Quartz inclusions in garnet are suitable for barometry because quartz is highly compressible relative to garnet, and the garnet host can maintain large stress differences generated as pressure-temperature conditions change. However, experimental validation of the quartz-in-garnet approach has been limited, raising questions concerning the accuracy of calculated entrapment pressure. Here we test the results of quartz-in-garnet barometry by conducting in situ Raman analysis of natural inclusions over a range of temperatures (−175 to 600 °C). We assess the temperature (T) dependence of inclusion pressure (Pincl) at 1 bar and compare calculated entrapment pressures derived from measurements at different temperatures. Experiments used two quartz standards (oriented [c] and <a>) and fully encapsulated quartz inclusions from three different terranes, retaining different Pincl at room T (−444, 296, and 755 MPa). The stretched quartz inclusion (Pincl < 0) had the greatest increase in Pincl (+264 MPa, during heating from 25 to 500 °C), whereas the high-P inclusions underwent less change (+138 MPa) in Pincl over the same T interval. The greater T sensitivity of inclusions with low Pincl reflects the greater thermal expansivity of quartz near the α to β quartz transition. While measured Pincl-T trends are consistent with predictions, numerical models tend to overestimate Pincl at elevated T, and calculated entrapment pressures show an unrealistic dependence on reference T. Raman spectroscopic measurements conducted in situ at elevated T provide optimal results. In addition, we have recalibrated the thermal portion of the numerical method based on the present results, and provide new empirical expressions for improved quartz-in-garnet barometry.