Once more with e-mail…

… sent off about 4 e-mail inquiries to police departments with a set of questions about police procedures and advice for an article I would like to write.  The initial responses were all out of office replies.  Not optimistic.

Advertisements

Editing patterns on English Wikipedia for cohorts with mental disorders

Contributing to English Wikipedia, unlike other social networks, does not allow self-selection of people one desires to interact, and does not allow users to easily move into different spaces within the same website as a way of removing oneself from stressers related to participation on the site.  People with different learning disabilities and mental health conditions are thus forced to interact and deal with people they might otherwise not wish or be able to functionally interact with if they were participating elsewhere.

image001This research and future research is important for potentially developing targeted approaches to better integration of people with disorders into the English Wikipedia community.   One of the issues in addressing this is which epistemology should be used, a clinical one grounded in psychology and solutions resulting from best practices for treatment, or an educational perspective grounded in learning theory and integration of these learners into an educational environment.  The perspective one takes necessitates the use of one wording selection over another.  For the most part, the emphasis has been placed on using a psychological approach, which defines the cohorts looked at as having mental disorders and relying on the Diagnostic and Statistical Manual of Mental Disorders (DSM) developed by the American Psychiatric Association for descriptions of these disorders.  As there is a desire on the part of the researcher for exploring in the future ways to better integrate these cohorts into English Wikipedia, educational terminology is at times used which defines these cohorts as having specific learning disabilities.  Some conditions, such as dyscalculia and dyslexia, are classified as learning disabilities or learning disorders but have not always been included in DSM listings for mental disorders.  This issue results in an uneasy tension in the research.image002

Wikipedia has userboxes, which allow users to tell other readers and contributors a little bit about themselves quickly.  English Wikipedia has a number of such userboxes where contributors can indicate their different types of learning disabilities and mental health issues.  The inclusion of such boxes is completely voluntary.  User boxes also allow for easily identifying unique user cohorts around certain conditions without relying on private, non-disclosed information.  The downside to this data is it is completely self-selecting and no research has been done to examine how representative these users are of the whole population.

https://en.wikipedia.org/wiki/Wikipedia:Userboxes/Health contains a list of user boxes which can be used to get lists of users based on different self-identified users.  From these, several disorders were selected including autism, Asperger Syndrome, ADD/ADHD, dyscalculia, dyslexia, OCD and social anxiety. There are 20 users who identify as having autism, 443 who identify as having Asperger, 405 who identify as having ADD/ADHD, 8 users who have dyscalculia, 139 who identify as having dyslexia, 238 who have obsessive compulsive disorder and 208 with social anxiety.

The Wikimedia Foundation provides some tools that allow easy comparison between cohorts called Wikimetrics, which can be located at https://metrics.wmflabs.org/ .  The user list from each learning disability or mental disorders was input as its own cohort, and data was then output based on the available options.  These metrics are ones the Wikimedia Foundation metrics staff have identified as either important or easy to attain data for the purpose of analysis.  The availability of the metrics is the constraints of this research.

The following analysis is broken down into two parts.  The first part compares the learning disability and mental disorder cohorts to each other based on metric type.  The second part looks at the specific learning disability and mental disorders individually to understand how the groups function in their own editing context.

While autism data is included in this data, the sample size is very small and likely does not reflect the broad array of people who have autism which includes four groups according to DSM-V.  At the same time, the symptoms of autism generally suggest that people with autism would have extreme difficulty effectively contributing to English Wikipedia.  This includes such problems as difficulty reading, difficulty with language acquisition and retention, and problems with motor coordination.  It seems likely, though it is not supported by observations, that some of the users who put the autism user boxes on their user pages are more likely to be people who support autism related topics given the visibility of and activism surrounding the disorder.  This sort of affiliation may be present for other mental disorders and learning disabilities but the size of the cohorts is likely to mitigate that effect somewhat and it seems likely that there would be similar activist to people having the disability across all disability types.

Research limitations

This research has some fundamental problems that cannot be resolved given the limited analysis, scope and constraints of data availability.  It is intended only as a potential starting point for further research, and to think about how formulate a research methodology that allows for a better understanding of this topic and its impact in recruiting and retaining users in these cohorts.

There is little available research on the accuracy of using user groups to represent the populations the userboxes indicate the population the user is part of.  This issue really needs to be addressed to begin to understand if the users are represented of the disorders.

For several of the disorders represented, the symptoms and the ability to function may manifest very differently from one individual to another.  This is especially true for disorders where the DSM has groupings inside the disorder or indicates the inability to function falls along a spectrum.  Disorders with this issue include ADHD, and autism.  In the case of autism, there are four diagnostic groups including autistic disorder, Asperger’s disorder, childhood disintegrative disorder, or pervasive development disorder not otherwise specified.

