• No results found

Limitations and Directions for Future Research

Duration of ad

5.9 Limitations and Directions for Future Research

A student sample was used for Studies 1 and 2 of this thesis. Such a sample is appropriate when internal validity is the main consideration as opposed to external

validity (Lynch, 1999). Generalisations about different target audiences and the general population cannot be made, however, given the results an extension of the study to other groups within the population would help address external validity. Study 3 used young drivers from the general population and thus overcame the limitation of an emphasis on internal validity of the first two studies of this thesis.

The age of the students directly corresponded with the high risk category of road users;

however, university students represented a more educated viewing audience. While there is an association between drink-driving, drinking, smoking, and low

socioeconomic status (Williams, Lund and Preusser, 1986), no such associations have been established for speeding behaviour.

T h e control and experimental groups in Study 2 contained uneven numbers of males

and females per advertisement group. Also, the prior driving behaviour of participants, in terms of speeders or non-speeders, was not used to quota sample participants into the advertisement groups. It is possible that even if quotas had initially been set, the unpredictable attrition rate of the three-week experiment would have resulted in dissimilar profiles of the advertisement groups in regard to gender and speeder classification. Again this limitation was addressed in Study 3, where quotas were assigned and achieved.

The CRM-dial used in Studies 1 and 2 could be classified as a mixture of both

emotional and cognitive responses to the advertisement. When attempting to separate participants' emotional reactions from their cognitive responses to these emotive advertisements, psychophysiological recording, such as EDR measurement, would be a purer way of obtaining emotional reactions. This technique was used in Study 3.

The VST was developed and validated in the United Kingdom and because of this it showed British road signs. This could have been a distraction when Australian drivers were undertaking the test. Thus the AVST was developed and validated to be used in the third and final study.

A controlled experiment allows for the effects of an individual advertisement to be measured rather than trying to measure wearout in the context of advertising campaigns and other real-life influences. An important qualification for Study 2 was the forced-exposure design. "On-air", these TV commercials may begin to lose attention

substantially after half a dozen or so viewings (although the results of the static

measures condition of the experiment, using the equivalent of six viewings, suggested that consumers still reported "watching most" of the commercial). After nine or more

on-air viewings that were artificially simulated in this experiment, loss of attention may cause the advertisements to also lose their influence on drivers' speed reduction.

The use of an experiment can be criticised for its superficial setting; for example, in this study, subjects paid absolute attention to the advertisement whereas in the real-life

situation of watching TV at home, they may not have paid attention to the advertisement to the same extent. It is suggested that three experimental advertising exposures for the dynamic rating component of Study 2 would be equivalent to approximately nine

exposures in-home viewing situations (Rossiter and Percy, 1997), with the results of this study showing that even with forced exposure to the advertisements there was a decline in the attention paid to the advertisements over the three weeks in the static measures component of Study 2. Most road safety advertisements are aired in short

bursts, for example, just before a holiday period, thus making nine exposures realistic.

Moreover, due to the intensive measurement requirements with CRM recording and participants tested individually, only one fear-relief commercial and one fear-only advertisement were investigated. The static measures condition of the experiment used two advertisements of each type and they produced similar results within type.

However, replication with a wider set of fear-appeal commercials is needed to rule out extraneous influences that may be due to the particular advertisements that were

selected for the study. Despite the measures for each advertisement of the various

advertisement execution variables, that revealed similar responses across advertisement groups for these control variables, the four advertisements used in Studies 1 and 2 still

differed significantly in their presentation that m a d e it difficult to compare results

between the advertisements. A cautionary note to be added to these findings is the possibility that the wear-in of the advertisements could be a characteristic of the advertisement rather than the participants.

Further development of the fear patterning theory is required by patterning and testing more anti-speeding advertisements.

In particular, paired counterparts of each of the original four advertisements used in Studies 1 and 2 need to be developed. This should then assist in overcoming the

potential issue of whether it may be variations in advertisement execution, rather than the fear patterns, influencing the outcome measures. That is, advertisements with the same executions, but different endings comprising fear or relief components need to be compared to be able to have more control over measurement of the persuasive

effectiveness of the patterns of fear used in advertisements. Again, Study 3 addresses this concern.

The effect of personality traits were not investigated in this thesis, and could be

included in future research on fear patterns. For example, trait anxiety is an individual difference variable that has been included in studies on fear appeals (Witte and

Morrison, 2000). Participants could be quota sampled according to high and low

anxiety, gender and speeder classification. However, for the purpose of this thesis, gender and speeder classifications were considered more influential variables on the AVST, particularly given budgetary and time constraints imposed on cell sizes, thus personality characteristics were not included in any of the studies. It is also less

practical to use personality traits, such as anxiety, to segment a market, whereas gender

and speeder classification are more identifiable and practical segmentation variables.

5.10 Conclusion

The major research objective of this chapter was to analyse how viewers reacted to successive exposures of different fear pattern TV advertisements, using both static and dynamic measures of advertising wearout. Emotion and attention wearout occurred

immediately for all fear patterns, using static measures of fear and relief. It was shown that fear-only pattern advertisements had higher rates of wearout than fear-relief

patterns. Additionally, fear-relief advertisements were more effective, using a behavioural dependent variable of speed-choice, the VST. The dynamic measures

condition of the study showed that the fear-only "shock" advertisement began to "wear-in" and became more effective with repetition but "fear-relief continued to be the most effective type of commercial for deterring speeding among young drivers.

The following chapter again tests fear patterns within anti-speeding advertisements.

However, for the final study additional advertisements were patterned using

psychophysiological recording and the AVST was applied to determine the effect of the various advertisements.