Advertising Research - Persuasion
Advertising Research - Persuasion
Forced Exposure Brand Preference Change
Theater testing, pioneered by Horace Schwerin and Paul Lazarsfeld in the 1950s and briefly described in Chapter 10, is now done by McCollum/ Spielman, ASI, and ARS.6
The McCollum/Spielman test uses a 450 person sample spread over four geographically dispersed locations.7 In forced exposure tests, the respondents are recruited by telephone to come to a central location in order to preview television programming. Seated in groups of 25 in front of television monitors, they respond to a set of demographic and brand/product usage questions that appear on the screen. The respondents view a half-hour variety program featuring four professional performers. At the midpoint in the program seven commercials including four test commercials are shown.
Performer Performer T C T C T C T Performer Performer
A B 1 2 3 4 C D
C = constant commercials T = test commercials
After audience reactions to the program are obtained, an unaided brand name recall question is asked that forms the basis of the clutter/ awareness score (the percentage who recalled that the brand was advertised). The clutter/awareness score (C/A) for 30-second commercials averages 56 percent for established brands, 40 percent for new brands.8 The four test commercials are then exposed a second time surrounded by program material:
Program T T T T
Intro. 1 Program 2 Program 3 Program 4 Program
T = test commercial
An attitude shift (AS) measure is obtained. For frequently purchased package goods, such as toiletries, the preexposure designation of brand purchased most often is compared with the postexposure brand selection in a market basket award situation. The respondents are asked to select brands they would like included if they were winners of a $25 basket of products. In product fields with multiple brand usage, such as soft drinks, a constant sum measure (i. e., ten points to be allocated to brands proportioned to how they are preferred) is employed before and after exposure. For durables and services, the pre- and postpreference is measured by determining:
The favorite brand
The next preferred alternative
Those brands which would not be considered
Those brands which are neither preferred nor rejected
An important element of the McCollum/Spielman test is the use of two exposures. McCollum/Spielman and many advertisers argue that less than two exposures represents an artificial and invalid test of most advertising.
Finally, diagnostic questions are asked. Some of the areas that are frequently explored include:
• Comprehension of message/slogan
• Communication of secondary copy ideas
• Evaluation of demonstrations, spokesman, message
• Perception of brand uniqueness/brand differentiation
• Irritating/confusing elements
• Viewer involvement
In a rare copy-test validity check, McCollum/Spielman asked advertisers of 412 campaigns (some campaigns consisted of several commercials) that were tested over a three-year period whether the brand has exceeded marketing objectives during the time that the campaign was being aired.9 These advertising campaigns were then divided into four groups.
High AS (attitude shift) and High A/C (awareness/communication) High AS and Low A/C Low AS and High A/C Low AS and Low A/C
9Peter R. Klein and Melvin Tainter, "Copy Research Validation: The Advertiser's Perspective," Journal of Advertising Research, 23 (October/November 1983), 9-18.
The results are shown in Figure 23-2. Clearly, the AS persuasion measure was a good predictor of campaign success. The A/C recall measure, on the other hand, may have diagnostic value, but it had little relationship to campaign success.
The ARS approach is similar, except that their proven recall measure is the percentage of respondents that claim to have seen the advertisement and can give some playback of it 72 hours later.10 ARS obtained a correlation of .78 with their proven recall measure and the unaided brand awareness level achieved by 24 new brands in test markets. Their pre- and postpersuasion measure had a correlation of .85 with the trial rate of 26 new brands in test markets. Further, on four occasions the ARS persuasion score correctly predicted which of two commercials achieved higher split-cable test market sales.
10"Advertising Quality Deserves More Weight!" Research Systems Corporation (August 1983). "ASI Laboratory Methodology, ASI Market Research, Inc., New York, Undated.
ASI, which uses a central Los Angeles location, relies on a pre- and postmeasure of brand selection in a prize/drawing context. Reliability studies across 100 commercials in 15 product categories yielded test—retest reliability correlations from .81 to .88. Fifteen hundred commercials per year are tested by ASI, so well developed and current norms are available.
The Buy Test design of the Sherman Group does not involve a central location, because the respondents are often recruited and exposed to advertising in shopping malls.12 A series of unaided questions on advertisement and copy recall identify those in the "Recall/Understand" group. The advertising "Involvement" group are those who had a favorable emotional response, believed that the brand positioning fit the execution, and felt that the advertisement was worth looking at (or reading). The "Buying Urgency" group is identified in part by intentions to buy, improved product opinion, and the motivation to tell someone something about it. A basic measure, the BUY score, is the percentage of those exposed who become part of all three groups.
On-Air Tests—Brand Preference Change
In a Mapes and Ross on-air test, commercials are aired in a preselected prime-time position on a UHF station in each of three major markets. Prior to the test, a sample of 200 viewers (150 if it is an all male target audience) are contacted by phone and asked to participate in a survey and cash award drawing that requires viewing the test program. During the telephone interview, respondents provide unaided brand name awareness and are questioned about their brand preferences for a number of different product categories. The day following the commercial exposure, the respondents again answer brand preference as well as DAR questions. The key Mapes and Ross measure is pre- and postbrand preference change.
