It’s one of those great paradoxes of the scientific world, particularly for those social scientists seeking to adopt a more natural science model of enquiry.
On the one hand, Skinner and others taught us that science is a process of exploration in which general laws are discovered through systematic replication of experimental results. At the same time, it is extremely difficult to get a paper published if it is a pure replication of past work because it is deemed to have no original “contribution” to make. Indeed, a survey in 2012 of papers accepted by psychology journals since 1900 found that only 2 studies out of every 1000 involve replication in any way, shape or form. Quite depressing! It is therefore refreshing to find that one leading journal, Social Psychology, has devoted its entire latest issue to systematic replications of well-known studies from the past. It’s a fascinating read and, I think, completely justified and long overdue because quite a number of studies failed to replicate past results, challenging some of our long-held assumptions.
A number of studies will be of interest to both consumer psychologists and marketing practitioners alike, covering topics as diverse as the use of stereotypes through to temperature effects. A paper that particularly caught my own attention was one which reports three experiments designed to recreate the conditioning of consumer product preferences through the use of popular music; a staple “we-all-know-that” in the retail marketing literature. Interestingly, only one of the experiments actually managed to achieve the “well-established” results, with no overall effect when the three studies are aggregated.
The authors challenge the findings of the past a little on the basis of these results, but I’m not convinced we should dismiss the previous work in this area so easily. The two studies that failed to replicate former ones used the exact same pieces of music, whereas the experiment that “did work” used contemporary music on the basis that this is exactly what the researchers in the past were doing at the time. Seems to me this dimension is worth further exploration too, another reason why we should be a little more open to engaging in replication in the best interests of science – journal editors, please take note!