Another potential issue impacting results based on reliance on userbox self-identification is the diagnostic criteria used to arrive at the determination that a person has the stated disorder.  People may incorrectly self-diagnosis.  Some diagnoses may originate in an educational setting.  Other people with a disorder may be diagnosed in a clinical setting.

In the case of social anxiety, ADHD and OCD, people do not always understand these explicitly as medical disorders and use the terms informally, in non-medical ways as verbal and written short hand for personality characteristics. This is at times manifested explicitly on English Wikipedia in various places including a template https://en.wikipedia.org/wiki/Template:Wikipediholism found on pages about Wikipedia such at the list of most prolific article creators at https://en.wikipedia.org/wiki/Wikipedia:List_of_Wikipedians_by_article_count. One could speculate that by owning being hyperactive, somewhat impulsive and obsessively focused on detail, the user may believe there is a benefit from the community in terms of positive discrimination. This issue is similar to problems with self-diagnosis in terms of potential reliability and all members of a cohort having diagnostic criteria for the stated disorder.

These issues all play into the reliability of userboxes as a way of cohort grouping around a disorder.  As previously stated, little research has been done into the reliability of userboxes for cohort grouping, which makes such analysis even more potentially tricky given the lack of common language, disability groupings and spectrums inside a disorder, and lack of understanding of diagnostic criteria leading to a person including the relevant userbox on their user page.

An ethical issue also exists.  All the data used for this analysis was publicly available, and were it conducted inside a university setting, may not require a human ethics research approval.  Public data is public data.  Given Wikipedia´s transparency in editing and the fully trackable nature of all edits on a project, there is even less of an expectation of privacy especially given the completely voluntary disclosures on the user pages.  Still, some people may have concerns about doing research about people with mental health disorders and learning disabilities without requesting they sign a consent form.

The use of Wikimetrics also brought limitations to the data.  The output is often one line per cohort.  The complete user list and their individual metrics that result in the final score are not available in the output.  The results are thus sum totals.  This becomes a problem in some cases such as content creation or total bytes added per year because the total cannot be factored based on the total number of people based on registration date.  For example, people with ADHD may have registered en masse in 2010, but the results for 2002 to 2009 all work based on averages for the total users.  This could have the net effect of pushing earlier numbers down and understating how early adopters behaved.

Wikimetrics also brought other limitations.  In the midst of conducting the research, Wikimetrics failed the research spectacularly.  Initially, administrators without a userbox identifying them as belonging to one or more of the cohorts was going to be used but a failure of Wikimetrics to process this data meant the information is presented without a control group.

While the researcher believes that given the research limitations provide some important insights into how different groups interact with English Wikipedia, readers should keep the stated issues in mind when drawing their own conclusions.

Editing Patterns

Total yearly edits

Across all groups, users with Asperger have been the most active in most years from 2003 to 2010.  Exceptions were 2004 when dyslexia had 57 average edits and obsessive compulsive disorder had 25 to Asperger’s 24.  The next was 2007 when obsessive compulsive disorder edits averaged 247 edits to Asperger editors with 208 edits. The pattern changed in 2011.  From that point on, users with dyslexia had the highest average yearly total edits.

image003

                In English Wikipedia, there has been a regular loss of total contributors. The period between 2005 and 2007 are viewed as key milestones for shedding new editors as a result of factors related to policy and technology. Some references that point to decline in these periods include https://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2012-09-24/Recent_researchhttp://strategy.wikimedia.org/wiki/Editor_Trends_Study/Results , http://www-users.cs.umn.edu/~halfak/publications/The_Rise_and_Decline/halfaker13rise-preprint.pdf .   Users with ADHD and OCD peaked in this period, and then their average yearly main space contributions went into decline.  Other groups peaked at other times or had multiple peaks.  Dyscalculia main space edits peaked in 2010. Asperger yearly editing peaked in 2010.  Social anxiety peaked in 2012 with 164 average edits per user per year, and had a smaller peak in 2007 of 110 before going into decline that saw a low of 35 total edits in 2010.  Dyslexia peaked in 2012 with 561.  Only autism showed an increase in total average yearly edits from 2012 to 2013.