A Mapes and Ross study involved 142 commercials from 55 product categories and 2241 respondents who were recontacted two weeks after participating in a test. Among those who bought the product category, purchases of the test brand were 3.3 times higher among those who changed their preference than among those who did not change.13
The ASI Apex system differs from the Mapes and Ross approach in several important ways.14 First, before exposure, brand preference is measured by determining the brand bought most often and the brand most likely to buy next. Thus, people who both use the brand and plan to buy it next are distinguished from those who answer only one or none of those questions positively. The impact of sample composition with respect to brand usage is thus potentially controlled (although the test samples of size 200 are limiting in this respect). Second, the after exposure brand preference measure, based on the brand to be selected if the respondent won a drawing and the brands they would consider if their preferred brand was unavailable, differs from the before measure. The use of different before/after measures reduces the likelihood that the before measure will influence the after measure. Third, the results are compared to a control group of 600 who goes through the complete procedure but does not see the
13Descriptive material from Mapes & Ross.
14APEX, ASI Market Research, Inc., New York, March 1984, 46.
test advertisements. Thus, the impact of the procedure itself on brand preference for a given product class can be determined.
Customized Measures of Communication/Attitude
Standardized copy test measures are useful because they come with norms sometimes based on thousands of past tests. Thus, the interpretation of the test becomes more meaningful. Some objectives, particularly communication objectives, are necessarily unique to a brand, and may require questions tailored to that brand.
For example, Chevron ran a series of 12 print ads in 1980 mostly telling people that Chevron made a lot less profit than people thought.15 A posttest sample of 380 respondents in seven major markets were interviewed. The percentage of those who were aware of at least one of the print ads that agreed with the statement "Chevron makes too much profit" fell from 81 to 72 percent.
Interestingly, however, data from the same study showed that people seeing these print ads and the very positive "Energy Frontier" television campaign actually had less attitude change toward Chevron than those seeing only the television ads. Thus, the print ads (20 percent of the budget) actually reduced the effect on the attitude toward the firm. Creating a positive attitude obviously had a positive impact on all belief dimensions. Calling attention to a source of irritation, namely, oil company profits, tended to counteract the positive attitude change. The Chevron experience graphically illustrates the risk of measuring a part of a campaign in isolation.
Forced Exposure Brand Preference Change
Theater testing, pioneered by Horace Schwerin and Paul Lazarsfeld in the 1950s and briefly described in Chapter 10, is now done by McCollum/ Spielman, ASI, and ARS.6
The McCollum/Spielman test uses a 450 person sample spread over four geographically dispersed locations.7 In forced exposure tests, the respondents are recruited by telephone to come to a central location in order to preview television programming. Seated in groups of 25 in front of television monitors, they respond to a set of demographic and brand/product usage questions that appear on the screen. The respondents view a half-hour variety program featuring four professional performers. At the midpoint in the program seven commercials including four test commercials are shown.
Performer Performer T C T C T C T Performer Performer
A B 1 2 3 4 C D
C = constant commercials T = test commercials
After audience reactions to the program are obtained, an unaided brand name recall question is asked that forms the basis of the clutter/ awareness score (the percentage who recalled that the brand was advertised). The clutter/awareness score (C/A) for 30-second commercials averages 56 percent for established brands, 40 percent for new brands.8 The four test commercials are then exposed a second time surrounded by program material:
Program T T T T
Intro. 1 Program 2 Program 3 Program 4 Program
T = test commercial
An attitude shift (AS) measure is obtained. For frequently purchased package goods, such as toiletries, the preexposure designation of brand purchased most often is compared with the postexposure brand selection in a market basket award situation. The respondents are asked to select brands they would like included if they were winners of a $25 basket of products. In product fields with multiple brand usage, such as soft drinks, a constant sum measure (i. e., ten points to be allocated to brands proportioned to how they are preferred) is employed before and after exposure. For durables and services, the pre- and postpreference is measured by determining:
The favorite brand
The next preferred alternative
Those brands which would not be considered
Those brands which are neither preferred nor rejected
An important element of the McCollum/Spielman test is the use of two exposures. McCollum/Spielman and many advertisers argue that less than two exposures represents an artificial and invalid test of most advertising.
Finally, diagnostic questions are asked. Some of the areas that are frequently explored include:
• Comprehension of message/slogan
• Communication of secondary copy ideas
• Evaluation of demonstrations, spokesman, message
• Perception of brand uniqueness/brand differentiation
• Irritating/confusing elements
• Viewer involvement
In a rare copy-test validity check, McCollum/Spielman asked advertisers of 412 campaigns (some campaigns consisted of several commercials) that were tested over a three-year period whether the brand has exceeded marketing objectives during the time that the campaign was being aired.9 These advertising campaigns were then divided into four groups.
