Research and analysis

Research with self-employed Universal Credit claimants: Technical note

Updated 5 April 2024

This technical note accompanies two slide decks presenting key findings from research with self-employed Universal Credit (UC) claimants. Self-employed workers on a low income may apply for support from Universal Credit. If a claimant is gainfully self-employed (their self-employment is their main activity and is regular, organised, developed and carried out in expectation of profit)^{[footnote 1]}, UC payments are calculated using an assumed level of earnings called a Minimum Income Floor (MIF). At the time of this research, due to the COVID-19 pandemic, the tests to assess gainful self-employment (i.e. the gateway interviews) were temporarily suspended and the MIF was not applied to UC claims.

This research consisted of a 2-wave tracking survey followed by in-depth interviews with survey respondents. This document details the methodology employed for the survey and the qualitative research, providing further insight into the sample size and research approach. This research was commissioned by DWP and conducted by Ipsos (at the time under the name Ipsos Mori).

1. Survey

1.1. Methodology

For the first wave of the survey a mixed mode online and telephone approach was used to survey 5,159 existing UC claimants and 5,062 new UC claimants. The second wave of the survey was conducted in the same way 12 months later. For Wave 2, Ipsos both recontacted claimants who took part in Wave 1 and contacted a DWP-provided booster sample to ensure the final survey population was representative of the UC self-employment population in the summer of 2021. Wave 2 achieved 2,742 responses from the recontact sample (1,411 existing UC claimants and 1,331 new claimants) and 1,953 from the booster sample.

Sample

The sample for Wave 1 was provided by DWP. The sample was split into two groups based on their claim start date:

• existing claimants who started claiming UC as a self-employed person before 16 March 2020. This group had their self-employment status assessed by Jobcentre Plus and attended a gateway interview, and as a result were deemed gainfully self-employed (GSE). Some claimants may have had the Minimum Income Floor (MIF) applied before it was suspended in March due to COVID-19.
• new (2020) claimants who started claiming UC as a self-employed person on or after 16th March 2020 (and up to 22nd June 2020 when the sample was drawn).This group did not have their self-employment status assessed by Jobcentre Plus and they had not attended a gateway interview, so may be deemed either GSE or non-GSE when gateway interviews resume in future. The MIF had not been applied to this group because the gateway interviews had not taken place, and because the MIF was suspended in March due to COVID-19.

The full population^{[footnote 2]} of existing claimants were invited to take part in the survey, whilst for new claimants a random sample was drawn. Further details on the achieved sample profile can be found in Tables 2.1 and 2.2.

The sample for Wave 2 consisted of the claimants that took part in Wave 1 – the recontact group – alongside the new 2021 group:

• new 2021 claimants who started claiming UC as a self-employed person between 23rd June 2020 and 7th July 2021 (when the booster sample was drawn). These claimants have the same position as the new 2020 group in terms of their interaction with UC policy: they had not attended a gateway interview, so may later have been deemed GSE or non-GSE and the MIF had not been applied.

To distinguish between the two ‘new’ groups this research refers to them as the new 2020 and new 2021 groups. Sample profile for the two recontact group and the new 2021 group are detailed in Tables 2.3, 2.4 and 2.5.

Questionnaire development

The questionnaire was developed in close collaboration with DWP. Before Wave 1, 10 cognitive interviews were carried out to ensure the questions were clear and interpreted in the same way by different people. A pilot of 30 interviews was also undertaken to test the survey functionality and interview length.

Fieldwork – Wave 1

Fieldwork took place between 18th September 2020 and 2nd November 2020.

Claimants were initially invited to take part in the online survey via email. Those who had not taken part received up to four reminders via email. After three weeks of fieldwork, those who had not responded were contacted by telephone to complete the survey with an interviewer.

Table 1.1 below shows a breakdown by mode.

Table 1.1: Achieved interviews by mode

Achieved interviews

Mode	Existing claimants	New claimants
Online	3,588	3,593
Telephone	1,571	1,469
Total	5,159	5,062

Fieldwork – Wave 2

Fieldwork for Wave 2 was conducted between 10th September 2021 and 21st November 2021, following a similar reminder emails strategy as Wave 1.

Claimants who took part in Wave 1 and claimants in the booster sample were invited to take part in the online survey via email and later followed up via telephone.

Table 1.2 below shows a breakdown by mode.