 Threshold

According to https://metrics.wmflabs.org/metrics/#Threshold , threshold is:

a metric that determines whether an editor has performed >= n edits  in a specified time window. It is used to measure early user activation.  It also computes the time it took a user to reach the threshold, if they did.   Time to Threshold is also computed by this metric.  This is the number of hours that it took an editor to reach exactly n edits.  If the editor did not reach the threshold, None is returned.

image004

                Simply put, this means the amount of hours it took a newly registered user to make an edit, and then consequently, how many edits during a certain time period.  For new users with ADHD, it took them an average of 6.8 days to make their first edit to the main space.  After they made that edit, they then made 143 more edits to the main space in the 24 hour period after that initial edit.  Asperger syndrome editors took an average of 8.7 days and then made 169 total edits in that period.  These two groups had the longest gap between registering and making an edit.  Coincidentally, once they made an edit after registering, the groups had the largest total number of edits in the first 24 hours after making that edit.  Dyscalculia and autism, two of the smallest cohorts, registered and then made their initial edits within four hours of registering.  These two groups also had the smallest total number of edits in the initial 24 hour period.  Dyslexia, social anxiety, and OCD all took between two and four days to make their initial edit, and in the subsequent 24 hour period made between 59 and 87 edits.

Bytes added and removed

Another way of understanding editor behavior that the metric tool provides is the total bytes added and removed from articles.  The average sum total bytes added and average sum total bytes removed from articles by user on a yearly basis per group was added.

image005

Total bytes added for the Asperger cohort peaked in 2007 at 67,000 bytes, went down and then remained relatively level at between 29,000 bytes and 41,000 bytes.  ADHD/ADD editing peaked at almost exactly the same level as the Asperger cohort.   While both groups had a fall off, excluding 2008, the ADHD fall off in total bytes added was much higher with a range of 13,000 to 26,000 total bytes added to main space per year on average.  Likewise, OCD also peaked in 2007 at 71,000 bytes before settling into a range of between 21,000 and 40,300 bytes added on average per year.  The low in that range of 21,000 was in 2011 and then trended upwards for the next two years.  Social anxiety has almost the same pattern as OCD but with a greater range in total bytes added on average, with a low of 9,300 bytes added in 2011, a high of 45,700 in 2008, and the upward trend from the low 2011 to 39,500 in 2013.

Dyslexia user average total bytes added peaked in 2008 at 106,000 bytes on average, over double the previous year.  In 2009, the average fell to 12,800, an amount not seen since 2005. The numbers went back up in 2011 and 2012 to the 50,000 bytes level.  In 2013, the levels were half 2012. The dyscalculia cohort peaked in 2010 at 70,000 bytes on average after an average of 14,000 bytes added in 2009.  The average bytes added per contributor dropped back down in 2011 to 5,600 bytes and stayed in that low range in the following years.

Bytes removed is another way of assessing content creation.  People can remove content for a variety of reasons including vandalism removal, article rewriting, fixing citations to make them shorter, removing extraneous tags and links and images, and removing uncited material.  It is a distinct type of editing compared to adding bytes.

 image006

ADHD average total content removal peaked in 2007 at 29,000 bytes removed.  Since 2007, the average year total content removal for ADHD has decreased every year at the same time that ADHD content additions have been in the decline. Social anxiety and Asperger Syndrome peaked in the same year with the average removal of 117,200 bytes and 67,000 bytes removed on average respectively.  Dyslexia average total content removal peaked in 2013 at 57,000 bytes, almost twice the number of bytes removed as the bytes added in the same year. OCD content removal also peaked in 2007.  Outside that year, from 2006 to 2013, the level of content removal has been relatively consistent in a range of 8,000 to 16,000 bytes. Dyscalculia peaked at content removal in 2010 with 70,000 bytes removed on average.

In contrast, dyslexia had two peak years for content removal, 2007 and 2013.  Since 2011, the total level of content removal increased.  In the same period since 2011, while content removal increased, content addition decreased. Social anxiety also had two peaks, 2007 and 2013. The increase in removed content started in 2011 and continued into 2010.

Survival

Another metric for determining how a cohort behaves is their survival rate on the projects.  The survival percentage for editor cohorts was found after the first 24 hours. That is, the contributor made another edit 24 hours after they made first edit hours on the project.

image007

All groups had relatively high survival rates, with the dyscalculia cohort having a 100% survival rate with all 8 editors having edited after the initial 24 hour period.  This contrasts to the social anxiety and OCD cohorts which have survival rates of approximately 86%.  ADHD had a survival rate of 88.4%.  Dyslexia and Autism are both at a 90% survival rate. The highest survival rate outside dyscalculia was Asperger Syndrome with 92.6%.

Page creation

One of the beautiful things about English Wikipedia is anyone can create an article.  The Wikimetrics tool provides information on how many pages were created by each cohort.  This information was broken down by cohort and then by year for each cohort.

