Experimentation Critique the Researchers Decided

Experimentation Critique

The researchers decided to do an initial test in an area which they felt was a new area of research, direct response radio advertising. Some research had been conducted on direct response television advertising, and this provided some of the basis for their study. The researchers first defined "direct response," based on the previous studies.

To conduct the study the researchers used two different ad campaigns and aired them on two different Dutch radio stations. The stations served different geographic regions but had similar audience demographics. The researchers chose three independent variables to study. They chose to study the day of the week, time of the day and the ad placement within the break. These variables were the three that the researchers felt would have the most impact based on previous research. They also felt that the length of the break may affect the results, but they were unable to control that variable. The researchers used two measures in this study - responses per rating point and responses per DFL 1000 media costs.

The survey found that the day of the week and the time of the day were most relevant. They found that 2pm-4pm was the best time to elicit action. They showed that 7am-9am and 4pm-6pm were the worst times to elicit action, which they had hypothesized since direct response is more difficult when listeners are driving. Thus, they found that for direct response advertisements these peak hours with the most expensive ad rates were also the poorest in terms of response. The survey found that the day of the week was less important than the time of day. However, they found that the particular day of the week was relevant. Monday, Tuesday and Wednesday were the strongest three days.

The survey found little evidence that the position of the ad in the break was relevant.

These three independent variables were chosen because they are three that in most cases can be controlled by the advertiser. The researchers found, however, that they did not have complete control over the time of day. This affected their sample size for the 7am-9am period, where they did not get the desired number of ad placements. Because of this, the study's findings for that time period should be considered to be of lower validity. Radio advertising, and especially direct response advertising, relies on a certain degree of repetition in order to be effective. With direct response, the listener must record a telephone number or postal address, for example. Repetition increases the odds of this. Because the 7am-9am time period did not receive the same degree of repetition as other time periods, the low results were compromised.

There were other potential independent variables that could have been chosen as well. For example, both of these ads featured direct action via telephone, but there are many different forms of direct response, such as Internet, mail or text message. This was the initial survey on the subject, however, so this should not be considered too strong a fault. This research is essentially providing the framework for more in-depth future study on this subject. The authors should have, however, spent more time explaining the limitations of their study and highlighted such areas of potential future research. They identified a handful of other potential independent variables at the end of their report but dismissed the potential for measuring these in the field. They did not cite any research showing that these factors should be dismissed as such. Therefore, they selected good independent variables but left many other potential, equally good ones, unselected. Future study could hopefully address some of these other variables.

The dependent variables in the study were the measures and observations. In this study, the calling attempts were measured over the 15 minutes following the broadcast. This is consistent with past research on the subject. The researchers determined that other measures such as sales were less valuable because of external factors such as the skill of the call center representative. This is a fair assessment because it separates the effectiveness of the ad from the effectiveness of the company's other activities. The use of time interval sampling is also appropriate, although this choice was only based on the slim body of previous literature on the subject.

The impact of repetition in radio advertising could result in different results if the dependent variable was measured using continuous measurement.

The sampling used to gather the subjects was relatively poor. Only two radio stations were used in the measurement and they both had the same homogenous demographic. The study is therefore limited to ads placed on stations with that demographic. Stations with different demographics, for example younger or gender-specific, may see completely different results. The objective of the study was fairly broad in focus, yet the subjects were chosen to be fairly narrow. This despite the researchers mentioning in the introduction to their paper that radio stations typically have narrow, homogenous markets. The researchers should have expanded the scope of their study in order to derive more accurate results, given that they are attempting to isolate results based on society as a whole.

Moreover, the sampling size was insufficient. The researchers received an abundance of zero responses. They tried to account for this with a Tobit analysis. It would have been more effective had they simply increased the sample size. This may not have provided a different percentage of zero responses, but it would have yielded more positive responses. In advertising, zero responses are a valid response. It is important to advertisers to understand when their ads are effective and when they are not. A larger sample size would give enough positive responses to derive stronger information, while the value of the zero responses is retained.