Degree day indices are a temperature index used for understanding impact of climate change. However, biases in climate models challenge our ability to make projections. Different methods to deal with these biases, termed bias correction or more generally calibration, yield different answers. This is not widely understood for temperature indices in many impact sectors. An analytical expression is derived for the expected value of degree days given parameters of the underlying statistical distribution. It is demonstrated that the uncertainty introduced by calibration methodology is driven by the magnitude of the nonlinearity in this expression. In a climate with mean temperature far from (approximately 3 standard deviations) the threshold used in definition of the index, the equation is approximately linear and methodological choice makes little difference. Case studies for London and Glasgow, two UK cities, for heating and cooling degree days (HDD and CDD) demonstrate the unrealistic results that may be produced if appropriate calibration is not performed. It can therefore be concluded that such inappropriate methods introduce error. Seasonally varying temperature biases in the 11-member perturbed parameter ensemble HadRM3-pertrubed physics ensemble are discussed, and different calibration strategies are applied to this ensemble. Projections of HDD over the UK are largely insensitive to method, due to the expression for HDD being approximately linear in many months and locations. In contrast, uncertainty in CDD projections is dominated by choice of method, due to nonlinearity of the CDD expression.