I’m not quite sure why but this blog has developed a bit of a research methods vibe over the past couple of weeks (I say “vibe” because it sounds so much less sinister than “obsession”) and this post continues the theme with a site I chanced across while searching for some sampling-related pictures (I know, it’s just one long fun parade working at shortcutstv.com).

It’s a bit of an oddity because although it’s a maths / statistics site there are a couple of areas on surveys and sampling that should be useful for both sociology and psychology teachers / students. These are illustrated by a mix of text, simple graphics and a couple of bits of optional video.

When all’s-said-and-done, however, it’s basically a pen-and-paper site, so if you don’t want your students sloping off to explore things like “binominals” (which sounded vaguely interesting but turned out not to be) you could easily copy the bits you wanted your students to cover, although it’s probably not worth the effort.

Since it’s a site aimed at American college students the examples it uses are rooted in American culture (school mascots!) and history but there’s nothing here that’s too alien to a British audience (and you can always substitute your own culturally-specific examples if necessary).

The two areas most-useful for sociology / psychology students are probably:

1. Sample Surveys

 This is based around the idea of social surveys and how you might select representative samples, with three broad techniques outlined:

Census – 100% sample

Simple random sampling

Stratified random sampling

Aside from brief text outlining different techniques the emphasis is on applying knowledge to particular situations – these are generally straightforward (although there’s a bit of maths included that sociologists can probably ignore and psychologists should find reasonably okay).

2. How Surveys Can Go Wrong

The focus here is on a range of sampling and survey errors: 

a. Sampling errors that make the research unrepresentative includes an example that sociologists can happily ignore (it involves a range of mathematical calculations that aren’t tested in the exam) and psychologists can happily embrace (Because, psychology!).

b. Survey Response bias gives examples of the different ways a survey may produce unreliable, invalid, data even where a sample is, in principle, representative. This section includes questions designed to get students applying what they’ve learned about response bias to some simple situations.


Share your thoughts on this post: