By: Richard Wilson (Wilson MLI Australia) and Geoff Isaac (Clemenger Australia)
With statistical analysis by: Michael Edwardson of the University of NSW
Presented by: Elizabeth Harley (Wilson MLI Australia)

| Background to the issue | Recent Research Findings | Research Approach | Results |
| Implications for syndicated electronic media research |

Foreword

The authors argue that electronic measurement alone is not enough as a basis for measuring audience delivery. If Byfield and Read (1994) and Gullen (1994) are correct and people who are paying a lot of attention to the programme they are viewing are 1.8 — 2.5 times more likely to recall advertising in that programme then there needs to be some continuous qualitative assessment of television in addition to electronic metering by which programmes can be adjusted for impact against a specified target audience.

This paper addresses whether this can be done in one dimension or whether any assessment of involvement needs to be multidimensional.

This paper explores the relationship between the two primary dimensions of viewing involvement, the Rational dimension (attention level) and the Emotional dimension programme interest). Previous work has demonstrated the value of attention levels as an indicator of advertising exposure. Similarly, positive attitudes toward specific programmes have been associated with higher levels of advertising recall. The authors look beyond recall to assess the effects of involvement on purchase intentions albeit only at a preliminary level in this paper.

Introduction

Australia is part of the worldwide trend away from paper-based diaries to 'people-meters' technology as the currency for TV audience measurement. The demand for, what promised to be, accurate and timely data was deemed to outweigh the benefits of larger scale diary based studies.

Since the changeover was implemented however, we have seen a growing demand for information that supplements basic ratings. In particular questions have focused in two major areas: do TARPS accurately reflect minute-by-minute audiences, and are all TARPS equal. For example:

While few question electronic measurement there is a growing realisation that ratings alone are insufficient for building efficient and effective media schedules. Added value measures of advertising effectiveness to date however have relied to a large degree on advertising recall. The researchers argue that recall alone is not enough and some purchase indicator should be included. In this exercise, we attempt to move beyond recall to the area of purchase intention and to ask: "Does higher programme involvement correlate with higher purchase intention for products advertised?" Henderson, Blair and Rosenberg (1994) reported that persuasion measures showed very strong predictive validity in IRI'S split cable experiment on advertising effectiveness reported in the Copy Testing issue of the Journal of Advertising Research in 1994.

In 1994 MLI began experimenting with a measure designed to assess quality of TV viewing (or viewer involvement), in addition to providing basic rating information. Respondents were asked to record their quarter hour TV viewing activities using a diary. An additional column was added to enable respondents to record how much attention was paid during each quarter hour TV viewing was recorded. Three levels of attention were offered to the respondent: a lot, some or not much.

In the course of completing a product consumption and attitude questionnaire respondents were asked to give details of their attitudes to the major TV programmes offered by each network. They are asked to indicate those programmes they especially choose to watch, those they watch because another person in the household wants to watch and those programmes they watch simply because there is nothing better to do. This measure is similar to that used by U.K. Researchers TGI.

These two potential indicators of involvement are included in the MLI database. The researchers argue that attention level is a "rational measure" of involvement whereas an attitudinal statement about interest in programmes is an "emotional indicator" involving feelings.

This experiment explores the possibility of a link between both dimensions of involvement and advertising effectiveness (as measured by purchase intention and advertising recall). It asks, is one indicator better than the other or is the combination of the two measures a more accurate predictor of involvement? Recent work in the print medium by Wilson & Isaac (Berlin '95) suggests that involvement measures such as Reader Trust are critical to determining the likely success of advertising effective in magazines.

---

 

Background to the issueBack to top

Does liking produce attention, or attention produce liking, or do both elements operate independently?

Early work conducted by the Home Testing Institute in the US (and Smith) and the Television
Audience Measurement in the UK clearly established the link between liking and attention (Television Audience Assessment, 1984). Involvement, as far as we are able to ascertain, has never been formally defined and there is no agreement as to the correct method for measuring this concept (TAA, 1984).

In Paris in 1994, Gullen of Carat Research reported that people who enjoyed programmes recalled an average of 15% of the adverts whereas "non-enjoyers recalled only 1%." Encouraged by this work we decided to combine both measures of liking and attention to assess their contribution to involvement.

In introducing the attention level measure two assumptions are made:

In support of these assumptions Phil Gullen of Carat Research in the UK reported, in his 1994 paper to the ESOMAR conference, "attention to the programme was, in fact, the biggest influence on ad recall."

The FCB Media Research Handbook under an article entitled "Environmental factors and commercial effectiveness" depicted by advertising recall. These are:

1) Ratings
The traditional indicator of effectiveness: programmes with higher ratings typically generate higher levels of recall.

However, the paper suggests, "Effectiveness is governed more by audience involvement than by rating size.... Perhaps it is not the overall rating, but instead the concentration of a particular group, that boosts commercial effectiveness."

A high rating quiz show may simply be used, by a large proportion of viewers, to provide background for other household activities.

2) Involvement or Impact
The majority of research conducted in this field (in the natural setting) indicates that programmes, which generate higher levels of involvement with their viewers, also generate higher levels of recall for their advertisers. These findings contrast with earlier laboratory based work that suggested the more impact a programme had the less effective the advertisements.

A clear definition of involvement has not been agreed upon however previous studies (such as Television Audience Assessment, 1984) have used attitudinal statements, for example: "It touched my feelings" or "I learned something" to identify programmes with high levels of involvement or impact.

These are similar expressions to those used by Olson (1983) to evaluate commercial effectiveness.

3) Program appeal
Program appeal, "is comprised of memorability, quality, credibility and persuasiveness... a measure of how much the viewer enjoys the programme".

Studies indicated the more a viewer likes a programme the more attention he or she pays to the commercial. However, viewers may like certain programmes as they merely provide a background for other activities. Involvement and liking are not always directly linked. A viewer may tune into a show every week but "involvement is needed to get the viewer to watch the programme and see the commercial".

4) Day-part
Generally prime time has the edge. In his 1994 ESOMAR paper "Who is really watching the commercials?" Phil Gullen reports on a Carat Research UK project which found ads shown in the late afternoon were recalled by far fewer viewers than at other times of the day (late evening performed best). Traditional early prime time slots were found to be less efficient as viewers were often performing other tasks (especially preparing or consuming meals), which had a negative impact on their attention levels.

5) Program type
Inconsistent results have been obtained from previous research to demonstrate the effect of programme type on effectiveness. News has generally been found to produce relatively low recall scores.

Yuspeh (1981) "found the same commercial can have radically different recall scores within different programmes, even when the programmes are of the same programme type".

The upshot of this is that it may not be possible to condense programmes into convenient classifications, something many media planners seem to want.

6) Independent stations vs Network affiliate stations
Only one study has looked at this dimension. It is not relevant for the Australian market (where all regional stations are affiliated to national networks).

7) Synergy
An advertisement for a product relevant to the content of a specific programme will have a positive impact on commercial effectiveness. In Australia we have noticed a trend for planners and buyers to give more consideration to content or environment when scheduling.

8) Holding Power
Holding power can be easily calculated from readily available data.

Limited research suggests as programme length increases, holding power decreases. The longer a show the lower the proportion of viewers who watch the entire programme. The hypothesis is that "viewers watching an entire episode are more likely to view the commercial."

However, the authors believe this element is accounted for by the rational measurement of attention included in the syndicated survey. Again the existing database has been used to explore this relationship further.

9) Clutter
Number of commercials per pod and per programme can have an impact on recall. As pod length increases average recall decreases.

10) Pod Position and Length
First and last positions are generally believed to achieve higher recall scores. As pod length increases the higher the index of non-watchers. This, however, was put into serious question by Gullen in his ESOMAR 1994 paper, "Who is really watching the commercials", reporting on an extensive British research project he challenges these older findings and suggests the first ad in the break could be the worst slot! Recall of ads in short breaks were, however, higher than longer breaks with more ads, but only when the pod length was substantially increased.

