"This report brings together disparate terminology, findings and recommendations from the private and public sectors and academia to synthesise a set of five general steps for practitioners when performing impact assessments. Impact assessment (IA) compares a variable of interest after an intervention (e.g. a communication campaign) to what it would it be if that intervention had never happened. Accompanying each of the five steps are examples of how migration communicators can assess the impact of campaigns to, first, change attitudes and, second, change behaviour.
Step 1: Set objectives for the intervention that define what the desired effect is. Ideally, this should (1) meet the SMART criteria of specific, measurable, attainable, relevant and timebound; (2) focus on ultimate outcomes, rather than only outputs; (3) avoid a number of common mistakes in setting objectives and (4) carefully consider what the—as specific as possible— target audience is.
Step 2: Identify a measure of the impact(s) that the intervention has as its objective. These should be (1) clearly defined and unambiguous; (2) externally valid (they act reliably if used for other interventions) and (3) internally valid (they measure what they claim to measure). There already exist long-lists of measures (also known as indicators) to choose from.
Step 3: Estimate what would have happened had there been no intervention, known as a ‘counterfactual’. Ideally, but not necessarily, this requires performing a pre-intervention measurement, also known as baseline assessment. However, this is not always possible. This should not dissuade practitioners but instead assumptions of any IA, as discussed below, should be made openly. A long list of methods for measurement, including sampling considerations, are provided.
Step 4: Perform the intervention. This may involve separating a randomised sample into a treatment group(s)—that receives the intervention(s)—and a control group—that does not. If this is not feasible, other forms of counterfactual are possible that ‘construct’ a control group. This section also overviews: (1) types of communication campaigns; (2) the MINDSPACE checklist of behavioural influences for interventions; (3) key recommendations on persuasive interventions from ICMPD’s previous work; (4) recommendations on campaigns deterring irregular migration.
Step 5: Post-intervention measurement and analysis: (1) those receiving the intervention should be measured post-intervention (ideally, as well as, either a genuine or constructed control group); (2) the impact must be calculated: usually the difference in the variable of interest pre- and post-intervention and/or between the treatment and control group; (3) theoretical consideration of why and what aspect of the campaign caused an impact; are the findings likely to be universalizable?; (4) creation (and, in some cases, incorporation) of recommendations based on combination of impact and theoretical considerations." (Executive summary)