{"id":41401,"date":"2026-02-19T13:34:19","date_gmt":"2026-02-19T14:34:19","guid":{"rendered":"https:\/\/euknow.it\/building-an-evaluation-system-the-case-of-the-family-star-project\/"},"modified":"2026-02-19T13:34:19","modified_gmt":"2026-02-19T14:34:19","slug":"building-an-evaluation-system-the-case-of-the-family-star-project","status":"publish","type":"post","link":"https:\/\/euknow.it\/en\/building-an-evaluation-system-the-case-of-the-family-star-project\/","title":{"rendered":"Building an evaluation system: the case of the &#8220;Family STAR&#8221; project"},"content":{"rendered":"<p>Let&#8217;s delve into what it means to design and implement an evaluation system in a European project with a concrete case.<!--more--><\/p>\n<p>This article is produced as part of our Guide&#8217;s collaboration with ASVAPP, Association for the Development of Public Policy Evaluation and Analysis. The article is edited by Lisa Zaquini and Francesco Tarantino. <\/p>\n<h2>Family STAR: an example of experimental evaluation of a European project<\/h2>\n<p>Following the introductory <a href=\"https:\/\/euknow.it\/en\/evaluating-interventions-what-how-and-why\/\">article<\/a> published a few months ago, in which we addressed why it is useful to evaluate and how or when it is possible to do so, today we delve into the applicability of some of the evaluation logic presented to a concrete case: the European Family STAR project.<\/p>\n<p>We will briefly introduce the intervention, the context in which it was implemented, and how its evaluation system was designed and implemented. Reviewing these activities, we will provide an example to understand the choices to be made, the tools available, and the challenges and benefits associated with building and implementing an evaluation design. <\/p>\n<p>The Family ST.A.R.<strong>(Family Group Conferences and Students at Risk<\/strong>) project was implemented between 2016 and 2018 under the EaSI (Employment and Social Innovation) Program, which is now one of the components of the <a href=\"https:\/\/euknow.it\/en\/guida\/funding-categories-2021-2027\/structural-and-rural-funds\/esf\/\">European Social Fund.<\/a> The call was about testing social policy innovations to support reforms in social services and specifically called for a rigorous evaluation of the project&#8217;s effects. The project, with a budget of more than 1 million euros, was aimed at testing the effectiveness of Family Reunions in promoting well-being and preventing school dropout for vulnerable secondary school students.  <\/p>\n<p>That of <a href=\"https:\/\/en.wikipedia.org\/wiki\/Family_Group_Conference\" target=\"_blank\" rel=\"noopener\">Family Reunions (FCRs)<\/a> is a model that originated in New Zealand in the 1980s in the field of child protection. It involves an organized series of meetings aimed at helping the family (broadly understood: children and parents, but also relatives, friends, and neighbors) make shared decisions for the safety and well-being of the child. This approach is used in a variety of settings. With Family STAR, DRFs have been applied to the relatively new and previously untested school setting: hence the importance of proving its effectiveness.   <\/p>\n<p>In response to this call, a <strong><a href=\"https:\/\/www.asvapp.org\/family-group-conferences-and-student-at-risk\/\" target=\"_blank\" rel=\"noopener\">partnership<\/a><\/strong> capable of taking up a twofold challenge: to <strong>experiment<\/strong> in several schools with the DRF model as a preventive intervention for hardship and school dropout, and at the same time to rigorously gather empirical evidence on the<strong>effectiveness of the<\/strong> proposed<strong>model<\/strong>. It therefore involved entities deputed to the provision of educational and social services and research centers with solid experience in conducting experiments and analyses in the educational and social fields. <\/p>\n<p>Family STAR was implemented in five locations, both urban and rural, in northern and southern Italy, involving <strong>55<\/strong> secondary <strong>schools<\/strong> in the pilot, with a total of <strong>540<\/strong> <strong>students<\/strong> nominated for an RFP. It had a duration of three years: the first deputed to the implementation of preparatory activities (involvement of schools, training of operators, development of data collection tools), the second and third to the implementation of interventions in schools and data collection and analysis. <\/p>\n<h2>Intervention as &#8220;a drug&#8221;: the choice of counterfactual approach<\/h2>\n<p>The project involved the adoption of two different evaluation logics, effects evaluation and implementation analysis, which (as we shall see) played a complementary role in describing project outcomes.<\/p>\n<p>The <strong>evaluation of effects<\/strong> aimed to answer the following question, &#8220;What level of well-being and what outcomes would the students who participated in the DRFs have had if they had not had the opportunity to take part in the DRFs?&#8221;<\/p>\n<p>He therefore wanted to estimate the impact of the intervention according to the <strong><a href=\"https:\/\/euknow.it\/en\/guida\/working-on-projects\/managing-the-project-monitoring-and-evaluation\/\">counterfactual logic<\/a><\/strong>, which defines the effect (or impact) as the difference between the conditions of the students targeted by the intervention (factual situation) and the conditions that would have occurred in the absence of the intervention (counterfactual situation). This is a particularly rigorous approach, which involves the challenge of artificially reconstructing a hypothetical situation: the same student cannot be both a participant (factual) and a nonparticipant (counterfactual) at the same time. <\/p>\n<p>The solution is to construct a credible <strong>control group<\/strong>, i.e., composed of students who are on average similar in everything to those targeted by the intervention: both in &#8220;observable&#8221; variables (age, gender, grades&#8230;) and in unobservable variables, i.e., capable of influencing the path and conditions exiting the project, but not directly detectable. A control group is credible if two conditions are met: a random assignment process (i.e., by drawing lots) and the presence of a large number of students to be assigned to the two groups. This method is called a randomized controlled experiment (or <strong>randomized<\/strong> controlled trial, RCT) in the literature and is the same method used in clinical settings to test the effectiveness of drugs and protocols.  <\/p>\n<p>The Family STAR &#8220;treated group&#8221; was proposed to participate in the DRF, while the &#8220;control group&#8221; was observed to describe the situation the treated would have been in if they had not accessed the intervention. By comparing the conditions of the treaties and controls, it was possible to obtain an <strong>estimate of the average effect<\/strong> of participation in a DRF on well-being and several other dimensions of interest. <\/p>\n<p>This is a <strong>rigorous<\/strong> and sometimes opposed <strong>method<\/strong> because it involves a random choice between who will or will not benefit from a potentially beneficial intervention. However, the estimates that are obtained are robust, i.e., following a counterfactual evaluation it can indeed be argued that the observed benefits have been achieved as a result of the intervention, and are not the result of &#8211; for example &#8211; spontaneous dynamics. <\/p>\n<p>Adopting this approach required setting a strict timetable for different phases of the experiment:<\/p>\n<ul>\n<li>the candidacy of students who could have potentially benefited from an RFP;<\/li>\n<li>The draw (randomization) into treaties and controls;<\/li>\n<li>The initiation of activities;<\/li>\n<li>The collection of descriptive data on conditions in and out of the project for all students involved.<\/li>\n<\/ul>\n<p>For the data collection, it was chosen to expand the student population to include the entire school population of the institutions involved, in order to be able to describe the characteristics of the students nominated by teachers also in comparison with the school population in general, and to be able to return to each school involved an overview of the conditions of all its students.<\/p>\n<h2>From the Family STAR Logical Framework to Implementation Analysis.<\/h2>\n<p>In parallel, the design for the so-called implementation analysis was prepared. This is an analysis aimed not so much at estimating the effect, but at verifying how an intervention was implemented: what <strong>obstacles<\/strong> arose in the implementation of the project? What <strong>strategies<\/strong> were implemented to deal with them? Overall, was the intervention implemented as planned?   <\/p>\n<p>The starting point for defining the research design is the reconstruction of the <a href=\"https:\/\/euknow.it\/en\/guida\/working-on-projects\/how-to-structure-a-project-the-process-and-tools\/\">intervention&#8217;s Logical Framework<\/a>, or a description of the implementation process that highlights any critical steps, challenges and threats to the effectiveness of the intervention itself. The Logical Framework then describes what <strong>is supposed to happen<\/strong>, and the implementation analysis goes to see if it actually happened on time and as planned. <\/p>\n<p>In the case of Family STAR, the critical points identified in the drafting phase of the logical framework concerned the weight of the &#8220;human variable&#8221; and, in particular, the possibility of <strong>poor adherence<\/strong> of the different actors (schools, families, teachers) to a very innovative design and evaluation approach-a risk that materialized and was addressed within the project (see next section).<\/p>\n<p>Given these critical issues, the <strong>most suitable tools<\/strong> for analyzing specific project phases were identified. The development of these tools took several months. Specifically, these were:  <\/p>\n<ul>\n<li>an effects estimation questionnaire to be administered to students, for which the paper machine-readable mode was chosen: a more expensive mode, but also easier for schools;<\/li>\n<li>A satisfaction questionnaire, to be administered to all participants in the RFPs, again paper-based and machine-readable;<\/li>\n<li>An online questionnaire for training participants;<\/li>\n<li>An online questionnaire to class coordinating teachers;<\/li>\n<li>A logbook, compiled by facilitators and spokespersons throughout the journey;<\/li>\n<li>several tracks for interviews and focus groups with local project coordinators, trainers, school leaders, school project liaisons, facilitators and spokespersons.<\/li>\n<\/ul>\n<p>The main difficulty encountered in the development phase of data collection instruments, particularly with regard to questionnaires, is finding the <strong>right balance<\/strong> between the comprehensiveness of the questions and the streamlining of the instrument itself: on the one hand, in fact, it is in the interest of researchers to collect as much information as possible; on the other hand, however, it is necessary to use parsimony so as not to overburden the respondent.<\/p>\n<p>Such a trade off is always valid, and in the Family STAR project it required several reflections: the research team opted to survey in <strong>greater detail<\/strong> the characteristics of the RFDs through the tool of the logbook, filled in by facilitators and spokespersons, who-as project workers-could count the time spent filling in in the total amount of time devoted to preparing and conducting each RFD.<\/p>\n<p>On the other hand, with regard to the instruments aimed at parents and-to a lesser extent-teachers, special attention was paid to <strong>ease of response<\/strong>, both in terms of content (avoiding asking the respondent questions that involved particularly difficult reflections) and in terms of the length of the questionnaire and the time required to complete it.<\/p>\n<h2>Realization vs Evaluation: opposing needs?<\/h2>\n<p>The design of the evaluation framework and development of the tools took about <strong>a year<\/strong> to complete, during which the implementers began the phase of engaging schools.