We have axial stress and strain data that we recorded for each cycle during the fatigue tests. We are trying to find out the Young's modulus of the loading and unloading part of each hysteresis curve. Visco-elastic effect of the materials provides us very short linear section which makes us difficult to calculate the slope of the loading and unloading section. We are using 95% linearity (as per ASTM standard) of the loading and unloading section, but we are getting a huge scatter in the data. I am wondering to know if there is any other criteria we can follow to calculate the Young's modulus. A sample data set and the schematic representation of the method we are using are attached.