\rmdefault}{ugm}
8 \usepackage[garamond]{mathdesign}
10 \renewcommand{\sfdefault}{phv}
14 \usepackage[utf8x]{inputenc}
15 \usepackage[T1]{fontenc}
18 % add tikz and a bunch of tikz foo
20 \usetikzlibrary{shapes,shapes.misc,backgrounds,fit,positioning}
21 \tikzstyle{every picture}+=[overlay,remember picture]
23 % add functions to circle parts of slides (e.g., in tables)
24 \newcommand\marktopleft[1]{%
25 \tikz[overlay,remember picture]
26 \node (marker-#1-a) at (0,1.5ex) {};%
28 \newcommand\markbottomright[1]{%
29 \tikz[overlay,remember picture]
30 \node (marker-#1-b) at (0,0) {};%
31 \tikz[overlay,remember picture,dashed,inner sep=3pt]
32 \node[violet!75,ultra thick,draw,rounded rectangle,fit=(marker-#1-a.center) (marker-#1-b.center)] {};%
35 % DEPRECATED function to build a huge centered dropshadow
36 \newcommand\dropshadow[3]{%
37 \node[black!30!white] at (#1+0.1,#2-0.1) {
38 \scalebox{2}{\Huge \textbf{#3}}
41 \scalebox{2}{\Huge \e{#3}}
45 % create an empty quotetxt so we can reuse it
46 \newcommand{\quotetxt}{}
48 % more flexible non-tikz alternative with no dropshadow
49 \newlength{\centertxtlen}
51 \newcommand\centertext[2]{%
52 \setlength{\centertxtlen}{#1}%
53 \setlength{\centertxtlen}{0.48\centertxtlen}%
55 \fontsize{#1}{2\centertxtlen}\selectfont
61 % add function to stop numbering appendix slides
62 \newcommand{\backupbegin}{
63 \newcounter{framenumberappendix}
64 \setcounter{framenumberappendix}{\value{framenumber}}
66 \newcommand{\backupend}{
67 \addtocounter{framenumberappendix}{-\value{framenumber}}
68 \addtocounter{framenumber}{\value{framenumberappendix}}
71 % packages i use in essentially every document
74 % \usepackage{dcolumn}
75 % \usepackage{booktabs}
77 % replace footnotes with symbols instead of numbers
78 \renewcommand*{\thefootnote}{\fnsymbol{footnote}}
80 \MakePerPage{footnote}
83 \usetheme[pageofpages=/,% String used between the current page and the
85 bullet=default,% Use circles instead of squares for bullets.
86 titleline=false,% Show a line below the frame title.
87 alternativetitlepage=true,% Use the fancy title page.
88 titlepagelogo=figures/logo.pdf,% Logo for the first page.
89 %watermark=watermark-polito,% Watermark used in every page.
90 watermarkheight=100px,% Height of the watermark.
91 watermarkheightmult=4,% The watermark image is 4 times bigger
92 % than watermarkheight.
