Table 1 Statistical models and variables used for hypotheses testing.

A. Thermostat settings are lower during the CoL than before the CoL.

Repeated measures t-test, two-sided

Thermostat setting (A1 CoL)

Thermostat setting (A5 BS)

In A1 CoL or A5 BS Do not know chosen

Values below 5 °C and above 30 °C

B. Likelihood to turn heating off when home is unoccupied for a day or more is higher during the CoL than before the Col.

Repeated measures t-test, two-sided

Turning heating off (A10 BS)

Turning heating off (A7 CoL)

In A10 BS or A7 CoL not applicable or cannot do this chosen

C. The effort to save energy is higher during the CoL than before the CoL.s

Repeated measures t-test, two-sided

Energy saving effort (A14 BS)

Energy saving effort (A12 CoL)

In A14 BS or A12 CoL Do not know chosen

D. Those using their IHD more during the CoL report greater energy saving effort than those using it less.

One-way Anova with posthoc Bonferroni correction

Outcome: Energy saving effort (A12 CoL)

Predictors: IHD usage (A11 CoL) coded as More often, less often, the same

In A11 CoL I do not have this, Prefer not to say, It is not working chosen

In A12 CoL Do not know chosen

E. Those are doing less well financially (a) report greater effort at saving energy and (b) perform energy saving actions more often.

Linear regression

Outcome a: energy saving effort A12 (CoL)

Outcome b: mean of frequency of energy saving actions (A6, CoL)

Predictors: financial status E1 (CoL)

Confounding variables

-tenure (B4 BS)

-anyone ≤ 16 years in the home (D2 CoL)

-anyone > 85 in the home (D2 CoL)

-anyone not working bc of disability (D3 CoL)

-anyone retired (D3 CoL)

-working from home (D4 CoL)

-household size (D1 CoL)

-number of bedrooms (B6 BS)

-number of bathrooms (B1 CoL)

-building age (B9 BS)

Building type (B1 BS)

In A12 CoL do not know chosen. I In A6 Not applicable, cannot do this chosen. In E1 CoL do not know or prefer not to say chosen

F. Those who find it difficult to meet their heating cost have lower wellbeing.

Spearman rank-order correlation

Heating costs (C5 CoL)

(a) Life satisfaction (E2 CoL); (b) worthwhile (E3 Col)

In C5 CoL do not know chosen

G. Those finding it difficult to keep warm have (a) lower life satisfaction & (b) find things less worthwhile.

Independent samples t-test, two-sided

Outcome: (a) life satisfaction (E2 CoL) and (b) worthwhile (E3 Col)

Grouping: keeping warm (C3 CoL) coded as yes, no

In C3 CoL do not know chosen