Three factors can significantly affect the measurement of air content in fresh concrete. Judgment is required when gauging foam levels and adding alcohol, which can result in small variations and widen the gap between the tests. Technique can also be a factor. For example, differences in agitating the volumetric test device may vary the extent to which all the entrained air is released.
The most significant factor, however, is likely to be the aggregate. The results of the pressure test, discussed by ASTM C 231, are much more sensitive to the effect of porous and lightweight aggregate than the volumetric test, also called the roll-a-meter and discussed by ASTM C 173. The volumetric test is preferred for testing mixes with porous or lightweight aggregate, but ASTM C 231 discusses how to determine the aggregate correction factor to adjust for this when using the pressure test. From the results you report it sounds like the aggregate might have some air in it, resulting in a higher reading with the pressure test.
Note also that the "Precision and Bias" section of ASTM C 231 puts the accuracy threshold at 0.8% when comparing results of two tests by different operators on the same material. Based on that, you are doing pretty well if you are consistently within 0.8% to 1%.