Photometric detections of dust circumstellar discs around pre-main sequence (PMS) stars, coupled with estimates of stellar ages, provide constraints on the time available for planet formation. Most previous studies on disc longevity, starting with Haisch, Lada & Lada, use star samples from PMS clusters but do not consider data sets with homogeneous photometric sensitivities and/or ages placed on a uniform time-scale. Here we conduct the largest study to date of the longevity of inner dust discs using X-ray and 1-8um infrared photometry from the MYStIX and SFiNCs projects for 69 young clusters in 32 nearby star-forming regions with ages t<=5Myr. Cluster ages are derived by combining the empirical Age_JX_ method with PMS evolutionary models, which treat dynamo-generated magnetic fields in different ways. Leveraging X-ray data to identify disc-free objects, we impose similar stellar mass sensitivity limits for disc-bearing and disc-free young stellar objects while extending the analysis to stellar masses as low as M~0.1M_{sun}_. We find that the disc longevity estimates are strongly affected by the choice of PMS evolutionary model. Assuming a disc fraction of 100 per cent at zero age, the inferred disc half-life changes significantly, from t_1/2_~1.3-2Myr to t_1/2_~3.5Myr when switching from non-magnetic to magnetic PMS models. In addition, we find no statistically significant evidence that disc fraction varies with stellar mass within the first few Myr of life for stars with masses <2M_{sun}_, but our samples may not be complete for more massive stars. The effects of initial disc fraction and star-forming environment are also explored.