Online Research… and All That Jazz! - The practical adaptation of old tunes to make new music

Steven Gittelman and Elaine Trimarchi


It is difficult, if not impossible, to determine the representative qualities of online data. The probabilistic models that once underpinned our research are re-visited. Here we take a hard look at the use of behavioral profiling of respondents to augment the old and create new representative sampling frames. We are at a crossroad; either we ground our research with proper theory or we lose to those who claim to read the wisdom of crowds.


We are in the midst of a social revolution. The online universe is changing our world. For us in market research, many are at a state of dissonance, unresolved and uncomfortable; the probabilistic sampling frames of the past have eroded. They were our structure. Where once we could rely on a theoretical underpinning to give our work a solid standard, we now float in an unstructured and anxious space where we cannot gain our footing. We question the quality of the samples and are seeking ways to reassure ourselves.