Table 1.2: Achieved interviews by mode

Achieved interviews

Mode	Existing claimants (recontact from W1)	New 2020 claimants (recontact from W1)	New 2021 claimants
Online	754	726	945
Telephone	657	605	1,008
Total	1,411	1,331	1,953

Weighting

For Wave 1, the data is weighted to be representative of the total populations of existing and new claimants, based on gender, age group, whether claimants have children, and whether they have a single or a couple claim. For Wave 2 the recontact sample is weighted to the original population profiles.

These population estimates were derived from the DWP dataset of all self-employed UC claimants, the dataset from which the sample was originally selected. For Wave 1 separate weights were generated for the three claimant groups: existing and new 2020.The New 2021 sample (for Wave 2) is weighted to the sample proportions supplied at the time it was drawn. The combined New sample has a secondary weight to reflect the profile of the new population.

Data analysis

All reported differences are statistically significant at the 95% confidence level, unless otherwise specified. This means, if this survey was run 100 times, each time with a different sample of people, the survey results would be similar to what one would see if the whole population was asked the survey question, 95 out of 100 times. Not all statistically significant differences are noted.

Following the survey, a segmentation analysis was conducted using a cluster analysis technique, drawing on 8 attitudinal statements measuring respondents’ interest in self-employment, knowledge of sources of advice and guidance to support self-employment, as well as their own financial health. Within each of the existing, new 2020, and new 2021 claimant groups, this identified five segments of claimants, where participants within each group held similar views.

Sample profile

Wave 1

Table 2.1 and Table 2.2 show the achieved sample profile for Wave 1 and the effect of weighting for new and existing claimants. Where an asterisk (*) is displayed, this means the value for that response is greater than zero, but less than 0.5%.

Table 2.2: Sample profile and weighting for existing claimants

Demographic	Unweighted number	Unweighted (%)	Weighted (%)
Total	5,159	100	100
Age: 17-24	117	2	3
Age: 25-34	1,094	21	28
Age: 35-44	1,857	36	37
Age: 45-54	1,378	27	22
Age: 55+	712	14	9
Age: Unknown	1	*	*
Gender: Female	2,198	42	36
Gender: Male	2,960	57	64
Gender: Unknown	1	*	*
Children: Yes	3,338	65	70
Children: No	1,821	35	30
Claim: single	3,023	59	51
Claim: couple	2,136	41	49

Table 2.3: Sample profile and weighting for new claimants

Demographic	Unweighted number	Unweighted (%)	Weighted (%)
Total	5,062	100	100
Age: 17-24	193	4	7
Age: 25-34	1,086	21	29
Age: 35-44	1,247	25	27
Age: 45-54	1,383	27	22
Age: 55+	1,149	23	14
Age: Unknown	4	*	*
Gender: Female	1,904	38	30
Gender: Male	3,154	62	70
Gender: Unknown	4	*	*
Children: Yes	1,648	33	34
Children: No	3,414	67	66
Claim: single	2,645	52	53
Claim: couple	2,417	48	47

Wave 2

Tables 2.3, 2.4 and 2.5 provide a breakdown of the Wave 2 achieved samples for each group (existing recontact, new 2020 recontact, and new 2021) and the effect of weighting for each. Where an asterisk (*) is displayed, this means the value for that response is greater than zero, but less than 0.5%.

Table 2.3: Sample profile and weighting for existing recontact claimants

Demographic	Unweighted number	Unweighted (%)	Weighted (%)
Total	1,411	100	100
Age: 17-24	23	2	3
Age: 25-34	238	17	28
Age: 35-44	485	34	37
Age: 45-54	422	30	22
Age: 55+	243	17	9
Age: Unknown	0	0	0
Gender: Female	631	45	36
Gender: Male	780	55	64
Gender: Unknown	19	1	1
Children: Yes	826	59	68
Children: No	566	40	31
Children: Unknown	0	0	–
Claim: single	874	62	51
Claim: couple	537	38	49

Table 2.4: Sample profile and weighting for new 2020 claimants

Demographic	Unweighted number	Unweighted (%)	Weighted (%)
Total	1,331	100	100
Age: 17-24	33	3	7
Age: 25-34	234	18	29
Age: 35-44	287	22	27
Age: 45-54	400	30	22
Age: 55+	376	28	14
Age: Unknown	1	0	0
Gender: Female	518	39	30
Gender: Male	812	61	70
Gender: Unknown	1	*	0
Children: Yes	407	31	38
Children: No	900	68	60
Children: Unknown	24	1	2
Claim: single	727	55	53
Claim: couple	604	45	47

