Context Effect in Fan Sound Evaluation by Jury Tests Using the Method of the Semantic Differential

Air delivering fans and fan systems are omnipresent in industry and private households. In general it is economically not feasible to reduce the sound emitted by technical devices below the threshold of hearing. Hence, end-users increasingly demand at least a high quality of the sound of a device. Jury tests of sound assessment in a laboratory environment are preferable with respect to cost and time, however, may lack external validity because of missing the real life context. Objective of this paper is to investigate whether the assessment of fan-related sounds in a laboratory is affected by test instructions which give information on the context in terms of the product to be assessed, the environment, the benefit to the user and the typical duration of sound exposure. Starting point is a semantic space with a total of 37 adjective scales, specifically created for a subjective description and assessment of fan and fan system sound in jury tests. The contexts given in the instruction to the subjects were designed to take into account different fan-related products, different environments and different ownerships. A principal component analysis of the data from comprehensive jury test (20 subjects × 10 sounds × 37 adjective scales x 3 sessions) revealed five perceptional dimensions "quality", "spectral content 1" and "2", "time structure" and "power". Although most trends observed are within the estimated uncertainty, the statistic analysis and the visualized results suggest a small context effect on the rating of various fan sounds. A more relevant finding, however, is that the two imaginary contexts imposed to the subjects did not change the overall order of ratings of the sounds. Hence, if for any reason jury tests with real life judgement are unfeasible, it is recommended to implement at least an imaginary context in the test instruction.