This answer is not correct but is close. The absorption between 3500 cm-1 and 2500 cm-1 is the key to this problem since all the possible compounds contain a benzene ring. The alcohol oxygen-hydrogen stretching absorption does show a strong broad band centered around 3400 cm-1, but that band is not as broad as the one shown in this spectrum. The alcohol band returns to the baseline before 3000 cm-1. You are on the right track, think about where the other compounds absorb and try again.