There's meter spectral response, human eye spectral response, film spectral response and the spectral distribution of light and the "color" of the test target. Lots of variables.
Many people inadvertently obtain a "Tungsten" speed rating by their tests without realizing it.
Others put an 80b filter over the light or lens to (at least partially) simulate daylight.
More important in my mind is to stick with a test plan where you understand it's limitations, and strive for consistency. Yes it's significant in your case, but if your test-to-test result stayed within 2/3 stop (or 2/3 N step), then you can use your test to control your processes.