An alternative is to use PRISM and so calculate the amplification efficiency ± CI. As Mikael points out, this can be quite an eye-opener. For one of our RNA-based dilution curves what was an acceptable slope of -3.389 turns out to actually range from -3.990 to -2.788, which of course needs to be accounted for when calculating copy numbers. And if you do a relative quantification, then you must make sure you carry over your errors appropriately.