@Halfak, a few comments:

Describe the steps you took to arrive at this edit
What do you mean by "arrive"? What do you mean by "edit"? Do you mean specifically clicking the "edit" or "save page" button on an article? What would be considered a "step"? For example, they might only consider wiki-related steps, but not looking up a reference in another page or grabbing a book from a bookshelf at home. If I am understanding the goal of the question, here are a couple for how to break this up better. One would be to have different open-ended questions based on different interfaces where behavior might happen, e.g. online and onwiki, online but offwiki (e.g. google search), offline and off-wiki. Another would be to bucket all the known activities into multiple choice, multiple response; but not sure if that is feasible, or the information you are looking for.

Describe the type of activity you were doing
This question is a bit unspecific and it feels very close to the previous question. What is the goal for this question? This will help me direct feedback. Also, do you users just do one activity when they click the "save page" button?

How much time did you spend working on this edit
Again, specify what you mean by "edit". I would make sure its related only to when the user clicked the save page button since they might jumble more than one; but depends on what you are going for here. I would make the responses more specific:

Less than 1 minute
More than 1 minute, less than 1 hour
More than 1 hour

Also; how can I make surveys like this?? =)

Hope this helps!

what is the plus/minus bar?

@egalvezwmf,

Thanks for the feedback!

I understand your point regarding the steps taken to arrive at an edit. I tried to clarify this question through the help button on the right. As you were going through the survey did you recognize or interact with this button? If not, I may need to consider finding a different way to present it. In this question I am trying to get all activities that the user feels were part of their contribution. What would think of a phrasing such as,

Describe all activities (online, offline, on-wiki, and off-wiki) that contributed to this edit. These can range from editing the wiki-text to getting a book from the library.

Regarding the type of activity, I was hoping to get responses such as copy editing, creating new content, fact checking, wikifying, cleanup, adding pictures, coordinating, etc. Do you think using a pull down menu with an allowance for free response would be more effective at communicating its purpose?

For the time question, I think that the design of the form may have been confusing. I am asking for a specific amount of time with the plus/minus bar and then am asking for the units of the number the user entered with the seconds/minutes/hours button set. Was this how you interpreted it or did it seem that I was just asking for the amount of time to be put in 3 general categories? If it is being interpreted as such I will need to look at tweaking it to make it's purpose clear.

So far as I'm aware there currently isn't a really easy way to make surveys like these, I ended up building it from some components in oo-js. If you think that it is a compelling format for a survey, then it might be worth looking into building a more robust framework to allow these to be produced with less effort. I rather like the format, but I'm biased.

Once again, thanks for the feedback!

Ah ok - I was not able to actively test the survey (not quite sure how to do that).

For the "describe all activities?" question you want to know if a user spend time researching/reading before they even logged into wikipedia right? If so, I would include a full example in order to encourage respondents to add the full range of tasks involved:

//

• Went to the library and checked out book
• Read book
• Added content
• Looked up resource link
• Added reference
• Searched for photos
• Added photo
• Clicked save page button

//

About the type of activity, yes; that would be a much better way to capture that information; its a bit vague at the moment. You can also use something like "Overall, what would you name the type of activity (e.g. copyedit, add link, formatting, etc.)" and let users specify on their own; you can then qualitatively code afterward. I worry about limiting people to a few responses; might be better to let them categorize their own work. The one caveat here is - has this already been done? (@Halfak?).. if so, then you should go with whatever that research found.

About the minutes, hours bar, I was only able to see the screenshot, so I can't see how it worked. I'm wondering if you might change it so you have a text field restricted to numeric characters, and then the seconds, minutes, hours buttons. That way they can just tell you how long it took rather than fussing with the slider - just a thought tho. not sure how it works.

@egalvezwmf, Thanks for the notes. Regarding your question "has [letting users categorize their work] already been done?" My recent lit review suggests: no. https://meta.wikimedia.org/wiki/Research:Automated_classification_of_edit_types

If we're going to impose some categorization scheme on editors, it probably makes sense to inform it with a grounded theory approach rather than just boldy making assumptions.

If you'd like to see the interface live in the Wiki, I've provided instructions in the description of the task:

add the following line to your common.js file.

importScript("User:Another_Article/protostudy.js");

See mine as an example: https://en.wikipedia.org/wiki/User:Halfak_%28WMF%29/common.js

@Halfak, I think I already did that, but I dont know how to actually view the survey :/
https://en.wikipedia.org/wiki/User:EGalvez_%28WMF%29/common.js

And yes, definitely agree with Grounded theory approach vs making assumptions.

Is the order of tasks that editors do to complete an edit important in the research? I have an idea about how to split the question, but we may lose chronology.