Ubuntu-devel-discuss signal-to-noise survey: possible analyses

This page discusses a few of the analyses of the data that I plan to do. I may not do all of these, and may do others - I will have to go where the data takes me.

Analysis of the signal-to-noise section should answer the questions "is the list too busy, or not busy enough?", "to what degree is noise a problem?", and "is the noise on the list actually noise, or just signal in the wrong place?". Bar graphs of answers to each question should help answer these questions. If there are enough responses, it may be possible to do individual graphs for developers, for regular contributors, etc.

Analysis of the what is signal? section should answer the questions "how confident are people about the purpose of the list?" and "what degree of (dis)agreement is there about the purpose of the list?". It should be possible to see confidence levels using a bar graph of answers to the first question in this section. To answer questions about (dis)agreement, I will have to categorise the examples that people give. I should then be able to produce a "top ten" list for each question, along with the overall numbers for each category of answer.

Analysis of the remedies section should answer the question "what types of things are people willing to do?" and "what are the barriers stopping people from doing things?". Answering the first question should be possible with bar graphs again, whereas answers to the second are likely to be less scientific. I hope to be able to pull some common themes from people's explanations, although this will depend entirely on the types of answers.

Analysis of the about you section should answer the question "what sort of person subscribes to this list?". If there are enough responses, it should also make it possible to provide a more detailed breakdown of answers in the previous sections.