11) Participating vs Break spot
Conflicting evidence exits on the relative merits of in-programme vs break spots. Gullen's more recent paper, 1995 based on a large sample, suggests there is no difference between centre breaks and end breaks.

12) Length of commercial
Does a 15 second commercial run twice lead to a better recall than a 30 second commercial run once?
No conclusive evidence was found in preparing this review. There is a growing trend within Australia to use a short commercial, for the same product, at the beginning and end of the same pod.

13) Viewer demographics
Only age is significant according to the review; older respondents have lower recall scores.

14) Measurement techniques
Traditional recall measurements screen out respondents who left the room during a commercial break (i.e. they are not included in the percentage base). Some results are presented as percentages of all claimed viewers. Respondents who left the room, for whatever reason, are not screened out.

15) Commercial Execution
Commercials can be categorised to some extent although the creative merits of a particular commercial are more subjective. As many of our clients are agencies we specifically decided not to enter in this debate!

16) Scheduling
Spreading exposure over time vs condensed exposure over a limited time can be expected to give very different recall scores. Zenith Media Research (1993) claims that the receptivity of the audience can also govern effectiveness. In particular the physical and emotional state of the viewer can be expected to have a significant impact on their ability to recall both programme content and adverts. Apart from work completed by Axelrod in 1963 our research review failed to find many significant contributions to the debate on mood. This paper does not concern itself with mood although the authors are aware of the possible significance of this factor.

17) Relevance
The relevance of the commercial (Olson, 1983 and many others) is also an important consideration. If the viewer is in the market for the advertised product category there is likely to be better response, and especially a higher level of recall.

The authors are well aware that involvement alone is not the only element of an efficient schedule. However the authors are not alone in thinking that next to ratings, involvement (and the potential for a syndicated, regularly updated measure of involvement) has a powerful impact on media selection.

---

 

Recent Research FindingsBack to top

Byfield and Read (1994) reported on a pilot study, Quality of Viewing, conducted by The Media Partnership. This paper used the British media and market database TGI to conduct an experiment similar to that conducted by MLI. Using a representative sample of TGI respondents a large number of appreciation scores were collected for individual programmes.

The survey found, "frequent correlations between appreciation scores and audience size." But, "the data really comes into its own when different target groups are examined and when it comes to making buying decisions between programmes of similar audience size and similar cost." The second part of the study reported, "a relationship between programme involvement and, not only commercial recall, but in commercial breaks being viewed at all. "Highly involved viewers were 49% more likely to have seen, and paid some attention to the commercial break than their low scoring counterparts." In general viewers do not use the ad breaks of programmes they are highly involved with to leave the room to complete other chores. On average highly involved viewers were 18% more likely to be able to correctly identify brands advertised in the experiment (using unprompted awareness, a prompted product field question and prompted brand recall).

Attention as an indicator of involvement
The two British Foretel studies (Gullen, 1994 and others) established that of the various involvement indicators tested, attention levels had the biggest influence on recall. In the first part of the study advertising recall was measured through a self-complete questionnaire on 1,000 respondents. In the second part of the survey programme quality of viewing was established through a telephone survey of 4,000 respondents who had previously completed the TGI Single Source Survey.

Enjoyment was also found to have a significant influence on recall with the second study reporting, "People who enjoyed the programme, on average, recalled 20% more commercials within or after it." (Gullen, 1994).

Hugh Johnson, (1992) gave details of a 1990 experiment with 400 respondents who were given a video tape of one of four test programmes with a single ad break, which was common to all tapes.

Program response was tested using 26 statements and advertising recall was measured at the same time. Respondents with high Program Appreciation Scores demonstrated higher levels of recall.

Enrique Domingo de Blas, (1992), reported a Spanish experiment on advertising recall. A CATI survey of 1076 respondents concluded, 'A clear relationship between the viewers' stated activities (i.e., their degree of attention) during commercial breaks and their degree of ad recall was also revealed."