High AS (attitude shift) and High A/C (awareness/communication) High AS and Low A/C Low AS and High A/C Low AS and Low A/C
9Peter R. Klein and Melvin Tainter, "Copy Research Validation: The Advertiser's Perspective," Journal of Advertising Research, 23 (October/November 1983), 9-18.
The results are shown in Figure 23-2. Clearly, the AS persuasion measure was a good predictor of campaign success. The A/C recall measure, on the other hand, may have diagnostic value, but it had little relationship to campaign success.
The ARS approach is similar, except that their proven recall measure is the percentage of respondents that claim to have seen the advertisement and can give some playback of it 72 hours later.10 ARS obtained a correlation of .78 with their proven recall measure and the unaided brand awareness level achieved by 24 new brands in test markets. Their pre- and postpersuasion measure had a correlation of .85 with the trial rate of 26 new brands in test markets. Further, on four occasions the ARS persuasion score correctly predicted which of two commercials achieved higher split-cable test market sales.
10"Advertising Quality Deserves More Weight!" Research Systems Corporation (August 1983). "ASI Laboratory Methodology, ASI Market Research, Inc., New York, Undated.
ASI, which uses a central Los Angeles location, relies on a pre- and postmeasure of brand selection in a prize/drawing context. Reliability studies across 100 commercials in 15 product categories yielded test—retest reliability correlations from .81 to .88. Fifteen hundred commercials per year are tested by ASI, so well developed and current norms are available.
The Buy Test design of the Sherman Group does not involve a central location, because the respondents are often recruited and exposed to advertising in shopping malls.12 A series of unaided questions on advertisement and copy recall identify those in the "Recall/Understand" group. The advertising "Involvement" group are those who had a favorable emotional response, believed that the brand positioning fit the execution, and felt that the advertisement was worth looking at (or reading). The "Buying Urgency" group is identified in part by intentions to buy, improved product opinion, and the motivation to tell someone something about it. A basic measure, the BUY score, is the percentage of those exposed who become part of all three groups.
On-Air Tests—Brand Preference Change
In a Mapes and Ross on-air test, commercials are aired in a preselected prime-time position on a UHF station in each of three major markets. Prior to the test, a sample of 200 viewers (150 if it is an all male target audience) are contacted by phone and asked to participate in a survey and cash award drawing that requires viewing the test program. During the telephone interview, respondents provide unaided brand name awareness and are questioned about their brand preferences for a number of different product categories. The day following the commercial exposure, the respondents again answer brand preference as well as DAR questions. The key Mapes and Ross measure is pre- and postbrand preference change.
A Mapes and Ross study involved 142 commercials from 55 product categories and 2241 respondents who were recontacted two weeks after participating in a test. Among those who bought the product category, purchases of the test brand were 3.3 times higher among those who changed their preference than among those who did not change.13
The ASI Apex system differs from the Mapes and Ross approach in several important ways.14 First, before exposure, brand preference is measured by determining the brand bought most often and the brand most likely to buy next. Thus, people who both use the brand and plan to buy it next are distinguished from those who answer only one or none of those questions positively. The impact of sample composition with respect to brand usage is thus potentially controlled (although the test samples of size 200 are limiting in this respect). Second, the after exposure brand preference measure, based on the brand to be selected if the respondent won a drawing and the brands they would consider if their preferred brand was unavailable, differs from the before measure. The use of different before/after measures reduces the likelihood that the before measure will influence the after measure. Third, the results are compared to a control group of 600 who goes through the complete procedure but does not see the
13Descriptive material from Mapes & Ross.
14APEX, ASI Market Research, Inc., New York, March 1984, 46.
test advertisements. Thus, the impact of the procedure itself on brand preference for a given product class can be determined.
Customized Measures of Communication/Attitude
Standardized copy test measures are useful because they come with norms sometimes based on thousands of past tests. Thus, the interpretation of the test becomes more meaningful. Some objectives, particularly communication objectives, are necessarily unique to a brand, and may require questions tailored to that brand.
For example, Chevron ran a series of 12 print ads in 1980 mostly telling people that Chevron made a lot less profit than people thought.15 A posttest sample of 380 respondents in seven major markets were interviewed. The percentage of those who were aware of at least one of the print ads that agreed with the statement "Chevron makes too much profit" fell from 81 to 72 percent.
Interestingly, however, data from the same study showed that people seeing these print ads and the very positive "Energy Frontier" television campaign actually had less attitude change toward Chevron than those seeing only the television ads. Thus, the print ads (20 percent of the budget) actually reduced the effect on the attitude toward the firm. Creating a positive attitude obviously had a positive impact on all belief dimensions. Calling attention to a source of irritation, namely, oil company profits, tended to counteract the positive attitude change. The Chevron experience graphically illustrates the risk of measuring a part of a campaign in isolation.
Comments
Post a Comment