96 \useinnertheme{rectangles}
97 %\setbeamertemplate{blocks}[rounded][]
98 \setbeamercolor{block title}{bg=makopurple3, fg=White}
100 \setbeamertemplate{items}[default]
101 \setbeamertemplate{blocks}[shadow=true]
103 \usepackage{tcolorbox}
104 % These options will be applied to all `tcolorboxes`
107 colback=makopurple5, %background color of the box
108 colframe=makopurple1, %color of frame and title background
109 coltext=black, %color of body text
110 coltitle=white, %color of title text
115 alerted/.style={coltitle=red,
117 example/.style={coltitle=black,
123 %\useoutertheme{infolines}
126 \hypersetup{colorlinks=true, linkcolor=Black, citecolor=Black, filecolor=makopurple1,
127 urlcolor=Plum, unicode=true}
129 % create a boldface version of the header
130 \setbeamerfont{frametitle}{series=\bfseries}
131 \setbeamerfont{title}{series=\bfseries}
133 % tweak the beamer font to make it a bit lists a bit smaller
134 \setbeamerfont*{itemize/enumerate body}{size=\small}
135 \setbeamerfont*{itemize/enumerate subbody}{size=\footnotesize}
136 \setbeamerfont*{itemize/enumerate subsubbody}{size=\footnotesize}
138 % indent the margins of the itemize lists a little bit
139 \setlength{\leftmargin}{0pt}
140 \setlength{\leftmargini}{0.7cm}
141 \setlength{\leftmarginii}{0.7cm}
143 % create a new \e{} command to make things purple and bold
144 \newcommand{\e}[1]{\textcolor{makopurple1}{\textbf{#1}}}
146 % remove the nagivation symbols
147 \setbeamertemplate{navigation symbols}{}
149 \title{Presentation Title}
150 % \subtitle{Presentation Subtitle}
151 \author[Benj. Mako Hill]{\textbf{Benjamin Mako Hill}\\ mako@mit.edu}
153 \institute[MIT/Harvard]{\textbf{Massachusetts Institute of Technology}\\
154 Sloan School of Management\\
157 \textbf{Harvard University}\\
158 Berkman Center for Internet and Society}
160 \date{December 2, 1980}
164 % remove some of the space in the itemize to make it quite compact
165 \let\olditemize\itemize
166 \renewcommand\itemize{\olditemize\itemsep-1pt}
168 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
169 \section{Introduction}
170 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
172 %% SLIDE: Title Slide
173 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
177 \node at (current page.center) [xshift=-3.5cm, yshift=0.5cm, opacity=0.4]
178 {\includegraphics[height=\paperheight]{figures/wikimedia_projects.png}};
181 \node at (current page.south east)
182 [anchor=south east,text width=1.8\paperwidth,align=right,color=black]
185 \fontsize{2.5em}{2.5em}
186 \selectfont {\bf \color{makopurple4} The State of Wikimedia\\
187 Research: 2013-2014} \par}
192 \fontsize{2.0em}{2.1em}
193 \selectfont {\bf \color{black} Benjamin Mako Hill\\
196 Wikimania 2014, London\\
197 August 8, 2014} \par}
204 \tikz[overlay,shift=(current page.south west)]{\node [xshift=5.6em,yshift=0.5em]{\colorbox{makopurple1}{\color{white} \tt \smaller \smaller \smaller revision:\ \VCRevision\ (\VCDateTEX)}};}
206 \note{I've been doing this for many years. I started in 2008 and
207 have done this almost every single year since.
209 This began as an excuse for me to make sure I was up to date on
214 %% SLIDE: Anecdote from Wikimania 2008
215 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
216 \renewcommand{\quotetxt}{``This talk will try to [provide] a quick
217 tour – a literature review in the scholarly parlance – of the last
218 year's academic landscape around Wikimedia and its projects geared
219 at non-academic editors and readers. It will try to categorize,
220 distill, and describe, from a birds eye view, the academic landscape
221 as it is shaping up around
223 \hfill – \e{From my Wikimania 2008 Submission}}
230 \includegraphics[width=\textwidth]{figures/google_scholar_result.png}
233 \tikz{\draw (current page.center) [xshift=-2.1cm, yshift=0.9cm, color=red]
234 ellipse (1.5cm and 0.5cm);}
236 \note<1>{Back in Wikimania 2008, I set out to run a session at
237 Wikimania that would provide a comprehensive literature review of
238 articles in Wikipedia published in the last year.
244 Then, about two weeks before Wikimania, I did the scholar search
245 so I could build the literature.}
247 \note<2->{I tried to import the whole list into Zotero and managed
248 to get banned for abusing the Google Scholar because they thought
249 that no human being could realistically consume the amount of
250 material published on Wikipedia that year.
