In this survey, we presented participating parents with descriptions about sets of parenting programmes (more info on this in the method section below) and asked them about which of these programmes they would choose to access. We used these findings to inform the predictions about likely parent engagement made in our Parenting Support Commissioning Assistant for Local Areas (PASCAL). 

PASCAL holds information about 22 different parenting programmes and it predicts which of these programmes may have the largest positive impact for families in a local area. The predictions made by PASCAL factor in estimates about the proportion of parents who may choose to take up each programme, which are derived from these survey findings.

Our survey was put together using Qualtrics software, and fielded online with participants recruited from the UK by a large survey company from 14 August to 13 September 2025. Individuals who completed the survey received a small payment as compensation for their time, and there was also a small bonus for answering the comprehension questions correctly. We employed a cross-sectional survey design that incorporated a discrete choice experiment. Participants needed to be a parent or guardian of a child born in 2019 or earlier. We oversampled families from low incomes (either claimants of Universal Credit, or had a household income of below £40,000) due to our particular interest in understanding the preferences of families from low incomes.

Here we present a table of key features of the survey sample demographics and the same demographics for a representative sample of the UK population (Understanding Society (2023) weighted sample). To make the sample comparable, we take only respondents with a child aged six or under.

The core of the methodology involved presenting each respondent with descriptions of five randomly-selected parenting programmes. These five programmes were drawn from a total pool of 35 programmes, filtered to ensure their relevance to the ages of the respondent's children. 

Each programme description included a small summary and specific details on its delivery mode (online or in person), its domain focus (eg, maths or behaviour), and whether a child could join. Respondents were then asked a series of questions designed to elicit their preferences over the presented programmes, which allowed us to construct a model revealing the characteristics they most prefer (using a random utility model).

Our goal was to pinpoint which programme features make them most attractive to parents. To do this, we used a robust statistical model from economics, designed to understand how people make decisions.

Simulating the decision-making process

At its core, the model works by assuming that people make choices based on a mental "score" they give to each option. This score is built from a programme's different features (eg, whether it's online, its duration, the topics covered). By analysing thousands of choices, the model learns which features contribute most to a high score and, therefore, a parent's decision to sign up.

Accounting for individual differences

Our analysis considered that some parents are naturally more inclined to sign up for any programme and some programmes might have a general appeal that isn't captured by its specific features alone. By accounting for these nuances, we could be more confident that we were measuring the true impact of each specific feature.

If you’d like to hear more about the methodology, such as the details of the random utility model, please feel free to contact the researchers at [email protected]

There are some limitations to bear in mind:

• Sample representativeness: We did not gather responses from a fully representative sample as we were limited to parents that we could recruit during our timescales. Features of our sample that are not fully representative of the overall UK population include a substantially higher rate of degree completion, and there is an indicative level of digital literacy to complete the survey. White and Mixed ethnicity are overrepresented in our sample, and Asian, Black, and Other are underrepresented. To account for this, when reporting descriptive statistics we report in subgroups, as the totals will not represent a typical population. Even within subgroup reporting, these figures may not represent the full UK population.
• Stated preferences: When respondents say they would sign up and attend a programme, this does not mean they’d actually attend. However, we can be confident that if a respondent said they’d attend Programme A and not Programme B, that they preferred Programme A to Programme B.
• Retention: We cannot say anything directly about what happens after sign-up, specifically concerning metrics like retention or completion rates for parents participating in programmes
• Limited programme features: This analysis looks only at a restricted number of programme features. We know there are additional features that may be important in a parent's preferences such as who is delivering the programme or whether the programme is delivered in a trusted space. It was not possible to examine these features which may vary between delivery locations.