Vocational interest theories imply a person-occupation cross-level interaction effect (e.g., artistic interests predict job performance better in artistic occupations), which has rarely if ever been tested as such. Using a large military sample, we findperson-occupation interest congruenceeffects are supported: (a) on core technical job performance for six of eight interest dimensions, and (b) on job performance ratings for structural/machines and rugged outdoors (i.e., Realistic) interests. Another cross-level interaction involves theperson-occupation gender congruenceeffect. Our data also confirm the job performance gap favors men when in male-dominated occupations, but favors women when in gender-balanced occupations. Due to strong overlap between vocational interests and gender, we conduct a critical test of whether person-occupation interest congruence might be due to person-occupation gender congruence. In only two of six cases (i.e., rugged outdoors and administrative interests), did the person-occupation interest congruence effect disappear after controlling for the person-occupation gender congruence effect; the gender congruence effect also remained significant after controlling for the interest congruence effect. Consequently, the two cross-level interactions on job performance (for vocational interests and for gender) appear to represent distinct effects. In a second, service organization sample, the person-occupation interest congruence effect (for Realistic interests) on job performance ratings and the person-occupation gender congruence effect were both replicated.