A set of randomised assignments related to payment of respondent incentives extends across waves 1-6 of the IP. At Wave 6, experimental allocation was nested within mixed mode treatments (See Mixed mode experiments). Households within primary sampling units (PSUs) were randomly allocated to treatments within PSUs. All enumerated adults within a household received the same incentive offer.
At Wave 1, each household received an initial unconditional £5 incentive sent with the advance letter, which was “topped-up” with the total incentive amount for the household after the interview. It contrasts lower and higher payments, uniform for the household, with a higher incentive if all eligible enumerated adults in the household were interviewed. The experimental treatment groups were as follows:
Group 1 – £5 per interviewed household member
Group 2 – £10 per interviewed household member
Group 3 – £5 per interviewed household member increasing to £10 per person if all eligible enumerated adults in the household were also interviewed
The variable in the data that controls allocation at the household level is a_groupincentive on the record a_hhsamp_ip.
At Wave 2, the incentive experiment had some groups where the payment amount was the same as at Wave 1 and some where the payment level was reduced. Incentives were sent in advance of fieldwork to named individuals through a postal mailing. Rising 16 year olds newly eligible for an adult interview at Wave 2 received an advance mailing with the appropriate adult incentive. The complete co-operation top-up (in Group 5) was posted to households once the final eligible person was interviewed. Within PSUs, households were randomly allocated to each of the 5 treatment groups. All adults within each household received the same treatment:
Group 1 – Receive £5 (as per IP1)
Group 2 – Receive £10 (as at IP1)
Group 3 – Receive £5 (reduction from £10 at IP1 to £5 at IP2)
Group 4 – Receive £5 rising to £10 if complete household co-operation (as per IP1)
Group 5 – Receive £5 (reduction from possible £10 at IP1 to £5)
The controlling variable is b_ff_incentw2 on the record b_hhsamp_ip.
In Wave 3, five of the six treatment groups retained the same treatment as at Wave 2. One of the two groups receiving £10 at both Wave 1 and Wave 2 had the incentive decreased to £5, to test the effect of a decrease at an early stage in the panel. All incentives were sent in advance of fieldwork to named individuals through a postal mailing. Rising 16-year-olds newly eligible for an adult interview at Wave 3 received an advance mailing with the appropriate adult incentive. The complete co-operation top-up was posted to households once the final eligible person was interviewed. Households within PSUs were allocated to treatments. There were 6 experimental groups and all adults within each household received the same treatment:
Group 1 – £5 (same at IP1 & IP2)
Group 2 – £10 (same at IP1 & IP2)
Group 3 – £5 (was £10 at IP1 & IP2)
Group 4 – £5 (was £10 at IP1)
Group 5 – £5 to £10 for complete cooperation (same at IP1 & IP2)
Group 6 – £5 (was £5 to £10 at IP1)
The controlling variable is c_ff_incentw3 on the record c_hhsamp_ip.
At Wave 4, some of the continuing sample households received an increase in incentive, from £5 to £10 (See group 2 and group 6 below.) Over the history of their participation, group 2 always received £5 in the past, whereas group 6 had started off with £10. All other groups for continuing households remained as at wave 3. The refreshment sample (groups 9, 10, and 11) received larger amounts than those traditional for the UK. See Section 4 about the refreshment sample.
Group 1 – £5 (same at IP1, IP2, IP3)
Group 2 – £10 (was £5 at IP1, IP2, IP3)
Group 3 – £10 (same at IP1, IP2, IP3)
Group 4 – £5 (was £10 at IP1, IP2)
Group 5 – £5 (was £10 at IP1)
Group 6 – £10 (was £10 at IP1 and £5 at IP2 & IP3)
Group 7 –£5 to £10 for complete cooperation (same at IP1, IP2, IP3)
Group 8 –£5 (was £5 to £10 at IP1)
Group 9 –£10 per interviewed household member
Group 10 –£20 per interviewed household member
Group 11 –£30 per interviewed household member
At Wave 4, the controlling variable is d_ff_incentw4 on the record d_hhsamp_ip.
