Messung kontinuierlicher Hörerbewertungen mit dem CRDI

Abstract

The following article describes the measurement of listeners’ reactions using Continuous Response Digital Interface (CRDI). This gage consists of several input devices, which are connected with computer software that lines the manual reactions of the test participant. The input devices are resistors with a turnable pointer on the front. Participants move the pointer between the two extreme points (amplitudes) while listening to an acoustic stimulus to express their assessment of the stimulus. The amplitudes are named with semantic differentials, e.g. “I like it very much“ vs. “I do not like it at all“. The CRDI was developed for research about the effect of music, and the manual operation is especially appropriate to express unreflected impressions from the participant. Unger (2012) and Altstadt (2013) were the first who applied the CRDI to speech signals. The CRDI-study by Unger (2012) had the aim to develop and test a setting in which listener reactions to radio news were measured. The study was separated into pretests (with eleven subjects in two groups) and main tests (with 18 subjects in three groups). The pretests showed that older as well as younger participants did not have problems with handling the input devices. However, participants indicated it was difficult to listen and to asses at the same time. In the main tests the three categories “entertainment value”, “informativeness” and “comprehensibility” were chosen for assessment with CRDI. The data were shown as graphs. An analysis according to paradigms of musicology was also made (correlation calculation, marking of characteristic “hills and valleys” in the graph). Altstadt (2013) continued this study. The aim of his research was to test more assessment categories and to optimize the CRDI-method in speech science. Smaller groups of participants assessed clippings of radio-moderation in five different categories (e.g. “sounds read – sounds spontaneous“; “sounds very put-on – sounds very natural“). Setting and data analysis could be improved. It was shown that only a few categories are suitable for measuring listeners’ reactions, because cognitively demanding categories overstrain subjects. Nonetheless, both Unger and Altstadt suggest a modification of the CRDI.