Some of the values fall in the ODs of around 0.009 and are close to the leveled lower limits of my standards. this for me i would want to set it as negative,. is there a statistical way to set cutoff to be only within the linear points of the standard curve?
You would have more faith in the values that are on the straight part of the curve but if there is a lower limit of detection supplied with your ELISA kit you can use this but this needs to be done consistently.
When in doubt, plot the data yourself. Look particularly at the spread of the data for each point. 0.009 as an absolute OD is not very much. What is the spread of the replicates of that test point, versus the spread of the reference points on either side of it?