Since 1980 the Television Audience Assessment has developed a qualitative ratings system to supplement traditional TV audience measurement in the United States, using two measures of involvement: Program Impact Index (measuring intellectual and emotional stimulation) and Program Appeal Index (measuring overall entertainment value). Early findings established "that viewers involved in a programme give more attention to the programme and are more likely to stay in the room for the programme's commercials". Significantly for our research, TAA contends increase in viewers' attention resulted from increased involvement, i.e. selecting programmes for viewing should be a good predictor of attention levels if this is correct.

The TAA summarise, "Greater involvement in a programme... translates into greater commercial effectiveness. Viewers giving a programme a high impact score had better commercial message recall than those giving low impact ratings, and high programme. Impact was also associated with more positive evaluation of a programme's commercials, greater credibility of the commercial messages, and even an increased likelihood that viewers would express a preference for purchasing the advertised product in the future."

The TAA also found a weaker relationship between the Program Appeal Index and commercial effectiveness scores. Viewers may well choose to watch a programme but it may be simply to provide background to other household activities. As discussed above measures of liking cannot simply provide the key to involvement in isolation.

In a related field the authors produced a paper on reader involvement for the Worldwide Readership Symposium held in Berlin last year. (Wilson & Isaac) The paper reported on an experiment that measured thirteen involvement indicators for two monthly publications. A total of 414 respondents, with different frequency of reading claims, were asked to give details of their attitudes to the two monthly titles.

It was hypothesised that Trust would be the central exogenous variable and that through various causal paths, would influence the other variables. The aim was to explore this theoretical hypothesis using the empirical data relationships found in the survey.

Wilson Involvement Model for magazine readership

---

 

Research ApproachBack to top
Sheila Byfield (O&M, UK) and Doug Read (JWT, USA) discuss the pros and cons of several methodologies in their paper "Program Involvement: Does it have any value in the commercial television marketplace?"

Natural viewing is identified as the 'ideal situation' as opposed to laboratory experiments. The weight of evidence is against day after interviewing suggesting respondents have great difficulty recalling advertising when this amount of time has lapsed.

"The ideal situation is to allow natural viewing to occur" say Byfield and Read. The main problem associated with this methodology is cost. Researchers often have problems finding sufficient viewers who claim to have little or no interest in a programme. The MLI database is sufficiently large enough to overcome this problem.

The use of a one-week diary to record information on quality of viewing is validated by a British experiment the results of which were presented at the 1992 Symposium. Gunter, Clemens, and Wober reported on work completed for the Independent Television Commission "which explored a number of techniques for measuring audience satisfaction with television in terms of the perceived quality and enjoyment of programmes watched."

Part of this experiment focused on comparing several different methodologies for collecting quality of viewing data; diaries, self-completion questionnaires and telephone interviewing. A sample of 1935 respondents completed one of seven different methodologies to allow different time frames to be compared at the same time as the different interview formats. They concluded, "In future we recommend using seven-day diaries since more information can be economically collected from the same respondents without any reduction in its completeness."

This paper uses data from the MLI seven day viewing diary where possible but also employs a post programme advertising recall experiment along the lines espoused by Byfield and Read.

---

 

ResultsBack to top
1) MLI Database

Attention Level
The researchers began by exploring the relationship between attention levels and ratings. For the majority of programmes we found an association between the two elements. Programs with high ratings generally attract high levels of attention. However, the correlation between these two elements was not as strong as might be expected.

Overall, for programmes watched by 5% or more of all people, we found a correlation co-efficient of 0.1413 (p=.007). Interestingly the relationship between ratings and attention would appear stronger for females (0.1494, p = .005) whereas males showed no significant correlation.

Averaging daily ratings and attention levels reveal prime time attracts both the highest ratings and the highest levels of attention:

Graph 1a
Overall males are slightly more likely to be paying full attention to the programmes they watch. Females pay less attention than males to breakfast and afternoon television whereas their attention during the 11am to 2pm slot was significantly above the males. After the evening meal (8.30-10.3Opm) females once again are more likely than males to be paying attention to their choice of programme.
At different times of the day the attention levels by age groups varied. This analysis reinforces previous work suggesting the more detailed a target becomes the more useful involvement indicators become.