252 So anyway, I had a 45 minute talk so it worked out to 3.45 seconds
255 And believe it or not, this year is even bigger.
257 And my talk is even shorter.}
261 %% SLIDE: Citations Per Year
262 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
265 \includegraphics[width=\textwidth]{figures/citations_by_year.pdf}
269 {\smaller \emph{Number of citation, per year, with the term
270 “wikipedia” in the title.\\
271 (Source: Google scholar results. Accessed: 2013-08-06)}}
273 \note{Academics have written \e{a lot} of papers about
274 Wikipedia. There are more than 500 papers published about
275 Wikipedia each year and although we've reached and moved past a
276 peak it seems, it's not slowing by much.}
284 \item \e{2968} Wikipedia-related publications in the Scopus database
287 \item \e{160} recent publications reviewed or mentioned in the 12 issues
288 of the Wikimedia Research Newsletter August 2013-July 2014.
293 %% SLIDE: My Scope Conditions
294 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
297 \includegraphics[width=\textwidth]{figures/multiple_issues.png}
300 In selecting papers for this session, the goal is always to choose
301 examples of work that:
305 \item Represent \e{important themes} from Wikipedia in the last year.
306 \item Research that is likely to be of \e{interest} to Wikimedians.
307 \item Research by people who are \e{not at Wikimania}.
310 \note{This is my disclaimer slide...
312 Within these goals, the selections are \e{incomplete}, and \e{wrong}.}
315 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
316 \section{Paper Summaries}
317 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
319 \subsection{Event Prediction}
322 \centertext{6em}{Event Prediction}
326 This was the year that studies of readership of Wikipedia really
327 blossomed. People figured out how to use the view data. Much of
328 what they used it for was prediction.}
333 \frametitle{Wikipedia Viewership and Flu Prediction}
335 \larger \larger McIver, David J., and John
336 S. Brownstein. ``\e{Wikipedia Usage Estimates Prevalence of
337 Influenza-Like Illness in the United States in Near Real-Time}.''
338 PLoS Comput Biol 10, no. 4 (April 17, 2014):
339 e1003581. \href{http://dx.doi.org/10.1371/journal.pcbi.1003581}{doi:10.1371/journal.pcbi.1003581}.
345 \frametitle{Wikipedia Viewership and Flu Prediction: Motivation}
349 \item \e{Google Flu Trends} uses search engine queries to try to
350 predict influenza epidemics more quickly than traditional methods.
351 \item ..but it has been criticized as being biased (e.g., by media coverage).
352 \item WP is freely available and viewership data is free, unlike
353 Google which is proprietary.
357 \note{2009 H1N1 Swine Flu broke GFT.}
362 \frametitle{Wikipedia Viewership and Flu Prediction: Methods}
365 \larger \larger \larger
366 \item Measure traffic to flu related articles on Wikipedia
367 \item Compare to the ``gold standard'' data from the Center for
368 Disease Control (CDC)
374 \frametitle{Wikipedia Viewership and Flu Prediction: Results}
377 \includegraphics[width=\textwidth]{figures/flu.png}
379 \note{\begin{itemize}
382 \item Wikipedia better than Google at predicting peak flu weeks.
383 \item Wikipedia better at predicting relative influenza rates.
390 \frametitle{Other things people have tried to predict include...}
394 \item \href{http://arxiv.org/abs/1405.3612}{Global disease forecasting}
396 \item \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2014/June\#.22Prediction_of_Foreign_Box_Office_Revenues_Based_on_Wikipedia_Page_Activity.22}{Box office revenue based on films}
397 \item \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2013/December\#Attempt_to_use_Wikipedia_pageviews_to_predict_election_results_in_Iran.2C_Germany_and_the_UK}{Election results in Iran, Germany and the UK}
398 \item \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2014/April\#cite_ref-10}{Breaking news stories}
399 \item Trending topics, general zeitgeist.