At Wave 5, the Wave 4 conditions were repeated with one exception. Respondents in the experimental treatment receiving £5 unconditionally, rising to £10 if all household members completed the interview, were randomly allocation to receive either a £5 or a £10 unconditional incentive. The result is that at Wave 5, 7/12 of the sample received a £5 incentive and 5/12 received a £10 incentive. All refreshment sample experimental allocations from Wave 4 were repeated at Wave 5.
Group 1 – £5 (same at IP1, IP2, IP3, IP4)
Group 2 – £10 (was £5 at IP1, IP2, IP3, IP4)
Group 3 – £10 (same at IP1, IP2, IP3, IP4)
Group 4 – £5 (was £10 at IP1, IP2)
Group 5 – £5 (was £10 at IP1)
Group 6 – £10 (was £10 at IP1 and £5 at IP2, IP3, IP4)
Group 7 –£5 (was £5 rising to £10 for complete cooperation at IP1, IP2, IP3, IP4)
Group 8 – £10 (was £5 rising to £10 for complete cooperation at IP1, IP2, IP3, IP4)
Group 9 – £5 (was £5 to £10 at IP1)
Group 10 –£10 per interviewed household member
Group 11 –£20 per interviewed household member
Group 12 –£30 per interviewed household member
At Wave 5, the controlling variable is e_ff_incentw5 on the record e_hhsamp_ip.
In addition, Wave 5 had a mixed mode survey interviewing experiment that examined whether incentives can be used to maximize take-up of the web survey. The design had two factors. The first factor was whether the day on which the email invitation to the web survey is sent. The second factor is whether conditional incentives – offered in addition to the existing unconditional incentives – can increase the take-up of the web survey.
Respondents were allocated to either receive the first email invitation for web interviewing on a Friday or a Monday. This was crossed with an incentive experiment where households were randomly allocated to receive an additional conditional web bonus or no bonus. In the web bonus group, the household reference person was offered a conditional bonus if they complete their part of the survey (household grid, household questionnaire and their individual questionnaire) within 3 days. Once the household grid was completed, all other household members were offered a conditional bonus for completing the web survey. In the control group all household members received the unconditional incentive only.
Group 1 — Monday, £5 bonus
Group 2 — Monday, No bonus
Group 3 — Friday, £5 bonus
Group 4 — Friday, No bonus
The controlling variable is e_ff_invitew5 on record e_hhsamp_ip.
Certain individuals not interviewed on the first visit to the household if face-to-face were approached to do their individual interview either online or face-to-face. If online, then the offer of a bonus to go online was controlled with experimental treatments 1 & 3 versus 2 & 4 on this item.
At Wave 6, experimentation with incentives was nested within mixed mode interviewing treatment (See Mixed Mode experiments: web and face-to-face). As with previous waves, allocation to experimental treatment groups was at the household level: all individuals within the household received the same experimental treatment and any split-off households retained experimental allocation from the previous household at both the current and all previous waves.
The one-third of households allocated to a face-to-face interviewing approach at Wave 6 were all allocated to a £10 unconditional incentive. For the five of the nine Wave 5 incentive groups in the original sample, this represented an increase from £5. For two of the Wave 5 refreshment sample incentive groups this was a decrease from £20 or £30.
The two-thirds of households allocated to a web interviewing approach at Wave 6 were allocated in equal proportions to three incentive experimental treatments. These were crossed with the Wave 5 incentive treatments. All treatments received an unconditional incentive with differences across them in the amount (£10 or £30) and whether an additional conditional incentive was offered for whole household completion within a two-week web-only interviewing period (See Mixed Mode Experiments). The groups are characterised as follows:
Group 1 – £10 unconditional incentive
Group 2 – £10 unconditional with a £20 conditional incentive for full-household completion by web in the allotted time
Group 3 – £30 unconditional incentive
The controlling variable is f_ff_incentw6 on record f_hhsamp_ip.
There were four households in the £10 treatment group who became aware of the £30 treatment group and were sent an extra £20. The variable f_incentcomp on the record f_hhsamp identifies these cases so that they can be excluded from analyses.
