In reading some of the other participants' blogs, I realized I'd gone WAY astray on question 5 about how many women were in the workforce in Anchorage. When I'd done my search and got 150K or so, I basically thought, yeah, that's half the population of Anchorage, and didn't really think it through. Like maybe a goodly portion of the female population in Anchorage might be underage and not technically part of the workforce. So Saturday at the ADR-Anchorage meetup, I brought the question to the group. And showed them my search which wound up basically the same way. This is on
ALARI: the Alaska Dept of Labor & Workforce Development's Research and Analysis website.
To show you what went wrong for me, I grabbed a series of screenshots.
So I went to the site, clicked the first dropdown, spotted Anchorage/MatSu but decided that wasn't what I wanted, and made the assumption that I should click the second drop down.
There was Anchorage! So I clicked on it.
And then I dropped to the bottom of the page, cleared out all the selected datasets, and just chose Worker characteristics. And hit Next.
And this is what I got. Note the circled items at the top: I did not notice that apparently my earlier chosen delimiters had not taken. This is the search I recreated for my colleagues on Saturday. And as we were viewing the initial webpage, I finally clicked on the Go button next to the Anchorage choice and got the following:
And behold, the Alaska heading changes to Anchorage, Municipality of. And I redid the rest of the search, clicked next, and got the correct answers. Or at least the answers everybody else got.
The conversation went for a while and kept getting more and more raucous as we tried to compare notes and show each other what we were doing on 4 different wireless devices scattered amongst the six of us. At least 2 of the others who got the answer right didn't recall having clicked on the Go button to save the delimiter. And one of them swore she'd never even noticed all three of the Go buttons (so you know she's not lying when she said she didn't recall clicking on it.) And as we played on the site, we also realized we didn't have to move L-R on the geographic choices: if you chose the appropriate locale in the 2nd or 3rd box, it automatically carried through to the larger geographic areas in the boxes to its left.
We all decided that DLWD needs to do some user testing on this site. Or at the very least, read this and put some directional info in like: "To narrow your search geographically, choose a location and click the Go ."
Arlene, thanks for posting this well illustrated example of how to arrive at the correct number of female workers. I have linked excercise 5 to your blog post.
ReplyDeleteSee, I just can't get past the idea that it shouldn't take this much explanation to function for people. Don't make me learn how to do it, fix the search mechanism so it doesn't have this difficulty! Obviously I don't expect you to do that, Daniel. :)
ReplyDeleteI just wonder, when really basic user testing is so, so easy and quick, why so few people do it. Recently I had great fun reading "It's not rocket surgery" by Steve Krug--http://www.sensible.com/--and honestly it couldn't be easier to figure out how people use your site and how they want to use your site. Fixing the interface might be a bit more work, but when 40% of your audience gets it wrong (based on our comparisons from the group on Saturday), I think maybe the development time needs to be spent.