PURPOSE The Alfonso et al. [Med. Phys. 35, 5179-5186 (2008)] formalism for small field dosimetry proposes a set of correction factors (kQclin,Qmsrfclin,fmsr) which account for differences between the detector response in nonstandard (clinical) and machine-specific-reference fields. In this study, the Monte Carlo method was used to investigate the viability of such small field correction factors for four different detectors irradiated under a variety of conditions. Because kQclin,Qmsrfclin,fmsr values for single detector position measurements are influenced by several factors, a new theoretical formalism for integrated-detector-position [dose area product (DAP)] measurements is also presented and was tested using Monte Carlo simulations. METHODS A BEAMnrc linac model was built and validated for a Varian Clinac iX accelerator. Using the egs++ geometry package, detailed virtual models were built for four different detectors: a PTW 60012 unshielded diode, a PTW 60003 Diamond detector, a PTW 31006 PinPoint (ionization chamber), and a PTW 31018 MicroLion (liquid-filled ionization chamber). The egs_chamber code was used to investigate the variation of kQclin,Qmsrfclin,fmsr with detector type, detector construction, field size, off-axis position, and the azimuthal angle between the detector and beam axis. Simulations were also used to consider the DAP obtained by each detector: virtual detectors and water voxels were scanned through high resolution grids of positions extending far beyond the boundaries of the fields under consideration. RESULTS For each detector, the correction factor (kQclin,Qmsrfclin,fmsr) was shown to depend strongly on detector off-axis position and detector azimuthal angle in addition to field size. In line with previous studies, substantial interdetector variation was also observed. However, it was demonstrated that by considering DAPs rather than single-detector-position dose measurements the high level of interdetector variation could be eliminated. Under small field conditions, mass density was found to be the principal determinant of water equivalence. Additionally, the mass densities of components outside the sensitive volumes were found to influence the detector response. CONCLUSIONS kQclin,Qmsrfclin,fmsr values for existing detector designs depend on a host of variables and their calculation typically relies on the use of time-intensive Monte Carlo methods. Future moves toward density-compensated detector designs or DAP based protocols may simplify the methodology of small field dosimetry.