Using GSS

Posted: 9th June 2015

This is a sister post to one on estimating the number of LGBTQ workers in technology. If you want to read about the value I got and how go read that. If you want to read about using GSS to get that data read this. Or read both.... Or neither, I guess.

Step 1: the obligatory sign up

Simple really go here and fill in the form.

Step 2: select the variables

From your projects page click 'search variables' on the right (we'll use your default project [1]). You'll now be presented with a huge list of the thousands of various 'variables' you can work with (the questions from the survey) and a search box. Search for what you're interested in and select those variables by clicking the '+' icon to the left of each entry you care about [2].

A tip on using the search system: pick a year you care about. The survey has changed regularly so you want to be precise (e.g. there are at least 3 'indus' codes as the 'occupation' question has changed twice since 1972). Read the description and the survey question (on the other 'tab', selected just below the search bar). Clicking on the variable will bring up more information as well as what values it can have (e.g. the different industries someone can say they work in).

Once you've chosen all the variables you care about you can go to...

Step 3: Extraction

In the top left corner, just below your user name, there should be an orange 'Actions' button. Clicking this will make a drop down appear with the option 'Extract Data', click that and you'll be whisked to a extraction form.

The first block of the extract form should be two boxes, one for a name and the other for a description. Name and describe your data, click 'next' (a new block will open).

Now choose the variables you want to export. These will be the columns of a spreadsheet (and I assume something analogous in the other formats selected at the end but I have no idea). The variables can either be dragged & dropped or you can click 'add all'. DONT click on the variable name, it'll take you to the variable info page; there is also checkbox which seems to do the same thing. Once you have the variables you want to extract, as ever, click 'next'.

Now you get the chance to filter your data based on your variables. For example if you only care about 1984 add the 'year' variable and (once the filter thing has loaded) select the years 1984 through to 1984. You can then add more variables by clicking below this box and each one will present the options for filtering (normally a huge list of check-boxes). One last time: click 'next'.

Select the format you want the data in (e.g. 'Excel Workbook (data + metadata)'. Click 'Create Extract', wait for the spinning to stop on the new page and the click the download icon.

Congratulations, you too can now (hopefully) replicate how I got my data.

[1] If you don't want to use your default project then this tutorial is no-where near advanced enough for you, sorry

[2] For some reason the whole site seems to use a shopping basket metaphor. As far as I know they don't expect pay so just ignore it but it almost stopped me bothering... [back]