Excluding autism, the OCD cohort were on average the most prolific article creator, with the average user creating 53 new articles compared to the second most active cohort of Asperger Syndrome with 48.  The next groupings are not close to these numbers with dyslexia averaging 36, social anxiety averaging 30, ADHD averaging 22 and dyscalculia averaging 4.

 image008

                Like other yearly metrics, there are peaks in 2007.  One peak for article is for social anxiety with users creating an average of 8 articles.  ADHD also peaked in 2007 but the total average new articles created was 6, the same number for the previous year. Dyscalculia maxed out with an average yearly high of one article for 2007, 2010, 2011 and 2012. 2009’s average is 0.5 and the other years are all decimals above 0.625.  Asperger peaked in 2007, 2010 and 2011 at seven.  The only cohorts not to have a peak in 2007 were dyslexia which peaked in 2004 and 2006 with 8 articles and OCD which peaked in 2011 with an average of 11 articles created.

Learning disability and mental disorders

Asperger Syndrome

image009               The DSM-V criteria for Asperger Syndrome, found at http://www.cdc.gov/ncbddd/autism/hcp-dsm.html , includes multiple impairments in ability to use non-verbal communication, failure to develop relationships with peers, lack of ability to spontaneously share joy with others, and a lack of social and emotional reciprocity in the sharing of feelings.

Of the seven disorders included in this research, this cohort of users was the largest with 443 people transcluding userboxes that identify them as having Asperger Syndrome.

A body of research on Asperger Syndrome strongly suggests that the Internet and social media provide a medium that allow people with the disorder to successfully form social connections in an environment that enables them to feel protected.  At the same time, this body of research suggests that while they may feel safe, there is little support for an idea that social media actually assists in people in maintaining successful social engagement.

The idea of extended successful social engagement may be key for successfully contributing to English Wikipedia in the long term.  At a certain point in contributing with a user focused on content creation, interacting with other contributors will likely happen as people attempt to collaborate on an article.  Failure to socially interact may cause problems for the contributor.  On the other hand, it may not if the person cannot pick up the social cues that indicate problems so long as the contributor engages in a space where they are generally left alone and are surrounded with like-minded individuals.  Much of the potential issues unique to people with Aspergers are speculative given the lack of the body of research, but in terms of best guess, the social interaction component would almost certainly cause the biggest problems given the diagnostic criteria.

 image010

                Wikipedia is generally viewed as having peaked around 2006 or 2007 and having been in decline since then, with a secondary peak, explored much less in literature, occurring around 2010/2011.  Yearly patterns would suggest this cohort followed that pattern, with the total average edits per year for it peaking in 2010 with 301 edits, and the yearly average for total bytes added peaking in 2007, with bytes added averaging 67,000 and then remaining relatively level at between 29,000 bytes and 41,000 bytes from 2008 to 2013.  Bytes removed peaked in 2008, with an average 18,700 bytes removed per user for the year.  Total pages created also peaked in 2007, and 2010 and 2011, with the average member of the cohort creating seven articles in each of those years.

As a group, this was one of the slowest in getting started, but once they started they appeared to stick with the project. On average, it took a member of this cohort 210 hours (8.7 days) to make their first edit.  In the 24 period following their first edit, they made an average of 169 total edits to the main space.  The group had one of the highest survival rates, trailing only dyscalculia. 92.6% members of this cohort made another main space edit 24 hours after they made first edit hours on the main space.  The group was also on average the second highest article creators, totaling an average of 48 new articles each.  The high average is impressive when one considers that English Wikipedia has 20 million registered users, and the top 5,000 article creators have all created 100 or more articles.

The constraints of the easily available metric data looking at social engagement limit the potential usefulness at looking at how Asperger Syndrome contributors interact with English Wikipedia.  As a cohort though, they appear much more likely to stay and contribute than other participants.  Of the groups included in this analysis, they appear to be in the middle of the pack when it comes to both adding and removing content.  Otherwise, they are leaders in other metric categories.  Additional research, specifically focusing on the type of content contributions and the type of non-space social engagement may provide a better picture.  Further analysis may want to explore the comparative level of content addition to talk pages when compared to their peers in other cohorts.

Attention Deficit Hyperactivity Disorder

image011                According to the DSM 5 for Attention Deficit Hyperactivity Disorder has two types, hyperactivity and inattention, with the criteria for each type found at http://www.cdc.gov/ncbddd/adhd/diagnosis.html . The diagnostic criteria involve meeting at least six of the characteristics in two distinct settings.  They include inability to sit still, losing things necessary to complete tasks, avoidance of tasks that involve extended mental effort, difficulty being organized, easily distracted, always on the go and moving, interrupting, having difficulty waiting for a turn, and talking excessively.

