I have been asked to provide within-day and between day CV% for data produced using ELISA. While the negative and positive controls appear to perform as they should, the absorbance values for the standard curve are not precisely the same each day. The curve shape and calculated value of positive controls is similar. Just enough difference between days to produce lower estimates of precision? I am a bit frustrated since the assay is not run using an automated system and there is some variation in several steps (washing, time for pipetting, time of color development, lab temperature, etc.). I have been calculating CV% within days using 3 different positive control samples (replicates of each) and then using mean square between days to calculate precision (CV%)

It seems to me that using a bioassay, there is going to be a lot of variation between plates (even run on same day) at least the way I run them. I use same reagents for the most part. My buffers are home made but I make them up in large volumes to extend from first to final plate in my study. All plates are coated and blocked at same time again using large batches of coating antibodies and block.

Be critical as it may spur me to get more precise?