Mean lifetimes of bound states in N-15 were inferred from the Doppler-shift-attenuation (DSA) of gamma rays produced in the inverse reaction H-2(N-14, pgamma) and from the primary gamma-ray-induced Doppler broadening (GRID) of secondary gamma rays iu the thermal-neutron capture reaction N-14(n, gamma). Targets for the DSA measurements were prepared by implanting first neon into a gold backing and then deuterium into the same region such that deuterium was trapped at the neon precipitates. To find out experimentally the initial velocity distributions of the recoiling N-15 nuclei, another target was prepared by implanting deuterium into silicon, which is a slow stopping-power medium. Computer simulations with the Monte Carlo method and experimental stopping powers were used in the DSA analysis of the gamma-ray line shapes. This analysis yielded the following lifetimes for eight bound levels in N-15: 43 +/- 4 fs (5.30 MeV), < 12 fs (6.32 MeV), 11 +/- 2 fs (7.16 MeV), < 3 fs (7.30 MeV), 129 +/- 6 fs (7.57 MeV), < 3 fs (8.31 MeV), < 10 fs (8.57 MeV), and < 4 fs (9.05 MeV). GRID measurements were made with melamine (C3H6N6) and with silicon nitride (Si3N4) as both targets and slowing-down media. Measurements were also made with air (80% nitrogen). The Doppler-broadened gamma-ray line shapes were analyzed by molecular-dynamics simulations of the slowing-down process to obtain the lifetime values of 40 +/- 3 fs (5.30 MeV), < 2 fs (6.32 MeV), < 3 fs (7.30 MeV), and < 3 fs (8.31 MeV) for levels in N-15. The extent to which the mirror symmetry of levels in N-15 and O-15 is valid is examined in some detail.