One of the challenges for genuine evaluation is striking a balance between being overly bold in statements and overly cautious.  An example of an  analysis that seems to strike this balance is in the Health: Best Treatments blog (a joint project of The British Medical Journal and The Guardian newspaper) of the limitations of recent research that reported:

About 1 in 100 people are ‘addicted’ to using the internet, and these people have a greater risk of becoming depressed.

The analysis discusses the limitations of the study in terms of sampling method (volunteer), sample size (especially of those who were classified as addicted), operationalization of key variables, and  of course the question of which causes which.

Popular reports of the study acknowledged the issue about causation (such as Reuters) but seemed to leave unchallenged the claims about incidence rates.