As school systems around the world continue trying to recover from learning losses caused by the pandemic, a new paper out of Germany gives us some historic context on what to look for and what to be wary of.
Researcher Kamila Cygan-Rehm from the Dresden University of Technology exploits a natural experiment that occurred when two school years were shortened in several West German states in the late 1960s. These were the result of a realignment of school start dates following several previous shifts during and immediately after World War II. Specifically, the two school years that ran from April to November 1966 and from December 1966 to July 1967 were each truncated by about one-third in order to ultimately move all schools’ starting dates from spring to fall. This shortened the amount of schooling received by all children attending school in ten states—excluding Bavaria, Hamburg, and Lower Saxony, which made the switch in other ways.
Although accommodations for the decreased instructional time varied from state to state, Cygan-Rehm notes important similarities. Emphasis was placed on maintaining the core curriculum, with priority given to mathematics, German, and modern foreign languages (mainly English). The weekly hours of instruction in these subjects increased when compared to a typical school year, as did the homework load. Consequently, instructional hours in non-core subjects like music, arts, and physical education were reduced. Extracurricular activities were not traditionally part of the school day, but certain annual school-based events like Christmas pageants and school trips—all of which took time away from instruction—were canceled until the new start dates were achieved. Many states also reduced the number of in-class tests and the requirements for final examinations to make up for the loss of instructional time. This effort, we are told, resulted at the time in massive outpourings of concern from academics, families, government, and media over the potential loss of student learning. However, short-term outcomes for students were not seen to be substantially negative, and thus no other remediation was taken when school years returned to the normal length. Yet students who graduated in either of those years did so having lost up to two-thirds of a year of instructional time.
This recent analysis, like those conducted in the 1960s, sees minimal short-term impact. Graduation and university enrollment did not change significantly in affected states from pre-1966/67 rates. Long-term outcomes, however, are a different story. Cygan-Rehm uses data from the Sample of Integrated Labor Market Biographies, a rich database of German employment data, as well as pension records, to identify over 278,000 individuals who were of compulsory school age during the 1966–1967 timeframe in the various states and whose work and earnings history are complete through 2017. While she cannot definitively state that her sample includes only workers who experienced shortened school years, she uses the same data from two of the unaffected states as a comparison and robustness check. As a result, she is satisfied that she has captured a large swath of impacted former students.
Her analysis finds that exposure to the shortened school years led to adverse labor market effects over almost the entirety of an individual’s working life. She estimates that one year of lost classroom instruction reduced lifetime earnings by almost 3 percent on average. The earnings losses were partly driven by higher rates of unemployment among affected individuals during the prime working ages (a 2 percent reduction in days worked) and partly by lower intelligence levels (yes, cognitive test scores were included in the data!) for individuals who ultimately received less instruction than their peers before and after the calendar changes. All this despite the fact that many students who experienced the shortened school years ended up graduating earlier in the year than they would have normally and actually experienced an initial earnings bump based on early entry into the workforce. The negative impacts of lost learning on lifetime income took hold shortly after that bump and persisted until retirement. The youngest and oldest students were impacted more strongly than those in the middle of the grade-level distribution. The largest earnings losses occurred at the bottom of the earnings distribution, suggesting that academically-disadvantaged students, already struggling while receiving a full year’s worth of instruction, fell even further behind in the shortened years.
What should we take from these historical data? Despite clear differences in context and setting, the West German example shares some obvious similarities to our current Covid-disrupted era: closed schools being replaced with a form of “remote learning” (extra homework), loss of instructional time being compensated for by narrowing the curriculum, and impacted students being given dispensation on work completion and assessment. All of these combined to lower the academic bar for students who moved on through the system with no accountability beyond the dates on the calendar. However, it is important to note one vital distinction: No remediation was provided for West German students in the 1960s. We know better now—and these new data show the dangers clearly—and we should expect better results. For those impacted students who remain in school today, we know what needs to be done to help avoid a repeat of history. And our schools still have time to do it.
SOURCE: Kamila Cygan-Rehm, “Lifetime Consequences of Lost Instructional Time in the Classroom: Evidence from Shortened School Years,” IZA Institute of Labor Economics (August 2024).
