Pore volume compressibility is a fundamental driver of production for unconsolidated sand reservoirs. Prediction of compressibility is desirable when direct measurements on core are not available. Many characteristics of reservoir sands change simultaneously. For this reason, the controls on compressibility are difficult to isolate and interpret. We present the results of compaction experiments using laboratory-created, unconsolidated sands. In these analog sands, we change one textural or mineralogical parameter at a time to investigate the influence of that parameter on the measured compaction properties. Initially, simple quartz grain packs of varying grain sizes were used. Subsequently, additional parameters were investigated, including grain packing, angularity, sorting, feldspar content, ductile grain content, small volumes of dispersed clay, and initial sample preconditioning at stress. Multiple samples of each type were created and tested. This allowed the testing to be halted at several intermediate stresses and the samples to be examined using 2D and 3D imaging and image analysis techniques. For monomineralic quartz sand packs, grain size is a principal control on compressibility. As mean size increases from 150 to 450 μm, peak compressibility increases from 6 to 24 microsips. The depletion stress at which peak compressibility occurs decreases from 8,000 to 2,500 psi. Increasing grain angularity also increases compressibility but with smaller effect. For 150-μm quartz sands, increasing the angularity resulted in an increase in compressibility from 6 microsips for round quartz to 10 microsips for angular quartz and decreased the depletion stress required to achieve peak compressibility from 8,000 to 7,000 psi. As sorting varies from well to moderately poorly sorted, compressibility decreases, and the curve broadens as a function of depletion stress. Adding small volumes of feldspar (or other minerals that cleave) increases the compressibility more than the change resulting from changes in grain size, illustrating the importance of framework grain composition. Adding similar volumes of ductile grains results in a similar increase in compressibility to that observed for feldspar. However, when the size of the ductile grains is larger than that of the associated quartz (e.g., locally derived rip-up clasts), the increase in compressibility is significantly larger. To validate the experimental work, we compare the results of uniaxial pore volume compressibility tests on laboratory-created sands with measurements made on subsurface samples of similar texture and mineralogy. Both the shape of the compressibility curves as well as the magnitude of the compressibility are successfully reproduced. We conclude that laboratory-created sands can provide reasonable proxies for estimating the compressibility of subsurface reservoirs when intact subsurface samples are not available for measurement (e.g., only percussion sidewall samples are acquired) as long as mineralogy and texture are known.