The demand for detecting food adulteration has recently grown, due to its economic and health implications. Infrared spectroscopy provides an efficient method of collecting data for use in food authenticity analyses. Statistical methods are routinely employed to analyze spectroscopy data in order to effectively detect adulterants in different food items and ensure food authenticity. This work presents a novel partial membership model for mid-infrared spectral data. Our approach not only detects the level of adulteration but also provides information on the spectral regions most affected by the adulterant. These insights can be used in combination with subject-matter expertise to characterize the chemical impact of the adulteration.