I investigate the origin of the observed correlation between a gamma-ray burst's ${\nu}F_{\nu}$ spectral peak $E_{\rm pk}$ and its isotropic equivalent energy $E_{\rm iso}$ through the use of a population synthesis code to model the prompt gamma-ray emission from GRBs. By using prescriptions for the distribution of prompt spectral parameters as well as the population's luminosity function and co-moving rate density, I generate a simulated population of GRBs and examine how bursts of varying spectral properties and redshift would appear to a gamma-ray detector here on Earth. I find that a strong observed correlation can be produced between the source frame $E_{\rm pk}$ and $E_{\rm iso}$ for the detected population despite the existence of only a weak and broad correlation in the original simulated population. The energy dependance of a gamma-ray detector's flux-limited detection threshold acts to produce a correlation between the source frame $E_{\rm pk}$ and $E_{\rm iso}$ for low luminosity GRBs, producing the left boundary of the observed correlation. Conversely, very luminous GRBs are found at higher redshifts than their low luminosity counterparts due to the standard Malquest bias, causing bursts in the low $E_{\rm pk}$, high $E_{\rm iso}$ regime to go undetected because their $E_{\rm pk}$ values would be redshifted to energies at which most gamma-ray detectors become less sensitive. I argue that it is this previously unexamined effect which produces the right boundary of the observed correlation. Therefore, the origin of the observed correlation is a complex combin Show Partial Abstract