More and more mixed-mode techniques in statistical surveys

Interview with Demetra Opinions.net, guest of the Assirm Marketing Research Forum of 29 October

In anticipation of its participation in theAssirm Marketing Research Forum, which will be held on October 29 2019 at Palazzo Mezzanotte in Milan, we interviewed Angelo Rodolfo Tomaselli, Scientific Manager.

The central theme of the 2019 edition of the Marketing Research Forum is “Design the future. Research, Strategy, Insight and Data Intelligence for business growth ". What is the vision of Demetra Opinions.net?

The changes, not only technological, are imposing the use of mixed-mode techniques to reach target populations otherwise difficult to reach or with low response rates; more generally, these techniques are used to try to contain the trend decrease in the response rate.

The main junction of the technological and methodological transformations of the last thirty years is constituted by the co-presence of fashions and therefore of frame population in the realization of the field and the consequent introduction of bias due to the greater complexity of the sample design. Having more frame population means having different possibilities of finding the statistical units that make up the different frames, different propensities to answer, different costs due to the use of different fashions.

What does this mean for statistical research activities?

This increase in complexity forces us to articulate structured survey procedures that involve the involvement of professionals and different technological structures. Furthermore, it forces us to make the research design more complex to take into account the different probabilities of contact of the statistical units belonging to the different frames. This implies the introduction of a posteriori weighting techniques that keep these aspects under control.

What difficulties are you facing today?

A first practical problem hardly surmountable is to foresee a priori response rates by mode and then allocate for the different fashions the amount of resources to be used for conducting a survey. This difficulty, within the framework that the recent technological revolutions are taking shape, is a difficulty generated by two converging aspects: on the one hand, a market for statistical research which is earmarking ever more scarce resources for field jobs; on the other hand, a growing complexity of sample designs that would require the use of professional figures and technological structures (hardware and software) whose costs, at times, are incompatible with these resources. This contradiction has only been partly absorbed by technological development, which has reduced the costs for the realization of telephone surveys and still remains the open ground of the comparison between market logics and the logic of scientific development.

What future is ahead, then?

These difficulties represent the new challenges to which the research methodology community is trying to respond even if the speed with which the content of the phenomena to be investigated changes and the way in which these phenomena are investigated makes it difficult to keep pace with changes or, for say it with Edith Desiree de Leeuw (2017): "never a quiet moment ... in the world of investigations".

Publication on ASSIRMFORUM