Tables 7.1 and 7.2 document the allocation of cases to experimental groups within mixed mode treatments. In order to reflect all historical incentive treatments, the controlling variable takes 48 unique values. However, only 3 different incentive amounts were used at Wave 6:
Groups 1-12, 13, 16, 19, 22, 25, 28, 31, 34, 37, 40, 43, 46 = £10 unconditional incentive
Groups 14, 17, 20, 23, 26, 29, 32, 35, 38, 41, 44, 47 = £10 unconditional + £20 conditioned on whole household web completion
Groups 15, 18, 21, 24, 27, 30, 33, 36, 39, 42, 45, 48 = £30 unconditional incentive
At Wave 7, the incentive experiment paralleled the experimental allocation carried at IP6 with a repeat of the elevated incentive amounts for the IP7 refreshment sample comparable to the IP4 incentive experiment the same purpose. Incentive allocation was nested within the mixed mode allocation repeated through from IP5. Households within IP1 (or IP4) PSUs were allocated the treatments such that all individuals within households received the same incentive, and all split-off households received the same incentive as their originating household. The continuing IP sample was allocated to the identical treatments as at IP6:
Face to Face sample (i.e., ff_gridmodew5 = 1) = £10
Web allocated (i.e., ff_gridmodew5 = 3) to three groups:
Group 1 – £10 unconditional incentive
Group 2 – £10 unconditional with a £20 conditional incentive for full-household completion by web in the allotted time
Group 3 – £30 unconditional incentive
The IP7 refreshment sample was allocated to the face-to-face experimental treatment (i.e., ff_gridmodew5 = 3), then divided evenly between three incentive experimental groups where each individual in the household received the same incentive amount:
Group 1 – £10 unconditional incentive
Group 2 – £20 unconditional incentive
Group 3 – £30 unconditional incentive
The controlling variable is ff_incentw7 on record hhsamp. The first 48 unique values were carried over from ff_incentw6 indicating a parallel to the IP6 approach. Codes 49, 50 and 51 were used to indicate the incentive amount for the IP7 refreshment sample. Tables 7.1 and 7.2 document the allocation of cases to experimental groups, though a shorthand allocation for the continuing sample is as follows:
Groups 1-12, 13, 16, 19, 22, 25, 28, 31, 34, 37, 40, 43, 46 = £10 unconditional incentive
Groups 14, 17, 20, 23, 26, 29, 32, 35, 38, 41, 44, 47 = £10 unconditional + £20 conditioned on whole HH completion
Groups 15, 18, 21, 24, 27, 30, 33, 36, 39, 42, 45, 48 = £30 unconditional incentive
Group 49 = IP7 refreshment sample £10 unconditional
Group 50 = IP7 refreshment sample £20 unconditional
Group 51 = IP7 refreshment sample £30 unconditional
At Wave 8, the incentive experiment paralleled the experimental allocation carried at previous waves with a repeat of the elevated incentive amounts for the IP7 refreshment sample comparable to the IP4 incentive experiment. Incentive allocation was nested within the mixed mode allocation at wave 8, with reference to the sample status vis., whether IP1/IP4 or IP7 refreshment such that the IP7 refreshment will maintain the incentive treatment it received at IP7. Households within IP1 (or IP4) PSUs were allocated the treatments such that all individuals within households received the same incentive, and all split-off households received the same incentive as their originating household.
IP1/IP4 continuing Face to Face sample (ff_gridmodew5 = 1 & ff_gridmodew8 = 1 & ff_hhorig = 7, 10) = £10 unconditional incentive.
IP1/IP4 former mixed-mode sample shifting to F2F (ff_gridmodew5 = 3 & ff_gridmodew8 = 1 & ff_hhorig = 7, 10) = £10 unconditional incentive.
