We study the evolution with redshift of three measures of gamma-ray burst (GRB) duration (T₉₀, T₅₀ and TR₄₅) in a fixed rest-frame energy band for a sample of 232 Swift/Burst Alert Telescope (BAT) detected GRBs. Binning the data in redshift we demonstrate a trend of increasing duration with increasing redshift that can be modelled with a power law for all three measures. Comparing redshift defined subsets of rest-frame duration reveals that the observed distributions of these durations are broadly consistent with cosmological time dilation. To ascertain if this is an instrumental effect, a similar analysis of Fermi/Gamma-ray Burst Monitor data for the 57 bursts detected by both instruments is conducted, but inconclusive due to small number statistics. We then investigate underpopulated regions of the duration redshift parameter space. We propose that the lack of low-redshift, long duration GRBs is a physical effect due to the sample being volume limited at such redshifts. However, we also find that the high-redshift, short duration region of parameter space suffers from censorship as any Swift GRB sample is fundamentally defined by trigger criteria determined in the observer frame energy band of Swift/BAT. As a result, we find that the significance of any evidence for cosmological time dilation in our sample of duration measures typically reduces to <2σ.

We present a template-fitting algorithm for determining photometric redshifts, z phot, of candidate high-redshift gamma-ray bursts (GRBs). Using afterglow photometry, obtained by the Reionization and Transients InfraRed (RATIR) camera, this algorithm accounts for the intrinsic GRB afterglow spectral energy distribution, host dust extinction, and the effect of neutral hydrogen (local and cosmological) along the line of sight. We present the results obtained by this algorithm and the RATIR photometry of GRB 130606A, finding a range of best-fit solutions, 5.6 < z _phot < 6.0, for models of several host dust extinction laws (none, the Milky Way, Large Magellanic Clouds, and Small Magellanic Clouds), consistent with spectroscopic measurements of the redshift of this GRB. Using simulated RATIR photometry, we find that our algorithm provides precise measures of z_phot in the ranges of 4 < z_phot lesssim 8 and 9 < z_phot < 10 and can robustly determine when z _phot > 4. Further testing highlights the required caution in cases of highly dust-extincted host galaxies. These tests also show that our algorithm does not erroneously find z _phot < 4 when z _sim > 4, thereby minimizing false negatives and allowing us to rapidly identify all potential high-redshift events.