An Interesting Method of Data Collection

Recently there was a commercial on the radio in San Jose looking for families that wanted to participate in an interesting type of data collection method. The advertisement – while potentially a scam, but interesting nonetheless – told San Jose families that if they qualify for the survey they will receive $1000. With that $1000 they can buy any and all groceries they want.

That was the entirety of the data collection, which the advertisement referred to as a "Survey," although it's questionable if that truly qualifies as a survey. The goal of this data collection methodology was to determine what San Jose families of various demographics spend their grocery money on. By handing them $1000 with the caveat that they must be spending it at various grocery stores, the company or government body (unclear by the advertisement) gets an opportunity to learn what people spend their money on without requiring they remember their own shopping habits.

Overall, collecting data in this manner is interesting, and potentially very effective. Normally a company would have to view receipts or hope that the person has a great memory, but by handing them $1000 to spend on groceries, they can learn any and all of the groceries that these individuals buy at each and every location which they shop, giving them a much more complete picture and a greater amount of data. It likely has some downsides, however.

For example: Price – This is a fairly expensive method of collecting data. $1000 per individual either limits the number of individuals you can use as a sample or causes you to spend a great deal of money. Reliability – One has to wonder how often someone would either buy more or buy different products if they were spending someone else's money, no matter how much the system promotes honesty. There are often times that people go for a cheaper or smaller or less varied amount of groceries because of cost.

Sample Size – Everyone eats different foods and cooks different meals. Some people base their meals on recipes while others have their routines. Can this survey or others get a large enough sample to be meaningful? Demographics – This was an opt-in survey.

As usual with Opt-in surveys, there is a lot of potential to have an unrepresentative sample. Still, it's certainly an interesting data collection method, and if it's both real and large enough it could be fairly successful at generating useful data.