One of the standard methods for determining atomic transition probabilities is to combine branching fractions measured using Fourier-transform spectrometry with radiative lifetimes measurements using laser-induced fluorescence (LIF). This combination of techniques provides an efficient method for measuring large sets of accurate, absolute transition probabilities. The radiative lifetimes, which provide the overall scaling for the transition probabilities, can be measured routinely to ± 5% accuracy using time-resolved LIF. Although the time-resolved LIF technique we use does not achieve the accuracy of fast-beam LIF, the time-resolved technique does enable us to make measurements at a far greater rate (hundreds of level lifetimes per year). Care must be taken, however, to understand and control the systematic effects in time-resolved LIF measurements to maintain ± 5% accuracy over a wide dynamic range and hundreds of lifetime measurements.Over the last 25 years, we have measured lifetimes for 47 spectra using time-resolved LIF. Our atomic beam source can produce a slow beam of neutral and singly ionized atoms of nearly any element. Lifetimes from 2 ns to ~2 µs can be measured for energy levels ranging from 15,000 to ~60,000 cm-1. In this review we will describe our method of measuring radiative lifetimes with an emphasis on possible errors and techniques used for controlling them. The electronic bandwidth, linearity, and overall fidelity of the fast photomultiplier, cable connections, and transient waveform digitizer are concerns. Possible errors from atomic collisions, radiation trapping, Zeeman quantum beats, hyperfine quantum beats, atoms/ions escaping from the observation region before radiating, and from radiative cascade through lower levels must be understood and controlled.We will then present a recent example of the application of our transition probability data to abundance determinations in the sun and in metal-poor halo stars (Den Hartog E A et al 2003 Astrophys. J. Suppl. 148 543). Our results have significantly reduced the uncertainty of abundance determinations for many elements including the rare-earths: La, Nd, Eu, Tb, Dy, Ho, Tm, and Lu. Better laboratory data have made it possible to cleanly separate the enhanced abundance of r(apid)-process neutron capture elements in metal poor Galactic halo stars. The long-term scientific payoff from this attention to detail in laboratory experiments is a better understanding of the r-process, of the Galactic chemical evolution, of the age of the oldest stars in the Galaxy, and indeed of the origins of the non-primordial chemical elements.