Re: [Usability] Gnome Research - Sociological Surveys

[Celeste Lyn Paul <celeste kde org>]

My .02...

On Friday 17 April 2009 09:38:41 am Anton Kerezov wrote:

> Well I haven't thought of them all just yet but here's a sample list:
>
> - How old are you? (determines the majority of users Gnome is used by)

Also should ask education, country, primary language, OS language, if they 
have a laptop/computer/both, etc. so you can get a profile of the type of user 
they might be.

> - How do you want to connect with others online? (Will show how the user
> interacts with other users online and if he wants the software integrated
> in the DE)

A better question would be "Do you want to connect with others online?" and 
"How have you connected with others online in the past?". You don't want users 
to design their own product.

> - What would you change in Gnome 3.0?

I would split this up into a few targeted questions about GNOME and then allow 
for additional comments.

> - Do you feel you can work faster if the software allows it (is more
> usable)?

This type of question is hard to for users to answer because it requires them 
to self assess without context which leads to speculation, not reporting.

> - Do you care how your desktop looks?

This is too general, ask more targeted questions.

> Some concrete, maybe:
> - Are you happy with the menubars and menu system usability?

Again, too general. ask a few targeted questions and then allow for general 
commenting. Also, try not to use the work "usability". Too many people think 
it is satisfaction and dont think of it as an engineering quality. Instead ask 
specifically if they think the menu system is difficult to use or learn, clumsy, 
inefficient, ugly or pretty, etc.

> - How many steps you want to open a file for edit?

This is a bad question. You are asking the user to design their own product. 
This isn't something a user can self assess, and it can't be generally applied 
to all users in all contexts and all tasks. Also, there is a lot of research 
which shows that the number of steps to do anything doesn't matter, it's 
whether the user is confident in the path they are tasking. 

When youre building your survey, instead start with what you want to learn 
from it instead of starting with the questions. The quality of data you get 
out of a survey is more important than the number of responses you get.

-- 
Celeste Lyn Paul
KDE Usability Project
usability.kde.org