Recent results indicate the stellar initial mass function is not a strong function of star-forming environment or ``initial conditions (e.g. Meyer et al. 2000). Some studies suggest that a universal IMF may extend to sub-stellar masses (see however Briceno et al. 2002). Yet most of this work is confined to star-forming environments within 1 kpc of the Sun. In order to probe the universality of the IMF over a wider range of parameter space (metalicity, ambient pressure, magnetic field strength) new techniques are required. We begin by summarizing our approach to deriving the sub-stellar IMF down to the opacity-limit for fragmentation using NGC 1333 as an example. Next, we describe results from simulations using the observed point-spread function of the new 6.5m MMT adaptive optics system and examine the confusion-limited sensitivity to low mass stars in rich star-forming clusters out to 0.5 Mpc. We also present preliminary results from observations with this system of the W51 star-forming complex. Finally, we outline a new technique to estimate the ratio of high to low mass stars in unresolved stellar populations, such as the massive star clusters observed in interacting galaxies (e.g. Mengel et al. 2002). While evidence for variations in the IMF remains inconclusive, new studies are required to rule them out and determine whether or not the IMF is universal over the range of parameter space relevant to star-forming galaxies over cosmic time.