Advertising Research - Copy Test Validity
Advertising Research - Copy Test Validity
The copy test validity refers to its ability to predict advertising response. Figure 23-4 provides an overview of some of the important ways in which copy tests can differ. Each dimension involves validity issues and trade-offs with cost. There follows a discussion of several important considerations involved in evaluating validity.
Advertisement Used
One issue is whether a rough mock-up of an ad or a finished ad is used. Several copy test firms have reported high correlations with mock-up measures and finished copy measures. The seriousness of the problem will depend on the difference between the mock-up and the finished commercial and the impact of this difference on audience response. For example, it is very difficult to test humor in rough form.
Another issue is the frequency of response. To what extent can a copy test predict the response to a campaign that will involve dozens or even hundreds of exposures? Can a single exposure provide meaningful results or should a minimum of two or three be used?
Still another issue is the context in which the test advertisement is embedded. The use of a clutter of advertisements embedded in a program or magazine is the most realistic but adds complexity and is possibly confounding.
Exposure
The approaches such as the theater tests or mall intercept exposure contexts are termed forced exposure tests because the setting is artificial and the respondent is required to watch. Respondents thus tend to pay more attention to the advertisement than under normal conditions where they are much more likely to ignore it completely. The others, such as the BehaviorScan split-cable testing are termed "on-air" tests because the exposure is a natural home setting in the context of watching a show. Approaches such as the ASI Apex method are on-air but the respondents realize they are in a test and are not watching a show they would watch at a time they would normally watch it. Thus, there is still concern that the exposure context may affect the results.
Sample
The sample should be representative of the target population. In all copy test approaches the biggest concern is with the bias introduced by nonresponse. The danger is that those who refuse to participate may respond differently than those who do agree to respond. In addition, mall intercept methods obviously access only mall shoppers, and cable-based tests miss those not connected to a cable. There is also the question whether one or even three or four cities can provide a representative sample.
Measuring Response
There is a concern with the reactive effect of the study. The respondent's reaction may be affected by knowing that actions and opinions are being recorded. There is evidence that this problem is minor in a system like BehaviorScan when the panel member becomes acclimatized to the system. However, it is of greatest concern in systems which demand that the respondent give an attitude response. Is the respondent willing and able to respond accurately?
Appropriateness of the Response Measure
Obviously, copy test validity will depend on the advertising response that is desired. A campaign that is designed to gain awareness may not be best measured by a test that focuses on immediate behavioral response. A campaign that attempts to create an image or an association with a feeling such as warmth might require many repetitions and a subtle measurement method—perhaps asking some questions directed at the use experience. A single exposure test with a coupon redemption measure may not be appropriate at all. Thus, the usefulness of the various criteria used in testing need to be evaluated in the context of the advertising objectives involved.
natural Versus Artificial
Running through the validity consideration is a spectrum from artificial to natural. At one extreme there would be a forced exposure to a commercial mock-up with a paper and pencil response using a convenience mall intercept sample. At the other would be the BehaviorScan system where the audience member realizes he or she is in a panel but everything else is completely natural including multiple exposures over time.
The copy test validity refers to its ability to predict advertising response. Figure 23-4 provides an overview of some of the important ways in which copy tests can differ. Each dimension involves validity issues and trade-offs with cost. There follows a discussion of several important considerations involved in evaluating validity.
Advertisement Used
One issue is whether a rough mock-up of an ad or a finished ad is used. Several copy test firms have reported high correlations with mock-up measures and finished copy measures. The seriousness of the problem will depend on the difference between the mock-up and the finished commercial and the impact of this difference on audience response. For example, it is very difficult to test humor in rough form.
Another issue is the frequency of response. To what extent can a copy test predict the response to a campaign that will involve dozens or even hundreds of exposures? Can a single exposure provide meaningful results or should a minimum of two or three be used?
Still another issue is the context in which the test advertisement is embedded. The use of a clutter of advertisements embedded in a program or magazine is the most realistic but adds complexity and is possibly confounding.
Exposure
The approaches such as the theater tests or mall intercept exposure contexts are termed forced exposure tests because the setting is artificial and the respondent is required to watch. Respondents thus tend to pay more attention to the advertisement than under normal conditions where they are much more likely to ignore it completely. The others, such as the BehaviorScan split-cable testing are termed "on-air" tests because the exposure is a natural home setting in the context of watching a show. Approaches such as the ASI Apex method are on-air but the respondents realize they are in a test and are not watching a show they would watch at a time they would normally watch it. Thus, there is still concern that the exposure context may affect the results.
Sample
The sample should be representative of the target population. In all copy test approaches the biggest concern is with the bias introduced by nonresponse. The danger is that those who refuse to participate may respond differently than those who do agree to respond. In addition, mall intercept methods obviously access only mall shoppers, and cable-based tests miss those not connected to a cable. There is also the question whether one or even three or four cities can provide a representative sample.
Measuring Response
There is a concern with the reactive effect of the study. The respondent's reaction may be affected by knowing that actions and opinions are being recorded. There is evidence that this problem is minor in a system like BehaviorScan when the panel member becomes acclimatized to the system. However, it is of greatest concern in systems which demand that the respondent give an attitude response. Is the respondent willing and able to respond accurately?
Appropriateness of the Response Measure
Obviously, copy test validity will depend on the advertising response that is desired. A campaign that is designed to gain awareness may not be best measured by a test that focuses on immediate behavioral response. A campaign that attempts to create an image or an association with a feeling such as warmth might require many repetitions and a subtle measurement method—perhaps asking some questions directed at the use experience. A single exposure test with a coupon redemption measure may not be appropriate at all. Thus, the usefulness of the various criteria used in testing need to be evaluated in the context of the advertising objectives involved.
natural Versus Artificial
Running through the validity consideration is a spectrum from artificial to natural. At one extreme there would be a forced exposure to a commercial mock-up with a paper and pencil response using a convenience mall intercept sample. At the other would be the BehaviorScan system where the audience member realizes he or she is in a panel but everything else is completely natural including multiple exposures over time.
Comments
Post a Comment