Hi. I am on my 7th day of MTurk. I downloaded several scripts to Chrome but I don't understand these: 1. I am using scripts to block certain hits and requesters. These work fine but they don't change what is listed on each page. In other words, now, only 1 or 2 hits might show on a page instead of 10. Is there any way to get 10 showing per page again so I don't have to go through so many pages all the time? 2. I don't understand how to export a CSV file from the Hit Database. If I check the box, then what? Just log out or update the database? There is nothing saying that it saved a file. And if a file was saved, where is it on my hard drive? Thanks very much for any assistance, Liz PS: Chrome just updated itself this morning, so I guess I have the latest version.
I'm not exactly sure what you're asking in the first question. If you block a requester, the script will simply collapse the hit so you won't see the details. It's possible to edit the script to eliminate it, but then you will be hiding hits that you may want to unblock later, so I think for that reason the programmer did it that way. If you wanted to completely block hits and still have 10 show, that's not possible. Scripts just edit html/css elements on a given page. We have no control over the search function of amazon considering it's on their server, something we don't have access to. We can only edit client-side functionality, which is the ten hits that are currently on the page. Or in other words, the server spits out 10 hits to you based on your search query, the script merely looks at the client's page elements that you're currently viewing. If you know how to program you could scrape their whole hit database and then design your own search feature, but considering amazon blocks multiple requests (you can hit a max request limit) that wouldn't really be possible on an efficient level. I could be mistaken though. On the second question, yeah, the hit database script is very buggy. I think there was one time a month ago where the export feature worked for me, but has since broke on me. It's not very maintained. I actually had to edit the script myself to even view my hits and use the reporting features on firefox. I wish there was more activity on an open source mturk project, but oh well. TLDR: So basically, first question, no it's not possible, second question, it's buggy, it doesn't work.
Thank you for responding. I found out by testing that I could get a CSV file to export on Chrome if I filled in the search boxes above the check mark. However, I don't see the file as very useful (in Notepad), as it's a mess to read it. I don't have Excel or anything to import it into to make it more useful. I understand better now why the hit blocking doesn't change the page viewed. What is really annoying to me is that every time I complete or even look at a hit, it returns to the first page of the hit list and I have to go back through a bunch of pages again. I guess this is an Amazon limitation?
There are a couple things you can do to view as a spreadsheet. There is a free suite that has been around for years called Open Office. It will do just about anything Excel will do. Google Docs also has the ability to upload and work with spreadsheets in a similar format and there are online excel viewers (skydrive has one built in as well). I had the same issues with the page resetting to 1 every time I viewed a HIT. Just had to get into the habit of always opening/viewing in a new tab. I will keep several tabs open with requestor specific searches and do the same thing just so I don't have to type it all in the search box again. After awhile it will just become second nature. Depending on your browser you can always change the settings to "always open in a new tab".