Table 2.5: Sample profile and weighting for new 2021 claimants

Demographic	Unweighted number	Unweighted (%)	Weighted (%)
Total	1,953	100	100
Age: 17-24	118	6	6
Age: 25-34	540	28	29
Age: 35-44	544	28	30
Age: 45-54	438	22	22
Age: 55+	313	16	13
Age: Unknown	0	0	0
Gender: Female	822	42	38
Gender: Male	1,131	58	62
Gender: Unknown	0	0	0
Children: Yes	841	43	45
Children: No	1,051	54	52
Children: Unknown	61	3	3
Claim: single	1,297	34	64
Claim: couple	656	66	36

2. Interviews

Qualitative research aims to explore experiences in-depth and to understand how and why issues occur; it does not seek to be statistically representative.

The design was small-scale and exploratory. A total of 55 in-depth interviews with participants from Wave 2 were conducted January to March 2022. The sample was organised into two groups: 28 Existing claimants and 27 New claimants (both recontact and New 2021) to enable us to understand the experiences of both groups.

The sample also comprised a mix of claimant demographics, including age, gender, age of children, and health. The achieved sample profile is outlined in table 3.1. Claimants were purposively selected from a sample of self-employed claimants who had taken part in Wave 2 of the survey. Participants were recruited to take part using specialist recruiters and guided by a screening questionnaire designed with DWP to ascertain eligibility and to check key quotas (as set in the sample profile).

The interviews took approximately 45 minutes and were undertaken by Ipsos researchers using a discussion guide agreed with DWP. The interview data was analysed using a robust inductive framework approach, as part of which the data was synthesised thematically and interrogated for patterns and relationships. As noted outputs are not statistically representative.

Table 3.1 Quotas and achieved sample for the qualitative element

Characteristic	Description	Quota	Achieved
Sample type	Recontact existing (W1)	28	28
Sample type	Recontact new (2020- W1)/ NEW (2021- W2)	27	27
Household composition	Single claim	Mix and monitor	29
Household composition	Joint claim	Mix and monitor	26
Parents	With children	Mix and monitor	29
Parents	Without children	Mix and monitor	26
Age	17-34	Mix and monitor	19
Age	35-44	Mix and monitor	20
Age	45+	Mix and monitor	16
Gender	Male	Min. 20	20
Gender	Female	Min. 20	35
Employment	Self-employed only	Min. 30	43
Employment	Employed and self-employed (dual income)	Min. 10	12
Employment	Move off UC	Min. 10	12
Long term health condition	Yes	Min. 10	20
Carers	Yes – those with caring responsibilities	Min. 5	9
Geography	Urban (in a town or city)	Good spread of urban/ rural locations	16
Geography	Suburban (near a town or city)	Good spread of urban/ rural locations	29
Geography	Rural or semi-rural (in the countryside)	Good spread of urban/ rural locations	10
Length of self-employment	Under 12 months	Spread of start-up and time in self-employment	6
Length of self-employment	12 months to less than 2 years	Spread of start-up and time in self-employment	11
Length of self-employment	2 years to less than 3 years	Spread of start-up and time in self-employment	10
Length of self-employment	3 years to less than 5 years	Spread of start-up and time in self-employment	13
Length of self-employment	5 years or more	Spread of start-up and time in self-employment	15
Sector	Food/ drink/ hospitality/ leisure	Good spread across sectors	3
Sector	Professional services	Good spread across sectors	9
Sector	Manufacturing/ construction/ engineering	Good spread across sectors	3
Sector	Media/ telecoms/ arts	Good spread across sectors	7
Sector	Transport/ distribution/ delivery	Good spread across sectors	2
Sector	Domestic/ Utilities/ vehicle services	Good spread across sectors	6
Sector	Personal services/ beauty/ sport & fitness	Good spread across sectors	7
Sector	Retail/ sales	Good spread across sectors	3
Sector	Social care/ healthcare/ education/ childcare	Good spread across sectors	6
Sector	Other	Good spread across sectors	4
Sector	Animals/ agriculture	Good spread across sectors	5
MIF applied	Yes	15 (will need to review)	24
Total interviews	–	55	55

1. Only claimants with a work expectation can be found GSE. Those who are not expected to look for work cannot be found GSE but are required to declare their self-employed earnings each assessment period. See further information on which claimants are expected to work ↩

2. Those who were invited to take part in the pilot were not re-contacted for the mainstage fieldwork, and those who did not have valid contact details were excluded from the sample. ↩