400 \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2014/June\#cite_ref-13}{[1]}
401 \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2013/August\#Collective_memories_in_Wikipedia}{[2]}
402 \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2013/November\#Twitter_activity_leads_Wikipedia_activity_by_an_hour}{[3]}
403 \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2013/September\#Identifying_trending_topics_of_yesteryear}{[4]}
409 \subsection{Wikimedia as a Corpus}
413 \centertext{5em}{Wikimedia as a Corpus}
417 From the Bar-Ilan lit review: 48\% of them about Wikipedia per se,
418 52\% are just using Wikipedia (e.g. as a text corpus)
420 Wiktionary, non Wikipedia projects.
422 Wiktionary as a source of data, not the substance/object of
423 analysis. Projects as amazing multilingual corpuses of natural
430 \frametitle{Multi-Lingual Dictionary from Wiktionary: Methods}
432 \larger \larger Ács, Judit. ``\e{Pivot-Based Multilingual Dictionary
433 Building Using Wiktionary}.'' In Proceedings of the Ninth
434 International Conference on Language Resources and Evaluation
435 (LREC'14). Rekyjavik, Iceland,
436 2014. \href{http://www.lrec-conf.org/proceedings/lrec2014/pdf/864\_Paper.pdf}{http://www.lrec-conf.org/proceedings/lrec2014/pdf/864\_Paper.pdf}.
438 \note{\e{Methods}: This paper uses a machine classifier to take
439 advantage of ``pivots'' --- words with common translations across
440 more than one language --- to automatically construct a
441 multilingual dictionary via triangulation! They then manually
442 evaluate the precision of this multilingual dictionary on a small
449 \frametitle{Multi-Lingual Dictionary from Wiktionary: Results}
453 \column{0.5\textwidth}
454 \includegraphics[height=0.85\textheight]{figures/Acs-fig5.pdf}
456 \column{0.5\textwidth}
458 \larger \e{Figure:} Translation graph with many pivots. The edge
459 labels denote the source Wiktionary and article of the translation
464 \note{\textbf{Finding:} So, kind of incredibly, this sort of
465 works. The author succeeds in constructing the multilingual
466 dictionary, but finds that problems like polysemy (one word
467 meaning multiple things) limit the precision of the resulting
468 output. Using the multilingual pivots offered by Wiktionary,
469 however, performs much better using ``triangles'' between three
474 \subsection{Content Quality}
477 \centertext{6em}{Content Quality}
481 Mostly focused on sub-areas. There was one this year we considered
482 discussion on hematology. Or information on infectious diseases.}
488 \frametitle{Comparing Wikipedia Quality to Britannica}
490 \larger \larger Nifrário Rodrigues, Fernando
491 Silvério. ``\e{Colaboração Em Massa Ou Amadorismo Em Massa? Um
492 Estudo Comparativo Da Qualidade Da Informação Científica Produzida
493 Utilizando Os Conceitos E Ferramentas Wiki}.'' Universidade de
495 2012. \href{http://massamateurism.blogspot.co.uk/p/synopsis.html}{http://massamateurism.blogspot.co.uk/p/synopsis.html}.
497 \note{A Portuguese-language dissertation at the Universidade de
498 Évora, titled "Colaboração em Massa ou Amadorismo em Massa?"
499 ("Mass collaboration or mass amateurism?")}
505 \frametitle{Comparing Wikipedia Quality to Britannica: Methods}
509 \item Random sample of 245 article pairs from both encyclopedias.
510 \item Graded by an expert in its subject area using a five-point scale.
511 \item Experts asked, ``to concentrate only on some [...] intrinsic
512 aspects of the articles' quality, namely accuracy and objectivity,
513 and discard the contextual, representational and accessibility
515 \item Experts were mostly university teachers.
