The Starch Tests are a classic series of tests pioneered during the 1920s by Daniel Starch (1883-1979), a psychologist who specialized in advertising research. The tests measure audience recall of advertisements in newspapers and magazines.
The tests were the first examples of what Mr. Starch named "recognition research," a method that is now widely accepted and used.
He founded Daniel Starch and Associates, which conducted the tests for decades. The firm is now part of United Business Media plc.
(Methods may vary from researcher to researcher.) The researcher interviews readers of print publications and asks each interviewee if he has recently read certain publications.
If an interviewee has recently read a publication, the researcher asks the interviewee which issue he read, and which ads he noticed in that issue (this is "unaided recall").
Then the researcher produces the issue and asks the interviewee to look inside it.
After the interviewee has looked, the researcher asks him about a certain advertisement in that issue (this is "aided recall").
The researcher keeps track of the percentage of subjects who:
This method enables the researcher to assess the effectiveness of various elements of ads, such as size (e.g., full-page vs. half or quarter), with fairly good accuracy.
Many advertisers value the classical form of the test because it can compare the effectiveness of their ads with the effectiveness of other advertisers' ads in the same issue of the same publication – a useful sort of benchmarking.
The Starch Tests are navigational tools – they can help you steer your ads toward greater profitability, by making your ads more effective at capturing and holding the reader's attention and sticking in his memory. However, the tests can't help you measure how profitable they are. To measure profitability, you need evaluative tools.
Experts (researchers, psychologists and advertisers) disagree about the research value of aided recall. This is a long-standing controversy.
Also, the testing may be affected by Nonresponse Bias. However, experts can often minimize the effects of this bias.