Comparison between multilevel and traditional logistic regression analysis of laboratory indoor annoyance caused by shaped sonic booms

In the 1970s commercial supersonic flight was forbidden over communities because conventional N-wave sonic booms caused excessive annoyance. Through recent advances in aircraft design techniques, conventional sonic booms are being modified into quieter, shaped sonic booms. As noise regulators consider allowing shaped booms over communities, NASA is providing data and expertise. Recently, a laboratory study quantified the additional annoyance caused by directly felt structural vibrations (J. Rathsam and J. Klos, “Vibration penalty estimates for indoor annoyance caused by sonic boom,” J. Acoust. Soc. Am. 139 (4), 2007 (2016)). To confirm experimental results and complement the traditional logistic regression analysis, a multilevel logistic regression model was fit to individual-level data. The multilevel analysis allows us to simultaneously model individual- and group-level factors. As a result, it is straightforward to test for relationships between annoyance and individual-level factors such as noise sensitivity.