This cohort was the second largest of the groups looked at with 405 contributors who identify as having ADD or ADHD.  A Canadian survey  of 723 people with disabilities, found at http://dc160.dawsoncollege.qc.ca/pubs/abPreliminaryFindingsOnSocialMediaUse.pdf , found that people with ADHD/ADD are one of the largest disability groups using social media, representing 18% of the population.  This compares to 21% with a chronic medical / health condition, 11% with visual impairment and 10% with an auditory impairment.  It also compares to 32% psychological / psychiatric disability and 29% self-reporting a learning disability.  In the United States, estimates put the percentage of people with ADHD in the population at approximately ten to twelve per cent.  In Europe, it is around five per cent.  Little research has been done though on issues surrounding how people specifically with ADHD or ADD interact with social media and potential challenges they face in using it.  There has been some broad generalizations made that suggest the Internet may play a contributing factor in the development of ADHD/ADD, and there have been a few texts written about Internet addiction by people who have ADHD/ADD.  Movers, Dreamers and Risk-Takers, Unlocking the power of ADHD by Kevin Roberts says the author finds that people with ADHD are often very good at finding entertaining things on the internet the make people laugh, and thinking outside the box in creating things. Like Aspergers, this appears to be another area where research is needed.

This means that at best with no research to fall back on, how ADHD people are likely to interact with English Wikipedia would be based on pure speculation and conjecture.  One such piece of conjecture is that ADHD users are likely to have attention problems when dealing with English Wikipedia.  They may be likely to super obsess and focus, creating large volumes of content and edits, or they are likely to not find English Wikipedia engaging, make a few edits, get distracted and not come back.  Given the speculative extremes, it would be hard to understand how contributors participate without individual contributor information to see if a bell curve exists.

Thus, surprisingly, on average, it takes a user 165 hours (6.8 days) to make their first edit.  ADHD users do not register and immediately start editing. Less surprisingly, in the 24h period following their first edit, they made an average of 144 total edits to the main space.  When they do start, they do a lot of editing. In terms of sticking with English Wikipedia, ADHD had a survival rate of 88.4%, relatively low amongst the disorders examined with only social anxiety at 85.6% and OCD at 86.1% being lower.  Most people with ADHD come back and make an edit after their initial 24 hour editing period.  As a group, ADHD contributors were amongst the least likely to create new articles with the average member of the cohort creating 22 articles.  Only dyscalculia with 8 had a lower number of the sampled disorders.  In contrast, Aspergers, which has a high rate of co-presenting with ADHD, contributors averaged 48 articles created.

 image012

                Like Aspergers, ADHD had peaks around the known and documented English Wikipedia peaks. Since 2007, the total average edits by users with ADHD have declined by year.  In 2007, the total average number of edits per year was 184, and by 2013 had declined to 50.  This drop off was more precipitous than other groups suggesting potentially that once a contributor with ADHD fails to engage, they really fail to engage and are much less likely to come back than other groups. In 2007, ADHD/ADD total byte added peaked at almost exactly the same level as the Asperger cohort.   While both groups had a fall off, excluding 2008, the ADHD fall off in total bytes added was much higher with a range of 13,000 to 26,000 total bytes added to main space per year on average.   This supports the premise that when ADHD users disengage, they disengage more rapidly than other disorders examined.

In conducting future research about English Wikipedia editors with ADD/ADHD, data collection should likely emphasize examining user specific patterns to see if editing mirrors expectations regarding early obsessive editing followed by a precipitous drop off.  It should also query editing pattern conditions where ADHD users manage to consistently engage the project over extended periods of time.

Autism

image013              The DSM-V criteria for autism, found at http://www.cdc.gov/ncbddd/autism/hcp-dsm.html , require that six or more criteria be met from three or more categories. One category includes the lack of ability to use non-verbal communication, failure to develop relationships with peers, lack of ability to spontaneously share joy with others, and a lack of social and emotional reciprocity in the sharing of feelings. A second category includes delays in the development of speech and motor functions, repetitive or idiosyncratic patterns of speech, inability to engage in sustained conversation and inability to play make believe.  The third category includes obsession with parts of objects instead of the whole, repetitive full or partial body movements, adherence to routines that are not required for functioning, and abnormal obsessions with patterns.

As stated earlier in the research, while autism is included here, the sample size is very small, second only to dycalculia in terms of size with 20.  This sample likely does not reflect the broad array of people who have autism which includes four groups according to DSM 5, and higher functional people with autism are almost certainly picked up in the Asperger Syndrome category.  Outside that group, the symptoms of autism generally suggest that people with autism would have extreme difficulty effectively contributing to English Wikipedia.  This includes such problems as difficulty reading, difficulty with language acquisition and retention, and problems with motor coordination.  It seems likely, though it is not supported by looking at the user pages of the 20 individuals with the autism user boxes, that some of the users who put the autism user boxes on their user pages are more likely to be people who support autism related topics given the visibility of and activism surround the disorder.

