fortunes.jpg
  • The randomized controlled trial is the 'gold standard' research design for assessing the impact of prevention programs. Drawing on articles from a special edition of the UK-based Journal of Children's Services, a series of stories explores the ins and outs of experimental evaluation.

Randomized Controlled Trials

Randomized trials, certain benefits

13 Oct 2008
As in medicine so in the fields of education and youth justice, randomized controlled trials may be difficult, expensive and, in some fields politically contentious, but the benefits to children and the taxpayer far outweigh the the costs.

Lifting the curse of the squabbling witch doctors

13 Oct 2008
In the battle for reliable program design and evaluation, the US Manpower Demonstration Research Corporation says high-quality policy research is constantly competing against the fatal attraction to politicians of those whose claims for greater success are invariably based on weaker evidence.

Good intention, right design – “wrong” answer?

15 Oct 2008
When a well-made randomized controlled trial proves that a well-liked school tutoring program makes no difference either way to student progress, is there any way back for it? Yes, say US researchers Gary Ritter and Rebecca Maynard, as long as program designers sit up and take notice.

Trialing fidelity among the Welsh faithful

16 Oct 2008
The introduction in Wales of the Incredible Years parenting program is continuing to provide researchers and service managers with insight into the kind of collaborative routines that should help to see evidence-based programs securely into mainstream practice.

Why must they break it before we’ve fixed it?

17 Oct 2008
From their survey of Pennsylvania’s $60m long-term investment in evidence based interventions, Brian Bumbarger and Daniel Perkins found that nearly 40% of the programs had not implemented according to the design specification and in more than half of those cases adaptations had been made that could reasonably be expected to reduce effectiveness.