Graph lb
Analysis by age reveals that in general attention and ratings are related. Where programme ratings dropped among a given age range attention levels also declined. However, the decrease in attention is often surprisingly small given the relative decrease in ratings.

Graph 2 (Seinfeld)
For a handful of programmes the researchers found significant differences in the behavior of these two measurements. Perhaps most interestingly popular youth drama series such as Melrose Place and Northern Exposure attracted high ratings among the 18-24 age group however, the proportion of viewers claiming to pay full attention was relatively low.

Graph 3 (Melrose Place)
Graph 4 (Northern Exposure)
In fact, for both shows viewers aged over 25 were more likely to be paying full attention to the programme than those in the expected target market. Perhaps peer pressure has a role in forcing people in this age group to stay abreast of popular culture. A high proportion of the age group might then be watching these shows to keep in touch but with insufficient interest to pay full attention to all subtle plot developments. However, we offer no proof for this hypothesis.

Graph 5 (Bold and Beautiful)
For some daytime soaps ratings were found to be relatively flat across age groups (especially to 35) but attention levels increased.

Graph 6 (Hey, Hey it's Saturday)
For one of Australia's longest running programmes, Hey, Hey, It's Saturday, the researchers found an extraordinary relationship. Attention levels trended upward with age while the audience peaked in the 25-34 age range. This magazine style comedy show has been screened (with the same host) for over fifteen years. The popularity of this show in the 25-34 age group could then be accounted for by the fact these viewers grew up with the show. The lower ratings and attention scores among the younger demographics must raise a question over the programme's future.

Attention Level and Program Selection
Next the researchers selected twenty programmes which are broadcast every day (Monday to Friday) and six programmes which are broadcast once a week to examine in detail the relationship between programmes especially selected for viewing and the level of attention paid. (Appendix 1 gives a full list of these programmes).

For each programme we compared a viewers level of attention with their attitude to the programme (whether they had especially selected the programme for viewing). Data was taken from our syndicated study and analysis was restricted to the Sydney market during the period June to September 1995.

An analysis of correlation co-efficients found a significant positive correlation for the following programmes:

---

 

CHART 1
Relationship Between Interest and Attention for selected programmes - All viewers

Program	Category	Correlation
Young and the Restless	Soap	0.4850
Days of Our Lives	Soap	0.4074
The Bold and The Beautiful	Soap	0.2656
Wheel of Fortune	Quiz	0.2367
Family Feud	Quiz	0.2222
A Current Affair	Current Affairs	0.2098
Today Tonight	Current Affairs	0.2080
Seven Nightly News	News	0.2002
Sale of the Century	Quiz	0.1736
Price is Right	Quiz	0.1476
Home and Away	Soap	0.1022

The highest correlations were found to exist between the daytime soaps. Viewers who especially chose to watch these daily programmes are significantly more likely to give their full attention to the show.

Early evening quiz shows, the most popular current affair shows and the news on one channel also demonstrated a strong positive correlation between these two measures of involvement.

The remaining fifteen shows demonstrated no significant correlation between the two involvement indicators. Chat shows, soaps broadcast in the evening, drama and infotainment shows failed to provide any significant link between attention and liking (i.e. attitude to the programme).

Sex Differences
An analysis by sex revealed significant variations. For both sexes positive correlations were found with nine programmes:

---

 

CHART 2
Relationship between interest and attention - Males

Program	Category	Correlation
Young and the Restless	Soap	0.5424
Wheel of Fortune	Quiz	0.2556
Sale of the Century	Quiz	0.2444
A Current Affair	Current Affairs	0.2398
Family Feud	Quiz	0.2327
National Nine News	News	0.2111
Channel Ten News	News	0.1871
Seven Nightly News	News	0.1807
Today Tonight	Current Affairs	0.1557

---

 

CHART 3
Relationship between interest and attention - Females

Program	Category	Correlation
Young and the Restless	Soap	0.4366
Days of Our Lives	Soap	0.3850
Bold and the Beautiful	Soap	0.2660
Today Tonight	Current Affairs	0.2496
Seven Nightly News	News	0.2198
Wheel of Fortune	Quiz	0.2060
Family Feud	Quiz	0.1871
A Current Affair	Current Affairs	0.1822
The Price is Right	Quiz	0.1802

While the Young and the Restless displayed the highest correlation for either sex, only females demonstrated a high correlation with the remaining day time soaps. For men, there was a relationship between the two involvement indicators for the news: offerings on the three main commercial networks. On the other hand this relationship with news only held true for only one channel among women.

