Television audience measurement has kept pace with technological developments and has become increasingly sophisticated over the decades, evolving from surveys based on respondents' statements to automatic measurement requiring virtually no human intervention.
The oldest technique, which is now the simplest to implement and by far the least expensive, consists of asking people what they watched on television the previous day. This is known as the declarative method.
Unfortunately, the advantages of simplicity and low implementation costs of this type of methodology are offset by various biases inherent in the way the information is collected.
The first of these biases is related to people's perceptions of television programs and channels. Respondents often tend to over-report their consumption of programs they consider prestigious and under-report programs with a poor image: for example, literary or political programs will score higher than reality TV shows, even if the latter are watched more than the former.
Another bias is linked to the fact that we rely on people's memories: we have all experienced this, it is often difficult to remember in detail what we did yesterday. Furthermore, the finest granularity that can be achieved with declarative questionnaires is generally 15 minutes. However, we know that there can be significant variations from one minute to the next, particularly when a program ends.
This type of methodology applied to television is therefore a reconstruction of reality, the low accuracy of which has justified, over the years, the implementation of more "technological" data collection techniques.
Beyond the biases associated with questioning, declarative studies require finding people who are willing to spend between 45 minutes and an hour or more answering questions, which can in itself cause another bias linked to the profile of the respondents.
Face-to-face, online, or by telephone?
There are several ways to collect declarative information: through a representative sample interviewed face-to-face, by telephone, – or online in countries where internet penetration is sufficient – or via a panel of viewers who complete a daily viewing diary over a period of several days, noting down when they and other members of their household watch television.
Face-to-face collection was historically the first method to be developed. Interviewers built their samples by approaching people on the street who were likely to be willing to be interviewed.
This method has the advantage of allowing good communication with respondents (good understanding of the questionnaire, visualization of logos or images, etc.) but has the disadvantage of low geographical dispersion, as interviewers can fill their quotas without moving from the place where they started working.
Until recently, telephone surveys were the preferred method of data collection because they were much less expensive and more efficient than other types of collection: everyone can be reached by phone, interviewers do not have to travel, samples can be widely dispersed, and quotas are easily met. On the other hand, questionnaires must be fairly short and must not include logos, images, or documents to show.
In recent years, this type of data collection has been supplanted by online questionnaires, which offer all the advantages of telephone surveys while correcting their shortcomings, at an even more competitive cost of collection and processing.
However, this type of method can only be systematized when internet penetration is sufficient to allow all population categories to be selected without difficulty.
As we can see, in the field of so-called declarative studies, there is no perfect collection method; each has its advantages and disadvantages. Depending on the level of funding that can be generated, the nature of the information required, and the level of digitization in the geographical area in question, one or other of these techniques can be implemented by adjusting sample sizes to improve the accuracy and precision of the analyses.
Automatic or semi-automatic techniques (audimetry, passive audimetry, etc.), which we will discuss in a future article, correct many of the shortcomings of declarative studies but require much heavier investment, necessitating a minimum advertising market size to enable long-term financing.
To be continued
See also: Definitions – Audience share, Cumulative audience, and Listening time