Ubuntu Membership Process Survey Report and Next Steps

The Ubuntu community, like many communities has different methods in which we identify and elevate the privileges of those who actively participate and contribute good work. As an example, if you do great packaging and development work, you can get upload access to the archive with MOTU or Core-Dev.

A core means of identifying such contributions is the Ubuntu Membership process. Ubuntu Members are those contributors who have made a significant and sustained contribution to the project across one or more of the many different ways you can participate (e.g. LoCo Teams, Documentation, Translations, Packaging, Testing etc).

Recently there was a little debate about some aspects of the process and how well it is functioning, and it struck me that we don’t have enough data to help get a good idea of the feedback and opinion of those folks who have been through the Ubuntu Membership process. As such, I put together a survey that went out to everyone who has applied to be an Ubuntu Member in the last few years to get their feedback. I am presenting the results of this survey today.

Generally the feedback is very positive, but there are definitely some areas where we can explore how to refine and improve the current process.

Download The Report

Tomorrow, Tues 6th Sep 2011, we have a meeting scheduled with the Community Council at 9pm UTC / 2pm Pacific / 5pm Eastern / 10pm UK / 11pm Europe in #ubuntu-meeting on freenode and we will discuss this topic there. I recommend everyone who wants to participate at the meeting downloads the report and familiarizes themselves with the findings.

  • htorqe

    Thanks for creating this, was an interesting read! :-)

  • Randall

    Thanks for conducting this survey and for documenting it so thoroughly, Jono.

    Your efforts are a big part of why this community is awesome. And with good data, it will become even more awesome. :)


  • http://bapoumba.wordpress.com/ bapoumba

    So this is only for the Regional Boards process, right ? Other boards can grant Ubuntu Memberships :)

  • http://bapoumba.wordpress.com/ bapoumba

    So this is only for the Regional Boards process, right ? Other boards can grant Ubuntu Memberships :)

  • http://twitter.com/jspaleta Jef Spaleta

    The report breaks out the demographics of pre-selected invitation only respondants correct? 

    How did you select who received and invite to participate in the survey? Was the invitation list randomly selected via some sort of algorithm from a pre-existing larger list of candidates or was these handpicked invites? Did you pre-select an equal number of Member non-member invitees? Did you pre-select an equal number of Canonical versus non-Canonical invitees?  Or perhaps did you select invitees along established demographic proportions?

    I am concerned that your invitation methodology might have introduced selection bias into your results. Accounting for demographic selection bias in the sampling of a diverse population is a big part of professional polling and it something that will really bite you in the end.  One way to account for it is to double check how far off your sampling group is in terms of demographic mix compared to the overall population.

    Can you provide corresponding information about the size and demographics of the entire Ubuntu Membership base? How many Ubuntu members are there in total, what is the breakdown of the Membership board approval percentages for the full Membership? etc…

    Also 5% of your respondants were not members. How many non-members did you invite? Was the response percentage lower or higher for the non-members than for the members?

    metrics are good, but sound methodology is vital. -jef

  • Omer Akram


  • Jchristym


  • Anonymous

    Whats the point of this report?  To be pretty and feel good?  

    Please don’t discount sound reasoning.  I didn’t take it as Jef attacking the Jono or the report, rather as very considerate questions and constructive ideas to help.  

    Being open and transparent has risks but part of the benefit is it allows for improvement.  In other words, everyone has their own domain of expertise.  Jef obviously knows something about statistics.  How is this different than a domain expert finding a bug in open-source code and submitting a bug?In other words, the report doesn’t actually “have meaning” if methodology is bogus.  It in fact can lead you to totally incorrect conclusions.

  • http://profiles.google.com/nathanlee2 nathan lee

    “metrics are good, but sound methodology is vital” – agreed. Fortunately, because the goal was nonspecific, they didn’t need an accurate and specific study adjusted for demographics. The purpose  of ‘assessing the quality of membership applications’ could have been achieved with as little as… well really 1 non-representative but intelligent person. Inviting every applicant in the last 2 years simply gives more chances to get useful feedback.

    I don’t think this claims to be a scientifically accurate assessment of what opinions are, in which case it would need more information to compare opinions adjusted for demographic consistency. But even then – it’s hard to beat a large sample size with invites sent to a non-random EVERYONE that applied. I know for a fact that radio station ratings are far less representative.