Our original plan was to use advertising response data from the MLI database but at the time of writing we were unable to match the weekly diary data with the weekly purchase data. Although this situation has been remedied through our "Closing the loop" model, the researchers chose to conduct a separate experiment to test the links between the Emotional and Rational dimensions of involvement and their relationship to advertising response.

The experiment was conducted over a one-week period and was restricted to the time period 6pm-8pm on weeknights and the weekend. A total of nine programmes were used for this experiment. News programmes were excluded. Following the screening of each targeted programme, a random cross section of telephone numbers was phoned in the Sydney Metropolitan area and viewers of the programmes identified. An approximate age and sex split was used to keep the data relatively representative.

A total of 368 respondents were surveyed across nine programmes with a sample per programme of between 40 and 60 respondents. The programmes surveyed were as follows:

Respondents once identified as having just watched the targeted programme were then asked their level of emotional involvement (i.e. especially choose to watch) and their rational involvement (i.e. level of attention paid tonight in particular - a lot, some, not much).

Respondents were then asked to identify all advertising that they could recall from the programme they just viewed. They were then asked if they were intending to buy in the product categories advertised when they next shopped. If so they were asked which brands they would be most likely to buy. They were then asked with prompts which brands they had seen advertised in the programme just viewed (aided advertising recall).

Intention to purchase scores and ad recall scores were averaged across all programmes and brands. Due to sample sizes individual brand and programme effects were not being evaluated in this particular research exercise.

Outcomes
Respondents were then allocated to dichotomous groups - High Interest (especially chose to watch n=208) and Lesser interest (any other reason, n=160) and High-average attention (n=280) and Low attention (not much, n=88) for the purposes of the analysis.

Mean unaided recall levels among High Interest (especially choose to watch) respondents and Low Interest (all other reasons) were different (i.e., 0.365 vs 0.281) The High interest group was 30% more likely to recall commercials in the programme viewed than the less interested group. These numbers are broadly in line with the earlier findings of Byfield and Read, but were not significant due to the sample size studied.

With aided recall we were unable to find any differences in mean recall levels between the two interest levels (i.e. 2.18 vs 2.04).

The situation with regard to attention level is more definitive along the lines of Gullen's 250% more likely to recall an ad. Mean unaided recall levels among Average to High attention respondents was .404 vs .09 for Low Attention respondents: i.e. a 400% plus difference. (F-test sig .01)

Clearly the findings on advertising recall have mirrored largely the earlier work. But this is really only the beginning, as we also need to discover the relationship between programme involvement and purchase intention.

Correlations were calculated for the relationship between interest, attention, aided and unaided advertising recall and intention to purchase. Those correlations are reported below.

---

 

Correlation Matrix

	Purchase Intention	Unaided Recall	Aided Recall	Interest (Emotional)	Attention (Rational)
Purchase Intention
Unaided Recall	NA **
Aided Recall	.021	.102
Interest	.044	.043	.098
Attention	.130*	.136*	.029	.344*

The correlation between intention to purchase and unaided recall is artifactual due to the methodology. However, other relationships appear significant. The researchers were still unclear at this point as to whether interest drives attention, attention drives interest on both variables work in a synergistic fashion.

The data for intention to buy brands is weak simply due to the fact that no attempt was made in the analysis to partial out non-target consumers, i.e. those who do not buy in the category. Indications are that this will produce a far tighter result in follow-up experimentation.

Aided recall does not correlate with any measure of involvement. The authors found similar in the print research reported in Berlin.

Path Analysis
LISREL 7 Path Analysis (Joreskog & Sorbom 1988) was conducted on the correlation matrix in order to establish the causal paths and these are represented in figures 1 and 2. As the researchers in this preliminary investigation were primarily interested in overall effects, data from all programmes areas combined.