IP7 refreshment sample issued Face-to-Face (ff_gridmodew8 = 1 and ff_hhorig = 11):
Group 1 – £10 unconditional incentive
Group 2 – £20 unconditional incentive
Group 3 – £30 unconditional incentive
Web allocated (i.e., ff_gridmodew8 = 3):
Group 1 – £10 unconditional incentive
Group 2 – £10 unconditional with a £20 conditional incentive for full-household completion by web in the allotted time
Group 3 – £30 unconditional incentive
The controlling variable is ff_incentw8 on record h_hhsamp_ip. It takes 87 unique values to reflect the historical incentive treatments for each case as determined by past incentive experiments and the current Wave 8 experimental allocation to incentive treatment. Tables 7.1-7.3 detail the meaning of each of these 87 codes which can be summarised as follows:
Groups 1-49, 52, 55, 58, 61, 64, 67, 70, 73, 76, 79, 82, 85 = £10 unconditional incentive
Groups 51, 54, 57, 60, 63, 66, 69, 72, 75, 78, 81, 84, 87 = £30 unconditional incentive
Group 50 = £20 unconditional incentive
Groups 53, 56, 59, 62, 65, 68, 71, 74, 77, 80, 83, 86 = £10 rising to £20 with full-household completion
At Wave 9, the incentive experiment has identical allocations to the incentive experiment at IP8. Households within IP1 (or IP4) PSUs were allocated to treatments such that all individuals within households received the same incentive, and all split-off households received the same incentive as their originating household.
IP1/IP4 continuing Face to Face sample (ff_gridmodew5 = 1 & ff_gridmodew8 = 1 & ff_hhorig = 7, 10) = £10 unconditional incentive.
IP1/IP4 former mixed-mode sample shifting to F2F (ff_gridmodew5 = 3 & ff_gridmodew8 = 1 & ff_hhorig = 7, 10) = £10 unconditional incentive.
IP7 refreshment sample issued Face-to-Face (ff_gridmodew8 = 1 and ff_hhorig = 11):
Group 1 – £10 unconditional incentive
Group 2 – £20 unconditional incentive
Group 3 – £30 unconditional incentive
Web allocated (i.e., ff_gridmodew8 = 3):
Group 1 – £10 unconditional incentive
Group 2 – £10 unconditional with a £20 conditional incentive for full-household completion by web in the allotted time
Group 3 – £30 unconditional incentive
The controlling variable is ff_incentw9. It takes 87 unique values to reflect the historical incentive treatments for each case as determined by past incentive experiments and the current IP9 experimental allocation to incentive treatment. Tables 1-3 (including amount of incentive for IP8 and IP9 only for space purposes) detail the meaning of each of these 87 codes which can be summarised as follows:
Groups 1-49, 52, 55, 58, 61, 64, 67, 70, 73, 76, 79, 82, 85 = £10 unconditional incentive
Groups 51, 54, 57, 60, 63, 66, 69, 72, 75, 78, 81, 84, 87 = £30 unconditional incentive
Group 50 = £20 unconditional incentive
Groups 53, 56, 59, 62, 65, 68, 71, 74, 77, 80, 83, 86 = £10 rising to £20 with full-household completion
At Wave 10, the incentive experiment had identical allocations as in wave 9. Households within IP1 (or IP4) PSUs were allocated to treatments such that all individuals within households received the same incentive, and all split-off households received the same incentive as their originating household.
IP1/IP4 continuing Face to Face sample (ff_gridmodew5 = 1 & ff_gridmodew8 = 1 & ff_hhorig = 11, 10) = £10 unconditional incentive.
IP1/IP4 former mixed-mode sample shifting to F2F (ff_gridmodew5 = 3 & ff_gridmodew8 = 1 & ff_hhorig = 11,12) = £10 unconditional incentive.