<\/p>\n<p>The <strong>second year<\/strong> of the project saw the start of the actual evaluation and project activities, immediately encountering a low adherence of schools to the project and the application of a small number of students by teachers, such that the achievement of the project objectives (number of participants in the RFPs and the control group) and the significance of the achievable estimates were jeopardized.<\/p>\n<p>The implementation analysis investigated the reasons underlying the low uptake of the project by schools and individual teachers. Project activities were <strong>reshaped<\/strong>, through: <\/p>\n<ul>\n<li>a lengthening of the implementation phase of the RoFs over two years and the engagement of new local organisations (and therefore new schools), to widen the pool of participants and compensate for the low take-up;<\/li>\n<li>the suggestion to teachers to report &#8220;twin cases,&#8221; i.e., students who-according to their responses to the initial questionnaire-appeared to be facing similar difficulties as the students nominated for RoFs. Once the list of twin cases was received from the research team, the decision on whether to nominate them for a RoF remained with the teachers, subject to the need to draw them into the treaty or control group; <\/li>\n<li>the possibility for teachers to indicate cases for which a DRF was urgent and which were therefore included in the treatment group without being randomized and thus effectively exiting the evaluation.<\/li>\n<\/ul>\n<p>Thus, an attempt was made to resolve the conflict between the needs of <strong>schools<\/strong> and those <strong>of evaluation<\/strong>. The needs of assessment have not always been well received by school leaders, who have said that they feel they have to devote more attention to teaching rather than research. Such divergence of interests is typical in the world of evaluation, where <strong>opposing needs<\/strong> meet (and sometimes clash). The evaluator is therefore called upon to carefully consider whether to make pressing demands of the stakeholders, yet without losing the scientific rigor required to provide answers to scientifically relevant questions.   <\/p>\n<p>A deep understanding of the context in which one operates, as well as a phase of shared involvement and planning certainly help to identify appropriate timeframes and tools that can meet the opposing needs.<\/p>\n<p>Relevant-though not always taken into account-is the aspect of future prospects: while the evaluator&#8217;s task ends with the presentation of the results of the analysis and policy directions, the actors involved in service design certainly have broader perspectives, and it is useful to take them into account to ensure that the proposed innovations can be followed up even after the research activities are completed.<\/p>\n<h2>Lessons learned: assessment outcomes<\/h2>\n<p>Despite the difficulties encountered, <strong>the effects analysis<\/strong> was able to show that the intervention was effective in improving students&#8217; perceived parental support and reducing their conflicts with teachers. However, it also showed that DRFs did not improve students&#8217; self-efficacy or make them more optimistic about their future educational paths. The effects seem more positive when considering students from families with higher cultural capital and when the relationship between parents and teachers is not conflictual.  <\/p>\n<p><strong>The implementation analysis<\/strong> revealed lights and shadows of the implementation process of the activities: there were the aforementioned obstacles to achieving full involvement of schools in the project, mainly due to lack of resources and skepticism towards innovative models; however, in specific circumstances, the data show a general appreciation for participatory and family empowerment models by all actors. Interesting information related to gender and father figure involvement was also collected. <\/p>\n<p>In conclusion, the counterfactual approach adopted in the research design developed for the Family STAR project is obviously not <strong>replicable<\/strong> to all European projects, as it requires dedicated resources, a structured approach, a large volume of potential beneficiaries, and the ability to build a control group with certain characteristics. In addition, the consistent adoption of a counterfactual approach made it necessary to overcome obstacles and reshape the project activities themselves. <\/p>\n<p>However, the combination of two different <strong>evaluation logics<\/strong> (effects evaluation and implementation analysis) allowed the project to produce solid<strong>evidence<\/strong> with respect to both its effects and the processes activated. The results of the evaluation made it possible to draw firm conclusions with respect to the effectiveness of the approach tested and can guide the implementation and progressive improvement of future interventions. <\/p>\n","protected":false},"excerpt":{"rendered":"<p>Let&#8217;s delve into what it means to design and implement an evaluation system in a European project with a concrete case.<\/p>\n","protected":false},"author":10,"featured_media":41231,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"inline_featured_image":false,"footnotes":""},"categories":[731,2189],"tags":[2314,2325],"class_list":["post-41401","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-archive-archive","category-archive-stories","tag-experiences","tag-how-to-methods-for"],"acf":[],"_links":{"self":[{"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/posts\/41401","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/users\/10"}],"replies":[{"embeddable":true,"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/comments?post=41401"}],"version-history":[{"count":0,"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/posts\/41401\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/media\/41231"}],"wp:attachment":[{"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/media?parent=41401"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/categories?post=41401"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/euknow.it\/en\/wp-json\/wp\/v2\/tags?post=41401"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}