Abstract

Hubble data has given us the first estimates of the intrinsic scatter of the main sequence of star-forming galaxies at redshifts >~3.5 (Salmon+15). In fact, Santini+17 find tentative evidence for increasing scatter to low stellar masses, which may indicate more stochastic star formation. However, significant broad-band flux contamination by emission lines, as well as sparse sampling of galaxy SEDs makes physical parameter estimation uncertain, and robust treatment of the uncertainties is key. Using Bayesian Hierarchical modelling teamed with the sophisticated nebular modelling of Gutkin+16 and SED -fitting using BEAGLE (Chevallard+16), I test the recovery of main-sequence parameters under standard assumptions used to analyse high redshift data, and discuss the implications for future constraints possible with JWST .