Calculated R-values for a variety of timber-framed, concrete masonry, and panel constructions have been compared to measured values, using results from 10 years of laboratory testing for industry and from field surveys of nearly 100 occupied houses. From this data it is concluded that R-volues can be realistically predicted by simple formulae with care and sound judgment.

The principal factors causing R-values to vary from ideal are argued to be the well-known and widely ignored trio of thermal bridging, gaps in insulant, and moisture (moisture effects are not examined here). Thermal bridging is shown to reduce realized R-values by up to 50% in common cases, even with timber framing, and the reduction becomes even more severe with increasing R-value. In many cases this loss can be largely avoided by using thermal breaks.

A recent trend towards wider acceptance of "zero lateral conductance" calculations is deprecated, and the alternative "infinite lateral conductance" is instead advocated for engineering purposes.

Units: SI