This study determined the learning styles of full-time post-secondary vocational education students and hotel general managers in the Bahamas. The study also examined differences in learning styles of students and managers, and examined differences in learning styles among students majoring in four full-time programs at the Bahamas Hotel Training College (BHTC) in Nassau, Bahamas;The sample contained two groups, 94 full-time post-secondary vocational education students at BHTC, and 24 hotel general managers in Nassau. The Marshall and Merritt (1984) Learning Style Instrument-Semantic Differential (LSI-SD) was used to measure learning styles based on Kolb's (1984) leading style model. Frequencies, percentages, and means were calculated for the demographic characteristics. The 40-item LSI-SD was used to measure four learning traits (active experimentation (AE), reflective observation (RO), abstract conceptualization (AC), and concrete experience (CE)). The learning traits formed two learning dimensions, AC-CE and AE-RO. Using the AC-CE and AE-RO dimensions as axes, individual scores were plotted so that each quadrant represented one of Kolb's four learning styles--converger, diverger, accommodator, and assimilator. Results indicated that the majority of the students (n = 44, 50.7%) and managers (n = 21, 87.5%) had a diverger learning style. No significant differences were found among students grouped by major;The independent variables of interest were gender, age, and Caribbean home. Of the total group of students and managers, 50 (42.4%) were male and 68 (57.6%) were female. Eighty-eight (74.6%) were less than 25 years old, and six (5.1%) were over 50 years old. One-hundred-four (88.1%) of the respondents identified the Caribbean as their home, and 14 (11.9%) reported their home outside the Caribbean. Discriminant analyses were used to determine whether the independent variables could be used to classify the learning styles of the respondents;Results of the discriminant analysis using prior probabilities indicated that all 109 respondents (excluding the nine cases that were not classifiable by learning style) were predicted as belonging to the diverger group. The prior probabilities were established as the actual group membership for each learning style. This resulted in correctly classifying 59.6% of the cases. When discriminant analysis was performed without establishing prior probabilities, 39.5% of the cases were correctly classified;Finally, factor analyses were used to cross-validate the learning traits (AC, CE, AE, and RO) and the AC-CE and AE-RO dimensions. Using a four-factor solution, three factors emerged which resembled three of the learning traits. Concrete experimentation was not represented in the fourth factor. Results of a two-factor solution suggested that axes for this sample did not resemble AC-CE and AE-RO, but rather the learning dimensions of AC-AE and CE-RO appeared to be better structures.