There is very little research on social media usage by people with autism, and an awareness amongst researchers of the lack of research in this area according to a July 2013 article published in Computers in Human Behavior available at http://www.sciencedirect.com/science/article/pii/S0747563213000708 .  What research appears to exist focuses on the use of social media by caretakers, or the use of audio-visual related imagery to assist people who have autism.  A 2011 study published in the Journal of Autism and Developmental Disorders available at http://link.springer.com/article/10.1007/s10803-011-1413-8 found that only 13.2% of young people with autistic spectrum disorder used social media, which the paper defined as e-mail, chatting and the internet.   They noted this rate of usage was much lower than other learning disability types. A similar conclusion was reached in a July 2013 also published in Journal of Autism and Developmental Disorders.

 image014

                The data around autism was at times so completely against pattern for other disorders that including it on graphs would have made the graphs difficult to read.  This includes average total edits, specifically for 2013 where they averaged 1,082, over double that of the next group, dyslexia with 512.  They were also one of the quickest groups to start editing once they registered: On average, it took a user in this cohort 3 hours (0.1 days) to make their first edit.  In the 24 period following their first edit, they made an average of 7 total edits to the main space.  They also had the highest rates of article creation, averaging 178 articles each.  This average would put all contributors on the list of top 5,000 article creators on English Wikipedia where the 5,000 mark involves creating at least 100 articles.  They led all disabilities on a yearly basis in article creation with an average of 8 articles in 2009, 84 in 2012 and 68 in 2013.  Similar skewing also happened for total bytes added, with autism leading all groups for bytes added in 2009, 2010, 2011, 2012 and 2013. Only in two other years did any other disability group match autism for average bytes added.  For 2013, the total average bytes added was six time the next highest group, 239,000 bytes on average compared to 32,000 for Aspergers Syndrome.  The only area where autism did not especially stand out was survival rate, which at 90% put it in the middle of all groups and about equal to that of dyslexia.

Any further research about users with autism needs to do a better job at identifying users, what type of autism they have and examine why there are these wild outliers.

Dyscalculia

Dyscalculia was not included as a separate disorder in the fifth edition of the DSM-V.  A lot of research material published by governmental bodies in the United States refer to it as a learning disability.  A paper published in the November 2012 Deutsch Ärzteblatt International found at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3514770/ defines dyscalculia as “as difficulty acquiring basic arithmetic skills that is not explained by low intelligence or inadequate schooling.”  Of the learning disabilities and disorders examined, this one is the one with the least documentation about it.

On the face of it, this particular learning disability appears unlikely to create any sort of difficulty in contributing to Wikipedia except possibly for maths related articles.  Most Wikipedia articles do not have math related problems.

image015

                With only 8 members of this cohort, it was the disorder with the smallest number of people of all the learning disabilities looked at.  The editing patterns for this group, like autism, were very distinct.  The peak for average total yearly edits, total bytes added and total bytes removed all peaked in 2010, not in 2007 when Wikipedia had its well documented peak. The dyscalculia cohort average total bytes added peaked in 2010 at 70,000 bytes on average after an average of 14,000 bytes added in 2009.  The average bytes added per contributor dropped back down in 2011 to 5,600 bytes and stayed in that low range in the following years.  Editors in this cohort were also the quickest of all groups to register and then make their first edits, on average making their first edit within two hours of registering.  They were also outliers in that the total edits they made in their first 24 hour editing period was 2, way below the average for other disability groups.   The cohort had the highest survival rate of all disorders at 100%, while at the same time having the lowest average total number of articles created at 4.

With a disability that appears unlikely to substantially impact their ability to edit, this group registered and then made a only few edits, peaked later than other groups, created much less content in terms of new articles and bytes added.  Further research is needed to see if this disability actually impacts a person’s ability to contribute productively to the project or if the patterns observed are actually more broadly reflective of editing populations without mental disorders or learning disabilities.

Dyslexia

Dyslexia was not included as a separate disorder in the fifth edition of the DSM 5.  A lot of research material published by governmental bodies in the United States refer to it as a learning disability/learning disorder.  The National Institute of Neurological Disorders and Stroke in the United States define dyslexia at http://www.ninds.nih.gov/disorders/dyslexia/dyslexia.htm as

a brain-based type of learning disability that specifically impairs a person’s ability to read. These individuals typically read at levels significantly lower than expected despite having normal intelligence. Although the disorder varies from person to person, common characteristics among people with dyslexia are difficulty with phonological processing (the manipulation of sounds), spelling, and/or rapid visual-verbal responding.

