Job‐search assistance using the internet: experiences from a Swedish randomised experiment

Purpose – The purpose of this paper is to investigate to what extent the evaluation design in this study manages to avoid common difficulties in experimental evaluation, and to assess the effect of an internet‐based job‐search assistance programme offered at the Swedish employment offices on employment outcome. Design/methodology/approach – In 2002, the Swedish Labour Market Board initiated a nation‐wide demonstration programme investigating the possibility of pursuing voluntary job‐search club activities on the internet. Participants were randomly selected from a group of voluntary job seekers registered at the employment offices. Findings – Many experiment‐related problems such as ethical concerns, bureaucratic resistance and randomisation bias were circumvented. With the services being voluntary, however, a considerable fraction of the applicants either failed to show up or dropped out early in the process. Neither the intent‐to‐treat impact estimate, estimating the effect of being offered the services, nor the treatment‐on‐the‐treated estimate, estimating the effect of actual treatment, shows any effect on subsequent job transitions. Practical implications – With fewer than 1,000 showing interest in the services, and with the large portions of no‐shows and dropouts among the applicants, it is likely that the Swedish Labour Market Board misjudged the clients' interest in this type of internet service. Originality/value – By using random assignment to these programme services, the study contributes to the sparse literature on experimental evaluation of labour market topics in Europe in general, and in Sweden in particular.