Evaluation of the Massachusetts Expanded Learning Time (ELT) Initiative: Final Study Findings

24 Expanded Learning Time schools against 25 matched comparison schools, over five years of implementation

Massachusetts elementary, middle and K-8 schools, predominantly under-performing with at least 50% low-income enrolment

SREE 2013 conference paper reporting the findings of Abt Associates' Year Five Final Report; ERIC ED563107. NOT an RCT - a comparative interrupted time series against matched comparison schools, hence grade C. The treatment is the largest dose increase measured anywhere in this topic: AT LEAST 300 ADDITIONAL INSTRUCTIONAL HOURS PER YEAR, sustained for five years. Outcome is MCAS, the state standardized assessment. The authors state the limits plainly: assignment was not random, 'results cannot be attributed solely to ELT', and 'this study was able to assess the quantity and allocation of time, but did not measure the quality of instruction'.

'Generally, there were no statistically significant effects of ELT on student achievement. There were no statistically significant effects of ELT in the first, second, or third years of implementation on student achievement, but in the fourth year of implementation 5th grade ELT students performed significantly higher on science tests than would be expected in the absence of ELT.' One significant result, in one subject, in one grade, in one of five years, from adding 300 hours annually. There were also 'no clear patterns or meaningful relationships' between implementation quality and outcomes, and 'minimal heterogeneity in the effect by the level of ELT implementation' - so the null is not explained by poor implementation. THE COST SIDE IS MEASURED AND IS RARELY REPORTED ELSEWHERE: more teachers in ELT schools reported that 'teacher and staff fatigue, as well as student fatigue, were problems', and 'more ELT students reported being tired in school and fewer ELT students looked forward to school, liked being in school, reported that all of their classes were important to them, and liked the length of their school day'. A large dose of extra time bought no learning and cost engagement.

Medium. Schools volunteered for the initiative and were then matched to comparison schools on observables, so selection into treatment is at the school level and only partially addressed; student composition could also shift over five years. The direction of any residual bias would favour ELT, since motivated schools volunteered, which makes the null more rather than less informative.

The large-dose null replicates. A Dutch randomised trial adding five weekly hours across seven elementary schools found 'no significant effect on math or language achievement'. Chile's national move to a full school day, roughly 27% more annual instructional time, bought 0.05-0.07 SD. Florida's extra literacy hour produced 0.05 SD that decayed to -0.004 SD within three years. Four settings, four large dose increases, effects between zero and 0.07 SD.