I conducted tensile test on glass fiber reinforced polyester composites wherein I found that the failure mainly occurs at the grip area (not within grip) and not at the gage area.
Hence I seek input from researchers on this issue.
The same problem faced by us as well. We works on natural fiber reinforced polymer composites. The samples break at grip area for us as well, due to the heavy pressure applied because of jaws on the composite sample. I think the grip area becomes weak due to the application of heavy load from jaws it breaks at that position.
The reason I mentioned above is based on my experience on type of failure. It may also be due to the misalignment of the grips (not axially). Anyhow have a look into the article which may give you some idea but not exactly.
Legerlotz, K., Riley, G. P., & Screen, H. R. (2010). Specimen dimensions influence the measurement of material properties in tendon fascicles. Journal of biomechanics, 43(12), 2274-2280.
If the local compression under the grips causes the local cross sectional area to be smaller than the tensile section then failure will occur at the grips, this can happen in metals as well as composites. Additionally localized constriction at the end of the grips can cause stress concentrations that promote failure in this area. Modifications to the grips can prevent premature failure
Increase the thickness of the gripping area or use additional piece of laminate on the gripping area if you are doing experiment with ASTM D3039/D3039M (see figure tensile specimen drawing). In other case, dog bone shape reduces the area in the middle so that the specimen is forced to fail in the middle. I hope you fine this useful.
If end tabs do not help, width tapering could. The use of a 'dogbone' shaped geometry will ensure a uniform stress in the central gauge section of the material. The values you have gotten could be conservative.