The disorder potentially can have a huge impact on a person’s ability to contribute to English Wikipedia.  Like problems faced by users with vision impairments, an easy technological fix has been implemented to provide an accommodation to make it easier to contribute.  This accommodation comes in the form of a font documented at https://en.wikipedia.org/wiki/Wikipedia:Dyslexic_readers though the implementation of it was very recent, only occurring in mid to late 2013.  Beyond Wikipedia, there has been extensive research done on how people with dyslexia use social media, and how the disability impacts on their social interactions.

 image016

                The cohort of users with dyslexia is one of the smallest of all groups looked, with 139 people including a userbox identifying themselves as having dyslexia.  Editing trends for the group buck some of the established trends, with bytes added peaks occurring in 2008, not 2007, and in 2012, not 2010 or 2011.  Bytes removed on average also peaked in 2013, which is particularly noteworthy because that peak followed a major dip in content added and the total bytes removed exceeded the average total bytes added. Dyslexia user average total bytes added peaked in 2008 at 106,000 bytes on average, over double the previous year.  In 2009, the average fell to 12,800, an amount not seen since 2005. The numbers went back up in 2011 and 2012 to the 50,000 bytes level.  In 2013, the levels were half 2012.  The total edits per year also did not follow trends, with a peak in 2012 of 561 edits per user on average for the year.  In other categories, dyslexia finished generally in the middle of the pack and not against trend.  It had a survival rate of 90%, which means that 9/10 editors made another edit after their initial 24 hour editing period.  The cohort members averaged 36 total articles created compared to 30 for people with social anxiety and 48 for people with Asperger Syndrome. On average, it takes a user with 66 hours (2.7 days) to make their first edit.  In the 24 period following their first edit, they made an average of 59 total edits to the main space.  This again put them in the middle of the all groups looked at.

Further research is needed, specifically in terms of the impact of the accommodation on ability of people with dyslexia to contribute.

Obsessive Compulsive Disorder

The DSM-V criteria for Obsessive Compulsive Disorder, mentioned at http://emedicine.medscape.com/article/1934139-overview , includes having obsessions and compulsions.  The obsessions include thoughts are recurrent and persistent, and the person with OCD actively tries to suppress them.  The compulsion part involves repetitive behaviors aimed at reducing anxiety caused by these thoughts, with the outcomes of the behaviors completely removed from the thing trying to be avoided.   Some research estimates that 2.5% of the US population has some form of the disorder.

Like the situation for ADHD, some research appears to suggest that the Internet has contributed to a rise in the number of diagnosis of OCD.  This particular disorder also has the problem that OCD has at times become shorthand for “obsessively checking social media sites”.  Outside of this, there again appears to be a research void on how people with OCD interact with the internet.

Given the lack of research on OCD and the disease manifest itself in terms of social media usage, how people with OCD contribute to English Wikipedia is a purely speculative exercise.  Wikipedia editors with OCD may approach editing from a ritualistic perspective by checking certain pages first, or needing to compulsively update their user page with their online editing status before starting and after they finish an editing session. Potentially, people with OCD may develop obsessive thoughts about articles they are working on or in relation to interactions they are having that put the person in a pattern loop of constantly needing to clarify a position or fix an article.  Further speculation might suggest for contributors where OCD manifest itself as part of their contributing to the project, the user may have higher levels of editing, and consistent levels of editing as the person engages in ritualistic behavior.

 image017

                This cohort was in the middle in terms its size relative to other disabilities examined with 238 included as having self-identified as having OCD. On average, it took a self-identified user with OCD 102 hours (4.2 days) to make their first edit.  In the 24 period following their first edit, they made an average of 87 total edits to the main space. Both results put the cohort in the middle of the pack for disorders examined. For bytes added, OCD peaked in 2007 at 71,000 bytes added before settling into a range of between 21,000 and 40,300 bytes added on average per year. The low in that range of 21,000 was in 2011 and then trended upwards for the next two years.  The 2007 peak was consistent with the historical peak on English Wikipedia. The peak for bytes removed was 2008, a year after the peak for bytes added. The OCD cohort was on average the most prolific article creator group, with the average user creating 53 new articles.  The cohort had a peak in 2004 and 2007 of an average of 6 new articles created per user, before going on to peak again in 2011 with 11 total articles created on average before going into decline again.  This trend roughly mirrors patterns on Wikipedia which had a second major editing spike around 2010/2011.  Despite what looks like high rates of editing and contributing, OCD users had the second lowest rate of survival on the project with 86.1% of contributors making an edit after their initial 24 hour editing period.  Only social anxiety had a lower rate.

