The collection and analysis of relevant data for evaluating public policies is not a straightforward task. An important type of such studies is the so-called ex-post evaluation. The main objective of ex-post evaluations is to determine to what extent a realized intervention is successful in tackling a societal challenge, e.g., youth unemployment. At a first glance an obvious method is to collect some baseline measurements for a set of relevant variables, apply the intervention for a while and collect the new measurement values for the same set of variables. Then, comparing the measurement values of the variables before and after the intervention provides an insight into the extent of successfulness of the intervention. This, however, is only true if the "ceteris paribus" condition holds. In practice it is infeasible to enforce this condition for societal challenges. Often, after having the baseline measurements, several phenomena emerge that may impact the new measurements without being taken into account. This makes it difficult to determine how much of the measured differences between the values of the variables before and after the intervention should be attributed to the emerging phenomena (or the so-called counterfactuals) and how much of the differences can be attributed to the applied intervention.
This paper discusses how exploiting big data may contribute to the task of elucidating the influences of counterfactuals (and interventions) in ex-post evaluation studies. The paper proposes a framework to utilize big data for accounting for the impact of emerging phenomena in ex-post evaluation studies.