518 \note{Compared the quality of English Wikipedia with that of
519 Encyclopedia Britannica}
523 \frametitle{Comparing Wikipedia Quality to Britannica: Results}
525 \includegraphics[width=\textwidth]{figures/synopsis4.jpg}
527 \note{They rated "\e{90\%} of the Wikipedia articles ... as having
528 \e{equivalent or better quality} than their Britannica
534 \frametitle{Comparing Wikipedia Quality to Britannica: Results}
536 \includegraphics[width=\textwidth]{figures/synopsis2.jpg}
541 % - https://meta.wikimedia.org/wiki/Research:Newsletter/2014/April#cite_ref-17
542 % - https://meta.wikimedia.org/wiki/Research:Newsletter/2014/February#.22World.E2.80.99s_largest_study_on_Wikipedia:_Better_than_its_reputation.22 (Finnish WP)
543 % - https://meta.wikimedia.org/wiki/Research:Newsletter/2014/March#.22Risk_factors_and_control_of_hospital_acquired_infections:_a_comparison_between_Wikipedia_and_scientific_literature.22
544 % - https://meta.wikimedia.org/wiki/Research:Newsletter/2013/December#.22Evaluation_of_gastroenterology_and_hepatology_articles_on_Wikipedia.22
546 \subsection{Controversy and Conflict}
549 \centertext{7em}{Controversy and Conflict}
556 \frametitle{Conflict, Consensus and Quality in Wikipedia}
558 \larger \larger Osman, Kim. ``\e{The Role of Conflict in Determining
559 Consensus on Quality in Wikipedia Articles}.'' In Proceedings of
560 the 9th International Symposium on Open Collaboration,
561 12:1–12:6. WikiSym ’13. New York, NY, USA: ACM,
562 2013. \href{http://dx.doi.org/10.1145/2491055.2491067}{doi:10.1145/2491055.2491067}.
568 \frametitle{Conflict, Consensus and Quality in Wikipedia: Methods}
571 \larger \larger \larger
572 \item Literature review of research on online communities
573 \item Discourse analysis of [[Talk:Australia]] (+ 17 archive pages),
574 coding 147 threads (156,112 words) in a grounded theory approach.
577 \note{Very simplified: Grounded theory is an approach in social
578 sciences where one starts from empirical data first and develops
579 hypotheses by coding... }
585 \frametitle{Conflict, Consensus and Quality in Wikipedia: Results}
589 ``\e{Conflict} was significantly more prevalent .. than \e{collaboration}.''
594 \item However: Personal attacks are rare (as opposed to
597 \item ``The four main themes that emerged as cause for
598 debate among the editorial community were \e{sources}, \e{wording},
599 \e{structure} and \e{content accuracy}".
601 \item In e.g. sourcing debates, "conflict ... had a role in developing a
602 mechanism to ensure the accuracy of information by prompting
603 participants to properly source and reference material."
605 \item Conflict is not always bad. ``\e{Generative friction}''
606 benefits Wikipedia quality.
609 \note{i.e. editors generally play the ball, not the man.}
615 \frametitle{Conflict, Consensus and Quality in Wikipedia: Results}
617 \includegraphics[width=0.85\textwidth]{figures/osman-fig.png}
621 \item Talk page discussions frequently contain references to
622 Wikipedia policy, both formal (directly linking a policy) and
623 informal (mentioning or quoting it).
625 \item 86\% of policy references are informal
629 \note{Bear in mind that while it was a sizable corpus, it still
630 pertained to only one article on the English Wikipedia.
632 Osman: "More than being a set of isolated rules for the community,
633 policies are part of the fabric of the culture of the talk
634 pages. They regulate both behavior and the production process and
635 manage conflict so that it remains a generative friction", e.g. by
636 discouraging personal attacks.}
640 \subsection{WikiProjects}
644 \centertext{7em}{WikiProjects}
646 \note{Another area of research focuses on understanding wikis and
647 other peer production communities as organization. Some of the
648 most interesting work in this area compares many projects in order
649 to better understand the characteristics that might lead them to
650 grow and attract contributors.}
654 \frametitle{Critical Mass in WikiProjects}
656 \larger \larger Solomon, Jacob, and Rick Wash. “\e{Critical Mass of
657 What? Exploring Community Growth in WikiProjects}.” In Eighth
658 International AAAI Conference on Weblogs and Social Media,
659 2014. \href{http://www.aaai.org/ocs/index.php/ICWSM/ICWSM14/paper/view/8104}{[1]}.
