no image

Cleaner more accurate tree for 2021

Privacy Level: Public (Green)
Date: 29 Nov 2020
This page has been accessed 172 times.

Under construction

This page is a summary of the ideas provided by the answers and comments (as of 28 November 2020) on the thread What is the best way to achieve a cleaner more accurate WikiTree for 2021?

Please note: This is a collection of ideas, summarized here for further consideration. Nothing here should be regarded as a policy proposal in its present form.



Thirty-three answers and dozens of comments were posted in response to the G2G question "What is the best way to achieve a cleaner more accurate WikiTree for 2021?" They can be sorted into three general groups dealing with Membership, Infrastructure, and Data. Within each of these groups are several subgroups, and in every case they include a variety of conflicting ideas of varying practicality.

Newcomer communication: Contributors mentioned that some newcomers do not understand WikiTree's most basic concepts of a shared tree and collaboration; that sourcing and biography writing should be emphasized; and that WikiTree's use of DNA should be better explained.

Member education: A variety of educational tools were suggested, some voluntary and some compulsory.

Problem members: Some punitive ideas for dealing with members with unsourced profiles and long suggestion lists were mentioned. One person suggested that the process for dealing with members needing help should be less punitive, i.e., not MIR-based.

Standards: Many people commented that sourcing standards should be stricter: That primary sources should be used; that personal knowledge and unsourced trees should not be accepted; that when creating profiles, the option to add sources later should not be allowed; and that FindAGrave and Wikipedia should be disallowed as sole sources for a profile. Streamlining rules and guidelines was also mentioned. (Here it should be noted that there is an existing proposal on G2G to extend the pre-1700 sourcing requirements to all pre-1800 profiles.)

Site organization and member support: Many people commented that the Help pages should be better written and easier to find. Other suggestions included new prompts and status summaries, automated tasks, and shortened wait times for trusted list and merge requests and Unresponsive Profile Manager filings.

Rewards systems: Most suggestions mentioned the adverse effects of rewarding quantity of contributions over quality. Not generally addressed was the possibility that loss of such incentives to members would discourage WikiTree activity.

Gedcoms: The idea of eliminating gedcom uploads was controversial. Other suggestions were made to limit their size, or require various pre-upload reviews or certifications. To clean up existing gedcom-generated profiles, targeted cleanup efforts were recommended.

Profiles: Suggestions related to both quality monitoring and cleanup. Some commented that new members should be restricted in the number of profiles they create until their work has been reviewed. A number of people suggested that certain unsourced or poorly documented profiles be deleted (which is contrary to current WikiTree policy).

Many suggestions in all categories would require changes in WikiTree policies, procedures, or official Help pages in order to be implemented. For members who want to contribute to the goal of making WikiTree cleaner and more accurate in 2021, individual efforts and participation in the organized WikiTree cleanup efforts would seem to be the most achievable.

Below is a listing of the answers and comments posted to G2G, noting the names of those posting, grouped by category.


Newcomer communication

• explain concepts of shared tree (including importance of not creating duplicates per Cooper answer) and collaboration (Carnahan, Morgan)
• emphasize importance of sources (Morgan) including sources that document relationships {McMichael}
• encourage biography writing, not just data (Barry answer); note that biography-writing app can be used (Corbin); see
• better explanation of how we use DNA (Black)
• create welcome message to new members explaining shared tree and collaboration and including list of Help page links (Kelts)
• be more welcoming to new members by refraining from criticizing minor compliance issues (Gitzl answer)

Member assistance, tools, and education

• provide better support to newcomers by re-thinking existing structure of Greeter/Ranger/Mentor/etc. system and by encouraging all experienced members to reach out to newcomers (Vorenhout answer); better access to Mentors (Sherwin answer)
• more examples of improved profiles (Gadsby); improve the most-visited profiles to serve as good examples (Gitzl answer)
• best practices self-quiz (Carnahan)
• have new members improve three existing profiles not related to themselves (Knight answer)
• mandatory sourcing tutorial (Tilton answer, Black, Thiessen, Knight answer)
• create and/or require a new member "Voyage" (Corbin, but not compulsory per L. Ford answer)
• mandatory Profile Improvement Project "Voyage" (Berryann, Jamison answer; or voluntary per Kelts, L. Ford)
• create more and better training videos and promote them more (Spires, Sherwin answer)
• make better use of weekly WikiTree e-mails for providing member education and tips (L. Ford)
• improve pre-1700 quiz (Carnahan)
• emphasize and enforce ethics (Corbin)
• emphasize overall quality and reasonableness of work and not simply numbers or types of sources (Barton answer)

Problem members

• flag members consistently creating unsourced profiles; consider rescinding membership (Tilton answer, Black, Vorenhout, Knight answer)
• follow up on long suggestion lists that are not dealt with (Corbin)
• members with unaddressed suggestion lists should be prevented from creating new profiles (Knight)
• create process other than Mentor Intervention Requests for dealing with members who simply need help understanding WikiTree (Andersson answer)



