Logo Programming and the Development of Planning Skills. Technical Report No. 16

Year 1: 50 experimental children in two classrooms (25 aged 8-9, 25 aged 11-12) plus same-grade control classrooms in the same school. Year 2: 32 students in each of the two age bands

Children aged 8-9 and 11-12 at a private school in Manhattan.

Two separate year-long longitudinal studies with same-school, same-grade control classrooms, not randomised. Six microcomputers per classroom, so exposure was substantial by the standards of the era. The outcome was a classroom chore-scheduling planning task with both product measures (route efficiency) and process measures (planning decision types and sequencing); the second year used a microcomputer version with a robot to remove objections about task format. This is the empirical core of the Logo-era null — the single most specific transfer prediction Papert made, tested twice, over a year each time.

Process and product measures of planning skill revealed NO benefits for students doing Logo programming. The second, redesigned study a year later reached the same conclusion: learning to program did not differentiate experimental from control group performance. A well-resourced, year-long, twice-run test of the flagship transfer prediction returned nothing both times.

Medium. Classrooms rather than children were assigned, and assignment was not random, so classroom composition is uncontrolled — but the direction of the finding is null, which selection would not typically produce in a school that chose to run Logo.

The second year-long study is an internal direct replication of the first with a redesigned computerised task. Both null.