Because the house runs differently. A sealed, heated winter house pulls soil gas in hard; an open-window summer house dilutes and draws less. Both results are real. Decisions should lean on heating-season numbers or a long-term average.
More detail
Neither test lied; they sampled different machines. From October to April a metro house is a chimney: furnace running, windows locked, warm air exfiltrating high and soil gas making up the difference low. In July the same house breathes through open windows and evaporative coolers, and the indoor-outdoor pressure difference nearly vanishes. Swings of two or three to one between seasons are ordinary, which has two practical consequences. A summer result in the 2-to-4 band is a yellow flag, not a pass, because the winter number is probably higher; and a mitigation decision made on a single July test underestimates the exposure the family actually averages. The clean answers are a heating-season short test or, better for borderline cases, an alpha-track device spanning the winter. The winter testing plan gives the calendar, and seasonal timing the reasoning.