IP7 refreshment sample issued Face-to-Face (ff_gridmodew8 = 1 and ff_hhorig = 13):
Group 1 – £10 unconditional incentive
Group 2 – £20 unconditional incentive
Group 3 – £30 unconditional incentive
IP10 refreshment sample issued Face-to-Face (ff_gridmodew10 = 1 and ff_hhorig = 14):
Group 1 – £10 unconditional incentive
Web allocated (i.e., ff_gridmodew8 = 3):
Group 1 – £10 unconditional incentive
Group 2 – £10 unconditional with a £20 conditional incentive for full-household completion by web in the allotted time
Group 3 – £30 unconditional incentive
The controlling variable is ff_incentw10 (in record HHSAMP). It takes 88 unique values to reflect the historical incentive treatments for each case as determined by past incentive experiments and the current experimental allocation to incentive treatment. Tables 1-3 (including amount of incentive for IP8 and IP9 only for space purposes) detail the meaning of each of these codes which can be summarised as follows:
Groups 1-49, 52, 55, 58, 61, 64, 67, 70, 73, 76, 79, 82, 85, 88 = £10 unconditional incentive
Groups 51, 54, 57, 60, 63, 66, 69, 72, 75, 78, 81, 84, 87 = £30 unconditional incentive
Group 50 = £20 unconditional incentive
Groups 53, 56, 59, 62, 65, 68, 71, 74, 77, 80, 83, 86 = £10 rising to £20 with full-household completion
Additionally, all previous wave non-responding households which were issued to field received an incentive conditional on their survey participation at the value of their previously assigned group value.
At Wave 11, the incentive experiment was identical to the incentive experiment at IP10 for the IP1, IP4 and IP7 samples, with the exception that now the £20 conditional incentives were conditional on the individual, not household completing the survey. In previous waves, respondents received the £20 extra conditional on entire household completion. IP10 households, allocated to the mixed-mode design, follow the same allocations as IP1/IP4/IP7 mixed-mode households. The IP11 refreshment sample was also allocated to the mixed-mode experiment. All IP11 sample members received £10 unconditionally, with individuals in the mixed-mode receiving an extra £15 conditional on their completion of the web survey within the first three weeks. Households within PSUs were allocated to treatments such that all individuals within households will received the same incentive, and all split-off households received the same incentive as their originating household.
IP1/IP4 continuing Face to Face sample (ff_gridmodew5 = 1 & ff_gridmodew11 = 1 & ff_hhorig = 11, 12) = £10 unconditional incentive.
IP1/IP4 former mixed-mode sample shifting to F2F (ff_gridmodew5 = 3 & ff_gridmodew11 = 1 & ff_hhorig = 11,12) = £10 unconditional incentive.
IP7 refreshment sample issued Face-to-Face (ff_gridmodew11 = 1 and ff_hhorig = 13):
Group 1 – £10 unconditional incentive
Group 2 – £20 unconditional incentive
Group 3 – £30 unconditional incentive
IP10 refreshment sample issued Face-to-Face (ff_gridmodew11 = 1 and ff_hhorig = 14):
Group 1 – £10 unconditional incentive
Web allocated for IP1/IP4/IP7/IP10 samples (i.e., ff_gridmodew11 = 3 and ff_hhorig = 11, 12, 13, 14):
Group 1 – £10 unconditional incentive
Group 2 – £10 unconditional with a £20 conditional incentive for individual completion by web in the allotted time
Group 3 – £30 unconditional incentive
IP11 refreshment sample issued Face-to-Face (ff_gridmodew11 = 1 and ff_hhorig = 18):
Group 1 – £10 unconditional incentive
Web allocated for IP11 (i.e., ff_gridmodew11 = 3 and ff_hhorig = 18):
Group 1 – £10 unconditional incentive with a £15 conditional incentive for each individual completion on web within three weeks
The controlling variable is ff_incentw11 on record K_HHSAMP_IP. It takes 93 unique values to reflect the historical incentive treatments for each case as determined by past incentive experiments and the current IP11 experimental allocation to incentive treatment. Tables 7.1-7.3 (including amount of incentive for IP8-IP11 only for space purposes) detail the meaning of each of these 93 codes which can be summarised as follows:
Groups 1-49, 52, 55, 58, 61, 64, 67, 70, 73, 76, 79, 82, 85, 88, 89, 92 = £10 unconditional incentive
Groups 51, 54, 57, 60, 63, 66, 69, 72, 75, 78, 81, 84, 87, 91 = £30 unconditional incentive
Group 50 = £20 unconditional incentive
Groups 53, 56, 59, 62, 65, 68, 71, 74, 77, 80, 83, 86, 90 = £10 rising to £20 with individual completion
Group 93= £10 rising to £15 with individual completion
Additionally, all previous wave non-responding households which are issued to field received an incentive conditional on their survey participation at the value of their previously assigned group value.
