Calculating the strain in three different ways - experimental error?

[Cottontails]

I've recently done an experiment where I've obtained the strain of a channel section beam in bending. I used a strain gauge to get the strain from the midspan and also used dial gauges to get the deflections of the beam under different loading conditions.

Along with the readings I've obtained from the strain gauge, I have also used the deflections to calculate the strain in two other ways. I've done this through the deflections (curvature) and the experimental stiffness. The strain for deflections and experimental stiffness are virtually alike (highest error is 10%) however, in comparison to the strain gauge readings, there is a greater error (highest error ir around 30%).

I ensured that I performed the experiment correctly and given my raw results, they do seem correct. So, I'm wondering why do the experimental and theoretical results vary so much?

 

[Aero_UoP]

uhm, I can't put my finger on it but have you calculated the strain in the same direction in which the strain gauge is oriented?