Key Features to look out for

1. The evaluation output relates to a plan for the service and the evaluation. Information about the service and evaluation planning are provided and findings are then linked to these
   This could be explicit reference to a logic model and evaluation plan, or more oblique reference to the aims of the service and evaluation.
   Eg., “[Better Start Imagine] has two components; a book gifting service that delivers a free book every month… and wrap-around sessions aimed at supporting families to feel confident in sharing books and stories with young children… The overall aim of the evaluation was to understand how both components of Better Start Imagine were implemented and received by the community.” (Better Start Imagine Final Report, January 2025)
   1. Reference to a plan for the service and the evaluation suggests (although it doesn’t guarantee) that both had been thought through in advance, suggesting robust service design and evaluation processes.  As emphasised throughout this framework it is essential to evaluation that you know what you are evaluating.
      
2. Data, whether quantitative or qualitative, is contextualised and gives sufficient detail to be useful (see also Interpreting and analysing basic data)
   This means that any data provided is explained sufficiently that the reader can understand whether this is a big or small, strong or weak, etc.
   Eg., “Qualitative interviews and focus groups were conducted with 37 participants at time 1 and a further 36 at time 2. Below are some of the key themes identified during analysis of the qualitative data… Training was generally viewed positively by participants. Most who had attended training valued it and believed it improved their confidence in their practice. Others described the benefits of more bespoke training packages”. (Little Minds Matter final report, October 2024)
   Eg., “Feeding intentions were recorded for 4524 of the contacted women. The majority of women, 55% (2472 women), said they intended to breastfeed, and 17% (770) said they intended to breastfeed alongside feeding formula milk”  (Breastfeeding Final Report, September 2023). 
   1. 1. 1. Context and detail like this means the reader can clearly understand the scale of the data being provided and the way in which it was gathered.
            
3. Recruitment of service users (See also Performance indicators)
   This means understanding how many people entered the service and may be defined in several ways, which should be set out in your service design/manual. This matters because it provides an indication of how well known the service is and its integration into the wider system in terms of referrals. Ideally recruitment data will examine numbers of people recruited, what their recruitment pathways looked like, and possibly the criteria under which they were recruited. 
   Eg., In total, 609 referrals were made to the Incredible Years – Toddler (IY-T) service. The most common referral source was the IY team (n=446, 75%) highlighting the success of their community engagement activities, but also IY-T’s reliance on gaining referrals through their own outreach work. The second most common referral source was schools/nurseries (n=53, 9%)… An enrollee is defined as someone who is seen face to face in at least one pre-course contact and booked onto a programme rota. The target set for enrollees was 828. There have been 417 enrollees over the entire delivery of IY-T (50% of the overall target). This represents approximately two thirds (64%) of all referrals. (Incredible Years Final Report, October 2023
   1. This example provides the definition of an enrollee and describes the ways in which people entered the service
4. Completion of support This means understanding how many people finished an episode of support and may be defined in several ways, which should be set out in your service design/manual. This matters because it provides an indication of how acceptable the service is to service users, even if recruitment is high, if the service is unacceptable to users in some way completion may be low. Eg., “Completion of the project was defined as attending at least 5 of the 8 sessions. Anticipated numbers were based on 75% of participants going on to complete. Overall, 626 families completed the Group programme, which represents 74% of all participants, very close to target. Rates of completion varied year on year, but most years were above 70% which demonstrated high levels of completion of the programme” (HENRY Final Report, December 2024)
   1. This example shows the detail provided in relation to completion, as a percentage, a number, and in relation to the target. It also provides a definition of completion to aid understanding of the statistics. Exactly the same can be done for recruitment and reach.
5. Reach
   This means understanding how the demographics of those recruited and completing compares to the demographics of the target population, whoever they are. This matters because a service will always have some target population, even if that is ‘everyone nearby’, but if you find that you are failing to recruit in a way that reflects that target population then changes to the service are required in order to address that. It may be that some aspect of the service or outreach is unacceptable to a certain group or is just not reaching them. It may also be that assumptions made in the service design about who needed the service were not accurate.
   Eg., “Of those with ethnicity data 57% were Asian/Asian British Pakistani, slightly over-representing this group. White British women accounted for 9%, close to the proportion seen in the maternal population. Women of White Other backgrounds were underrepresented (2%). However, nearly a third of women were from other minority backgrounds which may reflect the number of referrals recorded as recent migrants or asylum seekers.” (Bradford Doulas Final Report, September 2024). 
   1. This example compares the ethnicity of participating women with the maternal population of the area served. It provides both detail of how those differ and some explanations as to why. Here we may conclude that further investigation is needed to understand whether the need is greater amongst the group who are reaching the service than among the general maternal population or whether the service is not known amongst a certain demographic.  
6. Findings are not overstated
   This means that recognition is given to the limitations of the evaluation – a before and after evaluation, for example, does not prove that a service is the reason for a change, only that it could be. This does not mean that the evaluation is low quality or unhelpful, just that it can only tell us some things.
   Eg., “Findings suggest accessing Family Action Peer Support may lead to improvements in symptoms of depression and anxiety for women. However, given the relatively low number of pre and post measures included in the analysis, and the lack of a control group these findings should be interpreted with caution. These results do not tell us that the service is effective, however, they do give a promising indication that participation in the service may result in improved mental health outcomes.”  (Family Action Appendix to Final Report, July 2024)
   1. This is useful, it shows that there is promise, and that further investigation is worthwhile. Compared to a service with no evaluation, or poorly conducted evaluation this is significant.