The coefficients generated for independent variables (IVs) in a discriminant analysis are simply weights that help create a single score (linear composite) that maximally separates group means (group membership is the dependent variable). The sign of a coefficient will depend in part on how the groups are coded by the statistical software you have used, as well as on the sift of IVs you have included and how they are quantified.
A negative coefficient would be interpreted as indicating that, when the other IVs are held constant, and increase in the IV of interest would mean that the discriminant function score for a case is predicted to decrease. Therefore, the case would be somewhat less like a higher scored group (on the discriminant function scores) and more like a lower scored group). A decrease in the IV of interest would suggest that the case is more like a higher scored group and less like a lower scored group member.