Further research is necessary to understand to the extent that OCD impacts the ability of a person to contribute to English Wikipedia as it is not immediately clear how OCD would manifest in terms of contributing to Wikipedia.  Additional research is also needed to understand specific editing patterns to determine if people with OCD are engaging in ritualistic behavior on English Wikipedia, and, if they do, how that behavior impacts on their social interactions with other users.

Social Anxiety

Social anxiety is shorthand for social anxiety disorder or social phobia.  The DSM-V criteria found at http://www.dsm5.org/Documents/Social%20Anxiety%20Disorder%20Fact%20Sheet.pdf  state “a person’s fear or anxiety be out of proportion—in frequency and/or duration—to the actual situation. The symptoms must be persistent, lasting six months or longer.”  It goes on to state “To be diagnosed with social anxiety disorder, the person must suffer significant distress or impairment that interferes with his or her ordinary routine in social settings, at work or school, or during other everyday activities.”

Some of the research on Internet usage by people with social anxiety concludes that people with the disorder use the Internet as a way of more comfortably engaging with people.    At the same time, research suggests that social anxiety can be a predictor for decreased use of social media and that people with social anxiety are less likely to use the Internet to socialize.  Some research has found that anonymity in communication can assist a person in overcoming their social anxiety and enable them to contribute to social media. In terms of Wikipedia, little research appears to have been done about how social anxiety impacts an editor’s ability to contribute to the project. Again, this is an area that calls for pure speculation in terms of hypothesis formation.  Given the body of research related to this, contributors with social anxiety might contribute without any goal of trying socialize, and hoping to remain anonymous. When confronted with painful social interaction, the user may be more inclined to leave than other users.

 image018

                208 users with social anxiety were identified through inclusion of a userbox on their user page, putting them in the middle in terms of cohort size of all the groups examined.  As a group, they had the lowest survival rate following their first 24 hour editing period.  It was 85.6% compared to the next lowest group, OCD, which had a survival rate of 86.1%.  They were in the middle when it came to length of time after registering to make their first edit, 77 hours (3.2 days). Making 69 edits on the main space on average in that first 24 hour editing period, they were also in the middle of all the examined cohorts.  They were also pretty close to the middle when it came to average number of articles created at 30 since 2002.  Their peak yearly article creation  average of 8 articles in 2007 also followed the major trends, though the total number of articles topped all groups for that year. Social anxiety peaked in 2012 with 164 average edits per user per year, and had a smaller peak in 2007 of 110 before going into decline that saw a low of 35 total edits in 2010.  This did go against established trends. Social anxiety has almost the same pattern as OCD but with a greater range in total bytes added on average, with a low of 9,300 bytes added in 2011, a high of 45,700 in 2008, and the upward trend from the low 2011 to 39,500 in 2013.  The yearly trends suggest that the type of editing taking place changed from 2007 until the present.  Total bytes increased from 2012 to 2013, while total article created dropped and the total average yearly edits declined.

Additional research is required to understand more about how social interactions on English Wikipedia connect to and impact a user’s editing experience.  An explanation needs to be provided to explain why certain numbers went up while others went down to see what the causes for this were.  As it stands, it suggests that people with social anxiety changed their editing patterns between 2012 and 2013.

Conclusion

Too much of the potential conclusions are based on pure speculation because of the inability to get a control group and the existing research vacuum.

Different groups interact with English Wikipedia in different ways.  This is not unexpected given that each group has their own unique issues as to how they potentially relate to Wikipedia.  Some of these issues could potentially be useful if a way could be found to better utilize and attract users with these characteristics.  This research also highlights the existing void in research in this area, and the problems of some of the existing metric tools in terms of pure understanding of different types of editor actions as they relate to individual diagnostic criteria.  At the same time, it also should call attention to the problems with how to frame Wikipedia research in this area when there is tension between academic epistemologies when it comes to potential problem and solution identification.

At the end of the day, some of the problems faced by people with mental disorders and learning disabilities are minor ones that can be overcome with accommodation like dyslexia.  Some are unlikely to impact Wikipedia editing like dyscalculia.  The characteristics of some  mental disorders are ones that could be positive if harnessed appropriately for editing Wikipedia in a mentally healthy way.  This includes making large numbers of edits around a topic that a person knows well, or doing repetitive Wikipedia maintenance tasks.  The biggest challenges for editing for a number of mental disorders involve social interactions, and solutions for dealing with potential problems perceived as unique to these groups are likely to have positive flow on effects for the rest of the Wikipedia content creator community.

This post is available in PDF format at Learning disabilities and English Wikipedia .