Recently several methods for obtaining clinical measures of loudness growth through the use of categorical scaling (CS) have been proposed for the selection of hearing aids. These methods use differing test signals or suggest frequency-specific level corrections in an attempt to reflect the loudness perception of hearing aid-processed speech. While some decisions regarding the stimuli utilized for loudness perception procedures are based on measured relationships to speech signals, the effect on loudness perception of changing signal type and bandwidth (as measured by CS) remains unclear. The relationships between the CS loudness growth of signals of differing type, (pure tones, noise bands, filtered/temporally inverted/passband speech) and bandwidth were examined for subjects with both normal and impaired hearing. Results suggest that when the bandwidth is similar (e.g., pure tone and 1/3 oct), signal type does not have significant bearing on loudness perception. As expected, increasing the bandwidth beyond the critical band affected loudness growth, as wideband speech stimuli were judged to be significantly louder than narrow-band speech at equivalent overall sound levels. In this investigation, similar loudness growth patterns were noted across category ratings for all test signal bandwidths. In contrast, loudness growth data obtained using an intelligible speech signal [Cox et al., The American Academy of Audiology (1994a)], revealed loudness growth patterns for speech were fundamentally different than those obtained for tones. Implications for hearing aid fitting strategies are discussed.