Feasibility of directly measuring household energy use

This experiment assesses the feasibility of investigating certain key environmental behaviours and outcomes directly. Specifically, IP respondents were asked to report on two occasions meter readings for gas and electricity, and odometer readings for the motor vehicle used most often by the household. The experiment examines whether providing respondents with advance notice of the required data encourages the provision of the data within the interview. Collecting the readings on two occasions would allow the first reading to be subtracted from the second, thereby directly measuring vehicle and energy use over an identifiable period of time. Errors in implementing the experiment at Wave 5 meant that data from the experiment were not released. Instead, the experiment was re-implemented Wave 6.

The experiment takes a 2×2 design:

Table: Design of household energy use experiment

	Gas, electric and odometer reading	Odometer reading alone
Notification of meter readings in advance letter	Group 1	Group 2
No notification	Group 3	Group 4

Randomisation was of households within PSUs into one of the four groups for the experiment.

Approximately 4 weeks after the standard fieldwork period, a follow-up postal questionnaire collected the matching information gathered in the main Wave 6 interview from households providing at least one meter reading (N=825). The household questionnaire respondent was notified during the interview of this subsequent data collection and questionnaires were sent to this named individual. To be able to compare readings from two occasions, respondents to the follow-up confirmed their address and the make and mode of the vehicle about which they reported in the main interview.

After two-weeks, any non-response was followed-up with a reminder and a telephone follow-up was used after a further 10 days of non-response to collect information through the alternative mode.

Controlling variable if f_ff_metersw6 on record f_hhsamp_ip, and takes the values:

1          Gas, Electric, Odometer, advanced letter warning

2          Odometer only, advanced letter warning

3          Gas, Electric, Odometer, no advanced letter warning

4          Odometer only, no advanced letter warning

Substantive variables containing initial meter readings are f_gasuse, f_gasmeter, f_gasest, f_elecuse, f_elecmeter, f_varmeter1, f_varmeter2, f_elecest, f_meterfol, f_odouse, f_odometer, f_odoest, f_odofol on the record f_hhresp_ip.

Follow-up meter reading data are contained in variables f_intlen1, f_modetype, f_remind, f_ff_metersw6, f_qutype, f_mrfaddchck, f_mrfuelhave1, f_mrfuelhave2, f_mrfuelhave3, f_mrfuelhave4, f_mrfuelhave96, f_mrgasuse, f_mrgasmeter, f_mrgasest, f_mrelecuse, f_mrelecmeter, f_mrvarmeter1, f_mrvarmeter2, f_mrelecest, f_mrvehchck, f_mrodouse, f_mrodometer, f_mrodoest, f_readdated, f_readdatem, f_readdatey on record f_meterreading_ip.