A novel method to measure interlaminar shear strengths of composites is presented. This method uses the double notch shear (DNS) specimen and modifies the load introduction by use of a fixture. The main concept is to eliminate the compressive stress intensity at the notch roots, arising from the external load, by imposing a counteracting tensile load field there. This field is caused by a bending moment acting over the notched region. Proportional loading is achieved by holding the specimen in an inclined position with supports near its ends and behind the notches. The inclination angle determines the ratio between the compressive normal field and the tensile bending field. The required inclination angle depends on the notch distance, specimen thickness ratio, and to a lesser degree on any material anisotropy. Two test series, with four different notch distances, were carried out to compare the standard DNS test to the new method. The new method produced consistently higher strengths at all notch distances. For the shortest notch distance, which incidentally gives the highest strength values, the measured average interlaminar shear strength was almost doubled with the new method. As with the standard method, the new method still gives lower strength values with increasing notch distances and notable experimental scatter. Causes for this are discussed and some possibilities for improvements are outlined.