New video on Jamaican girls using LOGO
The comment by Artemis Papert in this video (at minute 3:40) is key. These girls are learning how to approach a problem and segment it into solvable chunks. Bravo! 
But we are still making major educational decisions on the basis of anecdotal evidence. We have now been using LOGO with kids for over 40 years. Where are the longitudinal studies to tell us what the outcomes have been in these children’s lives? It’s not enough just to notice that kids enjoy the activity, can generate artistic displays or to claim that it “works”. We need evidence of whether learning to code in this way correlates with changes in future education, work and leisure activities. 
The studies must be “goal free”. In other words, the research design cannot be biased so that negative findings are suppressed. Although I am a strong advocate for everyone to learn the rudiments of programming, I still want evidence that indicates that the proposition “there is no relationship between exposure to programming in LOGO and desired educational outcomes” is false.