The variables included in Analysis were Programme interest (Emotional dimension), Programme attention (Rational dimension), Unaided and Aided advertising recall and Purchase Intention at next shopping trip (Particular brand).

It was hypothesized that rather than there being an additive effect interest would drive attention and this would in term lead into advertising recall and purchase intention.

LISREL can be used for exploratory model development (Hayduk 1987) and (Wilson and Isaac 1995). Caution should be observed when finalising models that result from data driven modifications of models (MacCullum, Roznowski & Necowitz 1992). Nonetheless use of this exploratory technique is useful for theory building followed by subsequent confirmation testing.

Causal direction from interest to unaided recall

The path does not incorporate aided recall that appears not to have any significant relationship with programme involvement.

---

 

Figure 2 Causal Path for Intention to Purchase

Clearly in both the instance of unaided recall and the instance of purchase intention, programme interest (the emotional dimension) drives programme attention (the rational dimension) and this is likely to lead to a substantially greater level of unaided advertising recall and possibly purchase consideration.

The writers would like to emphasize that the results are at best indicative on purchase consideration, whereas they confirm what has gone before on advertising recall. However, it is the purchase intention and ultimately purchase at which we will be directing our investigations during 1996 and hope to be able to report more comprehensively at our next opportunity.

What we can however say at this point is that the Emotional dimension (interest) drives the Rational and the measures are not additive - in short there is no greater predictive power from a composite measure*. Although we will continue to measure programme interest, we can confidently say that programme attention measures will provide the overlay on programme ratings we need to determine potential sales conversion.

* A separate multiple regression analysis showed that the emotional dimension could add no more to explained variance after the rational dimension had been entered in the equation.

---

 

Implications for syndicated electronic media researchBack to top

All electronic media research providers are struggling to balance the demands from TV organisations and advertisers to provide depth of data against respondent fatigue. This is clearly the case if greater requirements were placed on 'people-meter' households. Notwithstanding this problem, the electronic media would value a measure of involvement that could be used to increase advertising revenue.

Media planners want access to an indicator, which could show what proportion of a programme audience has high attention, and might therefore be more likely to respond to the advertising appearing in that programme. We know that there are real differences in the edibility of certain programmes to capture viewer attention.

Everyone, especially media researchers, want the solution to be easily measurable. To be included in syndicated projects the selected indicators of involvement must not over burden the respondent. This study is, albeit tentatively, indicating that attention alone may be enough. As an industry we may be moving closer to an acceptable level of simplicity of measurement.

To become a valuable selling tool any measure of involvement must not only accurately reflect higher advertising recall among involved viewers but also demonstrate, a greater propensity to purchase advertised brands. This work we have begun and will continue to report on it over the next two years.

Although programme interest is the psychological or emotional driver behind attention level, it is not needed as a measure once we understand programme attention levels among the targeted consumers in the television audience. The researchers wish to emphasize that a critical element of this approach lies in the ability of the research to isolate attention levels with the target - not just the general audience as these may differ.

Further work
The sample for this pilot survey was relatively small (368 respondents). Analysis of the effects of certain variables was, therefore, not possible to complete with satisfactory reliability. For example, advertising recall and purchase intention were averaged areas all brands so many individual effects could not be fully explored. Fortunately the MLI database was able assist on many of the basic correlations.

Notwithstanding, from this initial analysis there is evidence to support the assertion that aided recall of advertising has little to offer as a measure of advertising effectiveness.

Some of the relationships referred to in our model are not especially strong, although they are statistically significant. Further work with a larger sample size needs to be completed to validate the results of this survey. In addition a refinement of these analyse to target audiences in relevant consumption categories is likely to substantially strengthen the relationships highlighted in this study.

---

 

Appendix 1Back to top

List of programmes included in analysis of comparison between programmes especially selected for viewing and attention level:

| Background to the issue | Recent Research Findings | Research Approach | Results |
| Implications for syndicated electronic media research |

Back to top