At Wave 12, the initial unconditional incentives sent to households were identical to those sent at IP11, regardless of mode allocation at IP12. This allocation was simplified at IP12 by using a four condition controlling variable which captures almost every of the above 93 conditions: £10, £20, £30, and £10+£20. Allocation to these groups mirror the allocation to these values in ff_incentw11. The final allocation variable is ff_incentw12 on recordl_hhsamp_ip:
Group 1 – £10
Group 2 – £20
Group 3 – £30
Group 4 – £10 + £20
In addition, respondents received conditional incentives for completion of several of the health-based measures collected at IP12. In particular, respondents received £5 conditional on completion of each of the biomarkers needing outside of the interview completion. These additional measures include participation in:
- Completing a blood pressure measure on their own prior to the interview
- Mailing back dried blood samples
- Mailing back a hair sample
There is no experimental aspect to these conditional incentives; all respondents received an incentive for completion of each of these.
At Wave 13, the initial unconditional incentives sent to households were identical to what was sent at IP12, which was a simplified version of what was done at IP11, regardless of mode allocation at IP13. Households within PSUs were allocated the treatments such that all individuals within households received the same incentive, and all split-off households received the same incentive as their originating household.
Additionally, all previous wave non-responding households which were issued to field received an incentive conditional on their survey participation at the value of their previously assigned group value.
In addition to the ongoing incentive experiment, as part of the Fieldwork Compression experiment at IP13 (see Section 7.19), there was an additional incentive given depending on experimental condition. 80% of respondents were offered a longer survey, with ½ of this (40%) told of the longer interview in the advance letter, and ½ (40%) asked to answer additional questions within the survey. Respondents asked in the advance letter were sent an additional unconditional incentive of £5; those asked in the survey to answer more questions were offered a conditional incentive of £5.
At Wave 14, the incentive experiment was largely ended, as much information has been gleaned from incentive experiments and in order to match the incentive structure on the main Understanding Society survey. Most respondents received a £20 unconditional voucher to complete the survey, matching what is done in the main survey. The exception to this amount was for IP respondents who in past waves have received £30; their incentive remained at £30 pounds, as reduction to their incentive may adversely impact response rates. This leaves two distinct incentive amounts to compare.
Respondents were allocated based on their previous wave allocation, last adjusted (through simplification) at IP12. The allocation at IP12 followed the controlling variable ff_incentw12 which had households allocated to the incentive levels £10, £20, £30 and £10 + £20. The three groups allocated to incentives of £10, £20, and £10 + £20 were allocated to receive an incentive of £20, with households previously receiving £30 continuing to receive £30. Given this is entirely based on previously allocation, no new allocation is needed, and those in about 90% of households received £20, with the remaining 10% receiving £30. The controlling variable is set to Inapplicable (-8) for households from the wave 14 refreshment sample.
n_ff_incentw14 on record n_hhsamp_ip:
1 £20
2 £30
The Wave 14 refreshment sample included an incentive experiment. Half the sample were offered £30 if they completed the survey online within a five-week deadline. The other half were offered £20 for completing the survey, plus £10 if they completed it within two weeks. The ‘early bird’ deadline was mentioned in all mailings. This experiment is controlled by the variable N_FF_EARLYBIRD on record N_HHSAMP_IP:
1 Two week time frame to complete (£20 plus £10 if early)
2 Five week time frame to complete (£30 at end)
At Wave 15, the incentive experiment comprised only the retention of the £30 condition for those who had previously received £30 incentives. Given this is entirely based on previous allocation, no new allocation was needed. Participants in about 90% of households received £20, with the remaining 10% receiving £30. The controlling variable is on record o_hhsamp_ip:
1 = £20
2 = £30
At Wave 16, the incentive treatments from wave 15 were carried forward. The controlling variable is ff_incentw15 on record p_hhsamp_ip:
1 = £20
2 = £30
Tables 7.1 to 7.3 document the incentives experimental allocation
