We study how the spectral fitting of galaxies, in terms of light fractions derived in one spectral region translates into another region, by using results from evolutionary synthesis models. In particular, we examine propagation dependencies on Evolutionary Population Synthesis (EPS, {\sc grasil}, {\sc galev}, Maraston and {\sc galaxev}) models, age, metallicity, and stellar evolution tracks over the near-UV---near infrared (NUV---NIR, 3500\AA\ to 2.5\mc) spectral region. Our main results are: as expected, young ($t \lesssim$ 400 Myr) stellar population fractions derived in the optical cannot be directly compared to those derived in the NIR, and vice versa. In contrast, intermediate to old age ($t \gtrsim$ 500 Myr) fractions are similar over the whole spectral region studied. The metallicity has a negligible effect on the propagation of the stellar population fractions derived from NUV --- NIR. The same applies to the different EPS models, but restricted to the range between 3800 \AA\ and 9000 \AA. However, a discrepancy between {\sc galev}/Maraston and {\sc grasil}/{\sc galaxev} models occurs in the NIR. Also, the initial mass function (IMF) is not important for the synthesis propagation. Compared to {\sc starlight} synthesis results, our propagation predictions agree at $\sim$95% confidence level in the optical, and $\sim$85% in the NIR. {\bf In summary, spectral fitting} performed in a restricted spectral range should not be directly propagated from the NIR to the UV/Optical, or vice versa. We provide equations and an on-line form ({\bf Pa}nchromatic {\bf A}veraged {\bf S}tellar {\bf P}opulation - \paasp) to be used for this purpose.