Four methods to determine composite interlaminar shear strength (ILSS) are evaluated. In particular, the recently devised inclined double-notch shear test (IDNS) is compared with three existing and more established methods: the Iosipescu test, the short three-point bending test (S3PB) and the double-notch compression tesl (DNC). The uniformity of strain field in the test region in a real test situation - which is the crucial test method quality parameter - is investigated by strain mapping using digital speckle photography. The measured strain fields are compared with FE-calculated strains representing ideal conditions and both known advantages and drawbacks of the different methods are confirmed. The IDNS test produces the most uniform strain fields and also consistently high ILSS values. A fractographic analysis indicates shear separation over a major part of the fracture surfaces of all specimen types; typical shear cusps were found over about 80% of the IDNS fracture surface and in about 50% to 70% in the other specimens. For the Iosipescu tests, failure initiation could be ascribed to initiation in tension at defects. Experimentally determined stress-strain responses in shear exhibit a distinct variation among the different methods. For the best methods, a notable material softening was observed prior to failure. Observed formation of shear cusps is believed to be the primary cause for this softening of the composite material studied here.