The term “copy test” refers to a group of different test methods designed to measure the usage of adverts or editorials in the print media. During the course of an interview, readers who state that they have read the test issue of a magazine or newspaper are shown an original copy of the issue in question. The interviewer then goes through this copy with the respondent page by page, asking questions about all, or certain, selected adverts or editorials to establish whether each one was “read in full,” “only glanced at” or “paid no attention at all”.
Although such tests are conducted as close to the actual reading event as can be arranged, normally within a time span of just one or two days, it is still impossible to completely avoid memory distortions. Tests that employ technical equipment, for example eye cameras, to monitor respondents’ field of vision while reading have shown that readers forget about some of the pieces that they in fact read (“underclaiming”). Other pieces are cited by readers even though they actually had no eye contact with them (“overclaiming”). In this respect, copy test findings often reflect respondents’ interests and habits. For example, readers who usually always read the lead article but are not sure in the case of the test article, are likely to say that they read it even though they cannot clearly remember actually having read that particular one. Although copy tests use a clear investigative approach and solid memory aids, giving the outward impression that findings must correct, it is evident that there are still limits on the validity of the information they can provide.
One of the most commonly used copy tests is the so-called Starch Test, which was developed by American media researcher Daniel Starch in 1931 and is still in use in America and around the world today as a standardized procedure for investigating ad noting and advertising effects. In line with copy tests in general, the Starch Test is a recognition test based on the ability of readers to remember advertisements presented to them in their original context in actual issues of publications (newspapers, magazines). Respondents who identify themselves as readers of a particular magazine or newspaper are posed the following three questions to establish the Starch categories:
1 Noted score: Question: “Have you seen or read any part of this item?” (percentage share of readers who noted this advertisement).
2 Associated score: Question: “Have you seen or read sufficient of this advertisement to know what it was advertising?” (percentage share of readers who say that they have noted this ad (ad noters) and are able to recall the product advertised, the advertiser, or the brand).
3 Read most score: Question: “Have you read more than half of the words, or less than half of the words?” (percentage share of readers who say they have read at least half ). In addition, it is possible to expand on these three standard measures by ascertaining two further parameters:
4 Involvement score: Question: “How involved do you feel with the ad?” (responses “involved” or “very involved” contribute to the score).
5 Persuasion score: Question: “How likely are you to try the brand?” (responses “likely” and “very likely” contribute to the score).
Responses are accepted on face value without any form of verification. Doubts thus exist as to whether all readers are equally able to provide sufficiently reliable responses considering the substantial demands the test places on their memories. During the course of a typical advertising campaign, an ad is often run either in one advertising medium consecutively or in several different media simultaneously (e.g., several different magazines), meaning that it is often not possible for respondents to recall with certainty whether they saw the advert in question in the test issue or actually in an earlier issue or in a different advertising medium altogether (multiple exposure, overlap). For this reason, the Starch Test has its limitations when it comes to drawing conclusions about the performance of one particular advertising medium.
In contrast, the Starch Test can clearly say a great deal about the quality of an advertisement itself. For example, investigations by Wells came to the conclusion that a Starch “noted score” in fact measures “whether, in the consumer’s judgment, the ad is worth at least a passing glance” (Wells 1964) and, according to Krugman (1985), what the Starch scores primarily convey is an advertisement’s “attention-getting quality.” Despite Starch scores being widely used by advertisers, for example also to optimize advertisement copy and form, there remains skepticism concerning the meaning of Starch data. Based on the findings of their own validity tests, Zinkhan and Gelb (1986) come to the conclusion that “however justified the qualms are about Starch scores as valid recognition measures, they are not useless numbers. They may only point out “ads the respondents believe they would have noticed”; but even if so, they predict, at least modestly, responses in which advertisers have a far greater stake.
- Krugman, H. E. (1985). Point of view: Measuring memory – an industry dilemma. Journal of Advertising Research, 25, 49 –51.
- Ostlund, L. (1978). Advertising copy testing: A review of current practices, problems and prospects. In J. Leigh & C. Martin (eds.), Current issues and research in advertising. Ann Arbor: University of Michigan, pp. 87–105.
- Starch, D. (1931). Buying power of the American market: National and local markets measured by family incomes in all states – a comprehensive measure of America’s buying power. New York: Daniel Starch.
- Starch, D. (1966). Measuring advertising readership and results. New York: McGraw-Hill.
- Wells, W. D. (1964). EQ, son of EQ and the reaction profile. Journal of Marketing, 28(4), 45 – 52.
- Zinkhan, G., & Gelb, B. (1986). What Starch scores predict. Journal of Advertising Research, 26(4), 45 –50.