• better definition of what a source is; stricter sourcing standards (Corbin, Morgan, Cooper answer, Steesy answer, Black answer); use primary sources (Heyman answer); see
• remove personal knowledge, unsourced trees, and "to be added later" from sourcing choices for newly created profiles (Forbes, Steesy, Dalton answer, L. Ford answer, Barry answer, Andersson answer re trees, Pickens); profiles documented only by "unsourced family trees" should be marked as unsourced (McMichael answer; response from Knight says can't be done by editbot)
• also disallow FindAGrave and Wikipedia as only sources (Andersson answer)
• adopt Robin Lee proposal to change pre-1700 standards to pre-1800 or even pre-1850 (Kelts, Andersson answer)
• prioritize quality over quantity (Morgan; but consider viability issues per Kelts)
• reduce the number of rules and guidelines, then enforce the ones remaining (Morgan); streamline rules and regulations (Gitzl answer)
• require place information on profiles (Schell)

Site organization and member support

• make Help pages easier to find; improve Help index (Corbin, Morgan, L. Ford answer, Knight answer)
• improve Help pages (L. Ford answer, Knight answer)
• add new Apps item(s) to Help menu (Knight answer, Andersson)
• improve Data Doctors' help pages and suggestion explanations (Morgan)
• develop pop-up prompt asking for source when profile data is changed (McMath answer)
• make it easier to add sources (Kelts, Sherwin answer)
• automate tasks wherever possible (L. Ford answer)
• improve the gedcom upload process (Sherwin answer)
• provide members with summaries of their profiles showing numbers of primary sources (Heyman, counter argument from L. Ford)
• shorten wait times for TL requests, merge requests, and UPM filings (L. Ford answer)

Rewards systems, contests, and -Thons

• evaluate adverse effects of rewarding quantity of contributions (Barton answer)
• get rid of contests that reward quantity (Morgan; counter argument from Kelts)
• get rid of Club 1000 badges, among other reasons because the contribution counting system includes trivial changes and can be gamed (Morgan; counter arguments from L. Ford, Gadsby, Paul)
• review quality of work before awarding Club 1000 badge (McMichael answer; response from Vorenhout that it can't be done)
• get rid of Club 100 and Club 1000 badges (L. Ford)
• create new badge to incentivize newcomers (B. Smith answer)
• thank people more often (L. Ford)


• decisions about cleanup strategies should be data-driven: sources of inaccuracies should be evaluated first (Weddington answer)


• qualifications/certification for importers (Carnahan, Cooper answer, Sell answer)
• size limits for imports (Carnahan, Cooper answer, Black answer, Conley, Andersson answer)
• pre-import review of trees before import is allowed (Kelts)
• discourage gedcom imports (Black answer); eliminate gedcom imports (Morgan, Berryann answer, Richardson, Zipperer, L. Ford; counter argument from MacBeath, Conley, Vorenhout)
• promote use of AGC app for cleaning up newly imported gedcom profiles (L. Ford)
• create list of existing gedcoms in order to facilitate better monitoring of their status (Knight answer, MacBeath, Starjak)
• create category for gedcom-generated profiles so they can be more easily found for cleanup (Cooper answer)
• focus -thons on gedcom cleanup (Corbin)
• "gedcom of the week" cleanup projects (Connolly answer)
• Correct-a-Ged instead of Correct-a-Thon in January (Haywood answer, Sheppard); see
• projects should have gedcom teams to do cleanup (L. Ford)


Quality monitoring
• label new profiles as "new" so that they may be more easily monitored by projects (Terink, Zipperer, Hall; Knight suggests volume could be problem)
• projects should have data quality teams (Terink)
• develop profile rating system and restrict members from adding new profiles if some percentage of their additions are of poor quality (Vorenhout answer, Hunt, Cooper answer and Spires also re rating system); see (Hunt answer)
• create category for "Maybe Sourced" (Cooper answer)
• limit number of profiles new members can create or adopt until review of their work shows conformance to WT rules (Staub answer)
• review new member work after first 100 profiles (Knight answer)
• review all members for profile style and sources (Knight answer)
• do not allow members without pre-1500 certification to adopt pre-1500 profiles
• neglected profiles - remove inactive PMs after one, three, or five years of inactivity (Corbin; Steesy answer, L. Ford answer, Tilton, Andersson)
• revise privacy standards: if deceased over 50 years and not attached to living people, should be open (L. Ford answer; counter argument per Paul)
• weekly challenges to improve profile accuracy (Gitzl answer)
• focus cleanup efforts on profiles created before 2016 (Andersson answer)
• delete abandoned junk profiles after six months' warning (Farler answer, Tilton, Gierszewski, L. Ford; counter argument from Knight, Peterson, Nelson)
• delete unsourced unconnected profiles (Fuller; counter argument from Nelson--and Nelson says it's not going to happen)
• delete post-1925 unsourced profiles after one year (Carnahan)
• delete profiles with no date or location information (L. Ford answer; counter argument from Knight; hide option per Vorenhout)
• hide pre-1500 profiles until adopted by projects and brought up to acceptable standards (Andersson answer)


Comments: 4

Leave a message for others who see this profile.
There are no comments yet.
Login to post a comment.
I see no mention of

"Accuracy can only be improved by adding primary sources to profiles of which the Birth, Marriage and Death records are the most important three"

posted by Louis Heyman
Thank you. I have added a mention of primary sources to the first bullet point under Standards.
posted by [Living Kelts]
I don't think you could ever make it a requierment, but you can encourage it and provide incentives like the contribution badges are an incentive to make contributions. So no, I do not think you can require it.
posted by Louis Heyman
Further clarified in accordance with your comment.
posted by [Living Kelts]
edited by [Living Kelts]