661 \note{\e{Methods:} Analyzed data on 1069 EN:WP WikiProjects. Fit
662 models to describe the projects' growth curves in terms of editors
663 and project edits. They then use these growth curves to estimate
664 the relationship between adding editors or edits (at an early
665 stage) on subsequent project growth.}
670 \frametitle{Critical Mass in WikiProjects: Results}
672 \includegraphics[width=\textwidth]{figures/Solomon_Walsh-2014-critical_mass_wikiprojects-FIG1.pdf}
674 \note{Results: The authors find that projects with more contributors
675 are more likely to experience subsequent growth in contributions
676 and contributors.They also find that contributions from both
677 "power users" and more casual, one-off contributors predicts
678 subsequent growth. These findings (like others in this area) have
679 important implications for project leaders and designers.}
683 \subsection{Vandalism}
686 \centertext{7em}{Vandalism}
692 \frametitle{Edit Patterns and Vandalism Detection}
694 \larger \larger Sethi, Deepika. \e{A Large Scale Study of Edit
695 Patterns in Wikipedia and its Applications to Vandalism
696 Detection}. M. Sc. thesis, University of Georgia. Submitted
703 \frametitle{Edit Patterns and Vandalism Detection: Methods}
705 \larger \larger Used the \e{PAN Wikipedia vandalism corpus 2010 } of
706 32,452 edits, classified as vandalism (2,391 edits) or non-vandalism
707 by Mechanical Turk workers.
709 \note{Corpus was created to train vandalism detectors and
710 formed the basis of several competitions.}
712 Identified vandalism based on:
720 \item \e{Article content domain} using 12 ``classes'' from DBpedia
721 (e.g., 1. Person, 2. Work, 3. Sports, 4. Places, 5. Food ...)
722 \item Content of edits
729 \frametitle{Edit Patterns and Vandalism Detection: Results}
731 \larger \larger \larger "\e{Vandalism occurs the most during office
732 hours} while \e{non-vandalism occurs the most during late
738 \frametitle{Edit Patterns and Vandalism Detection: Results}
740 \larger "Hostilities among the countries are one major cause of
745 \includegraphics[width=\textwidth]{figures/sethi-countries.pdf}
746 \note{\e{Country} where the vandal is based..
748 Methodology: Get Geolocation of IP editors to articles about
749 countries Result: image : Figure 20: India’s Vandal Contributions
750 (caption: Where Indians vandalize most)
752 [[India]] most frequently vandalized from India, Bahrain, Bangladesh, Pakistan, ...
754 [[Pakistan]] most frequently vandalized from India, Pakistan, Poland, Sweden}
757 \note{Other example: [[Taiwan]] from Taiwan, China, Germany ... / [[China]] from Ireland, Germany, Poland, ... }
762 \frametitle{Edit Patterns and Vandalism Detection: Results}
764 % \includegraphics[width=\textwidth]{figures/sethi-46-figure.pdf}
766 \larger \larger \larger \e{Common Vandalism Words}
768 Ball, chicken, British, woman, hole, handicap, meat, kiss, play,
769 old, love, death, course, kick, American, bomb
771 \note{Methodology: Word most frequently occurring in vandalized versions
773 Result: Ball, chicken, British, woman, hole, handicap, meat, kiss,
774 play, old, love, death, course, kick, American, bomb, ...}
777 %\note{Intriguing... A vandal who only added words might aim to get visibility for them; a vandal who only deleted words might dislike a particular statement}
782 \subsection{Editor Motivation}
786 \centertext{7em}{Editor Motivation}
792 \frametitle{When do barnstars increase edits?}
794 \larger \larger Restivo, Michael, and Arnout van de Rijt. “\e{No Praise
795 without Effort: Experimental Evidence on How Rewards Affect
796 Wikipedia’s Contributor Community}.” Information, Communication \&
797 Society 0, no. 0 (0):
798 1–12. \href{http://dx.doi.org/10.1080/1369118X.2014.888459}{doi:10.1080/1369118X.2014.888459}.
