This experiment has, we believe for the first time, demonstrated that it is possible (in fact without having to "mount eye cameras onto magazines or their readers" to take a technique on which over the years millions of dollars and pounds have been spent to little avail, and with the help of laboratory equipment, to improve the technique, and to bring it into line with (advertising) reality. Clearly, in view of the differences we have in the past found to hold for claims for different types of advertisement (colour is mono, size and so on) it will be necessary to conduct further experiments in order to allow for the effects of such variables in the weighting system. It might ultimately be necessary in the future, before conducting large-scale reading and noting studies, to precede them with a small pilot study using DEMOS to ascertain the correct weights to be used for readers of the particular publication(s) concerned.
The process here described is quite straightforward in concept, although highly complex mathematically. It is not the purpose of this paper to discuss in detail the technicalities, and it will be sufficient to remark that the factor Analysis used in MASA uses the FA6 programme of Cybernetic Research Consultants Ltd., providing oblique factors . An addition to the programme has been written by C.R.C., allowing for the calculation of variable shifts by calculating differences for occasional two and occasional three from a normalised score. Factor movements are calculated by summing the individuals' scores, each weighted by the regression weight of the variable in the factor. The method described provides precise, statistically valid, measurements of attitude change caused by an advertisement.
This seminar revealed the gulf between the passionate non-verbalists who believe structured verbal response is inadequate to describe meaning (for them the Black Box, the Group Discussion, the indirect inference) and those who seem to believe that thoughts can be put into words and measured, and that the resultant words lead to ideas. So, though we would never have guessed it, we turn out to be not just market researchers, but two schools of philosophers.
One of the realities of advertising research is continuous, usually syndicated, research on recognition and recall, and - roughly these last five years - of intent-to-buy. That seems to go on whatever the latest fashion in schools of thought (or schools of action) about producing advertising.
Some people make a distinction between "advertising theories" and "research models". I am not going to do that. I am interested in theories of persuasion and behaviour which underly both advertising planning and market research. One of the aims of this paper will be to demonstrate the constant interaction between advertising and research theories. The kind of research done depends on what theories of advertising are current; the kind of advertising fashionable depends at least partly on the kind of research information available. My own experience has been very much within the English advertising world. So I am going to centre my discussion on what has been going on in London in the past decade or so.
In this short contribution it has been necessary to state theoretical arguments briefly rather than argue them out as fully as would be possible. The main contention is that if we are going to assess advertisements, then the complexity of advertising theory ought to make us use a black box, 'anything-can-happen' approach. It is possible to go on from such an approach and consider how advertising has worked and examples have been given. If we can find cases where the attitude change model is appropriate it can be used very effectively, which makes attitude work worthwhile . It is suggested, however, that to assess all advertisements in terms of a rational attitude model is unprofitable, inappropriate to the kind of learning involved in many market/media situations, and unfair to advertisement creators , who intuitively may have hit upon a better way of communicating. Whilst, it is argued that assessment should be on such a 'black box' basis, it is also valuable to be able to provide some indications to creators of advertisements as to how their advertisement was working, since it is only by such a feedback process that communication skills can be acquired and reinforced .
I think seminars are of two kinds - there are those which confirm one's prejudices, and those which enable one to re-organize and rearrange one's prejudices, and I would say that this is certainly of the second kind - that I have rearranged all of my prejudices and dusted them up a little. Certainly, being serious, this seminar has made me think, and I have found it very valuable. I'd like today to speak under four headings, which I hope will do justice to most, though not all, of the topics that we've discussed. I'm not going to refer back to all of the papers in the seminar, by any means.
Within this paper I will be concerned with two specific aspects of evaluating this particular Fishbein study: A) Does the theory work? is there a clear relationship between the measures taken and behavioural intention? For if there is we can isolate and assess the worth of contributory components; B) Is it useful? Does it contribute or is it likely to contribute to advertising or marketing planning?
In this paper we would like to: 1) initially show some examples of the results we have obtained using the model, indicating the various degrees of success and failure we have achieved; 2) briefly discuss some of the factors which TNA has found critical (i.e. variations in the application which can and do substantially affect the results); 3) finally, go on to indicate some of the conclusions we have come to in order to make our Fishbein results more reliable, and we believe more meaningful to advertisers .
There are some criticisms and comments we would like to append to our paper. The criticisms are self-directed, since I feel that we may not have fully communicated some elements of the T-Meter technique.
The author comments some of the main points raised in the papers on Fishbein Theory.