Objective: To investigate explanatory factors for persistent cold temperatures in homes which have received heating improvements. Design: Analysis of data from a national survey of dwellings and households (in England occupied by low-income residents) that had received heating improvements or repairs under the Warm Front Scheme. Methods: Over the winters of 2001-02 and 2002-03, householders recorded living room and main bedroom temperatures in a diary. Entries were examined for 888 households, which had received high level heating interventions. Two hundred and twenty-two households were identified as occupying cold homes, with mean bedroom temperature below 16 degrees C or mean living room temperatures below 18 degrees C. Binary logistic regression was used to model dwelling and household features and then occupants' behaviour and attitudes in the 'cold homes' sub-set compared with the remainder of the high intervention group. Seventy-nine supplementary, structured telephone interviews explored reasons given for lower temperatures. Using graphical and tabular methods, householders preferring cooler homes were distinguished from those who felt constrained in some way. Results: Cold homes predominate in pre-1930 properties where the householder remains dissatisfied with the heating system despite major improvements funded by Warm Front. Residents of cold homes are less likely to have long-standing illness or disability, but more likely to experience anxiety or depression. A small sample of telephone interviews reveals those preferring lower temperatures for health or other reasons, report less anxiety and depression than those with limited control over their home environment. Their 'thermal resistance' to higher temperatures challenges orthodox definitions of comfort and fuel poverty. (c) 2006 Elsevier Ltd. All rights reserved.