Examines calculation methods and insulation techniques applied to a construction project that comprises 1,200 two-storey housing units built in New York City. The R-value is calculated using several methods. Two-dimensional finite-difference results for a stud-and-tracks assembley are first fine-tuned against a three-dimensional model. The two-dimensional computer program is then used to analyse the exact configuration of the steel frame construction of the entire building accounting for all double studs, cripple studs, jambs, lintels, and beams, as well as metal ties and fasteners. The results emphasise the need for a whole-wall R-value calculation that accounts for all framing in the wall, rather than a clear-wall approximation. The error of the clear-wall calculation is in the 40 to 50% range for this type of house. Thermally improved wall constructions are examined. For the same centre-of-cavity R-value, a wall with insulating sheathing has a total R-value that is 13% higher than one without insulating sheathing. The risks of moisture condensation is significantly reduced with the improvements.KEYWORDS: year 1997, Calculating, thermal insulation, walls, steel, flats, housing, thermal resistance, prefabricated construction, moisture, condensation, relative humidity, comparing