804 \frametitle{When do barnstars increase edits? Methods}
806 \larger \larger Prior work by the authors showed that randomly given
807 barnstars to very active editors (top 1\% in a month) resulted in
808 more edits and a positive feedback loop in English Wikipedia.
812 \item Like before, gives barnstars to \e{randomly} selected users --
813 but this time to editors of varying activity level (e.g.,
814 91-95$^{\mathrm{th}}$, 96-99$^{\mathrm{th}}$, and 100$^{th}$
815 percentile of editing in the month). Plus a ``control'' group of
816 other editors who do not receive the award.
817 \item Follow post-award activity on Wikipedia.
824 \frametitle{Do barnstars encourage editing? Results for Edits}
826 \includegraphics[width=\textwidth]{figures/barnstars-1.pdf}
828 \note{Positive effect only in the case of the very active -- top 1\%
829 of editors. No significant difference in the other two groups.}
835 \frametitle{Do barnstars encourage editing? Results for Retention}
837 \includegraphics[width=\textwidth]{figures/barnstars-2.pdf}
839 \note{Lower retention among award recipients in the less active
840 group! No significant difference in the other two groups.}
844 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
846 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
848 %% SLIDE: Other Resources
849 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
850 \begin{frame}{More Resources}
854 \item \e{Wikimedia Research Newsletter} [[:meta:Research:Newsletter]]
855 \item \e{WikiSym} (Later this month in Berlin!)
856 \item \e{WikiPapers Repository} [http://wikipapers.referata.com]
861 \includegraphics[width=0.25\textwidth]{figures/Wikimedia_Research_Newsletter_Logo.png}
865 \note{Those are my six exemplary studies from the past year.
867 There has been just tons and tons of work in this area. Trying to
868 talk about this in 20 minutes strikes me as increasingly crazy
869 every year I try to do it.
871 The most important source, now going for a couple years, is the
872 Wikimedia Research Newsletter which is published monthly in the (English)
873 Signpost and syndicated on the Wikimedia Research.
875 But there are other resources as well. And I encourage you to get
880 \subsection{Meta-Analyses}
884 \frametitle{Meta-Analyses}
891 \href{https://spectrum.library.concordia.ca/978618/}{``The sum of
892 all human knowledge'': a systematic review of scholarly research
893 on the content of Wikipedia}.
895 \item Bar-Ilan and Aharony,
896 \href{http://dl.acm.org/citation.cfm?doid=2615569.2615643}{Twelve
897 years of Wikipedia research}.
899 \item Taraborelli. \href{https://meta.wikimedia.org/wiki/Research:Newsletter/2013/August\#Keynote\_on\_applicable\_Wikipedia\_research}{Keynote
900 on Wikipedia Research}. OpenSym 2013. Hong Kong.
902 \item Benkler, Shaw, and Hill,
903 \href{http://mako.cc/academic/benkler\_shaw\_hill-peer\_production\_ci.pdf}{Peer
904 Production: A Modality of Collective Intelligence}.
913 % LocalWords: xshift yshift makopurple Tilman Wikimedians Okoli al
914 % LocalWords: Ilan Aharony Taraborelli OpenSym Hong shaw ci pdf GFT
915 % LocalWords: McIver Brownstein Comput Wiktionary Acs Judit LREC Ou
916 % LocalWords: Rekyjavik Multi polysemy Nifrário Rodrigues Silvério
917 % LocalWords: Colaboração Massa Amadorismo Estudo Comparativo Da ou
918 % LocalWords: Qualidade Informação Científica Produzida Utilizando
919 % LocalWords: Conceitos Ferramentas Universidade Évora WikiProjects
920 % LocalWords: Weblogs Sethi Deepika DBpedia Restivo Arnout Rijt th
921 % LocalWords: WikiPapers