Ideally, they should all be the same. The fact of the matter is that we canÆt measure the properties of materials in ideal states. Because of the way we have to test materials, not everything can be held constant --- things change --- and these changes, many of which occur during the test, alter the outcome of the test.
A case in point, flexural modulus is generally measured in a 3-point bend test. A rectangular specimen is supported horizontally by two steel pins and the plastic bar is loaded at the midpoint of the two supports.
One limitation with flex test is that it is not a ôpureö stressed state. The stress is calculated as the maximum ôfiberö stress that occurs directly under the load on the underside surface of the bar --- this is the only point at which that maximum fiber stress exists. Actually, the stress distribution through the bar varies from tensile on the underside surface of the specimen to compressive stress on the topside surface. The compressive stress tends to inhibit the deflection of the specimen, artificially raising the apparent modulus of the material.
Also, as the specimen deflects, the bar must move along the supports to accommodate the deflection. If the calculated fiber strain on the underside surface exceeds about 5%, a significant portion of the constant load is consumed as the driving force to ôpushö the specimen through the supports --- rather than merely bend the specimen.
In addition, at some strain level, probably around 5%, the actual strain on the bar at the point of loading stops increasing --- no further curvature occurs in the central area of the specimen --- but the specimen continues bend toward the ends of the bar as it slides through the supports. Inasmuch as the apparent fiber strain is calculated based on the amount of deflection from the original horizontal position, the method begins to yield erroneous data.
The latter two points most likely would not affect the reported modulus, but should be considered in any other data reported for the test.