On cheating in Indian education, India’s economic woes, understanding Chinese education, and the Opium Wars

Interesting Links: September 6, 2020

1. Cheating in Indian schools - Great paper on how school-reported levels of student learning in India can be highly inflated using Madhya Pradesh data as an example. The author’s website also lists links to the working paper in case the link above stops working. Karthik Muralidharan recently tweeted about this paper.

   The paper focuses on the state of Madhya Pradesh (MP), one of the largest states in India with a population of over 80 million, and the test “Pratibha Parv” which is an annual diagnostic test administered to all students from Grades 1-8 in public schools (~7 million students). Importantly, this test is not linked to student grade progression, teacher salaries or promotion, or other such incentives.

   The basic methodology was that the researchers sample 283 primary and middle schools in 10 districts, and had the students retake the exact same test a month later, which was independently proctored and graded.

   Punchline: the researchers find that the ‘official’ school test results were more than twice as high as the correct test results (as determined by the independently proctored retest). The correct test results averaged 25.1% in Math and 37.9% in Hindi. However, the school reported results were 38.9% higher on average in Math and 33.8% higher in Hindi.

   These inflated scores are found to reflect a combination of student copying and teacher-assisted manipulation of scores. Interestingly, in schools which used multiple booklets and external grading (to combat such copying/cheating) the inflation is reduced by one-half in Math and three-quarters in Hindi.

   Possible solutions reported: Digital testing which makes student copying harder (by providing different questions to students) and automatic grading (removes the discretion of individual teachers to inflate grades).

   Digital testing was studied using the following field experiment: 2,400 schools in one district in Andhra Pradesh (AP) were used. AP is also a fairly large state (over 50 million people). However, note that AP is considerably richer, with over 1.5x the GDP per capita of MP, to the extent that this might have any bearing on comparing outcomes across the two states.

   In the AP field experiment, schools were divided into two treatment arms: one benchmark case (768 schools) administered paper-based tests (using multiple test booklets to curb student copying) and a treatment arm (1,694 schools) administering digital tablet-based tests. And then a retest was done later, similar to the procedure used in MP.

   In AP, the researchers flag 38-43% of the schools in the paper-based testing arm for cheating. And on average, students scored 16-20% higher in the earlier teacher-administered tests than in the independently-proctored retests. But in contrast, only 2-5% of the tablet-based testing schools were flagged for cheating. And the difference in scores between the earlier teacher-administered tests and the later independently-proctored retests was not significant.

2. I’d be remiss if I didn’t mention India’s massive GDP contraction in the Mar-Jun quarter of a gazillion percent. You can read all about it elsewhere, but on the heels of the article above about mismeasurement, it’s time to revisit Arvind Subramnian’s article about inflated GDP figures. Finception has a decent summary of it. As they note:

   Arvind Subramanian’s argument throughout the paper is that our measurement of real GDP is suspect because our knob is broken i.e. the deflators we use to convert nominal GDP to real GDP is misleading us into thinking we are growing at 7% when we are not. In fact, there is almost no mention of misestimation of nominal GDP in the paper and AS leaves this exercise to future researchers who might want to dig deeper.

   Similarly, Finshots writes about the GST debacle.

3. And while we’re on the topic of the disaster of the Indian education system, check out this great article reviewing a book about the Chinese education system. This quote for example:

   Lenora even witnesses Rainey’s teachers cheating, and not even at something that matters. Lenora and her husband attend a day-long athletic competition between all the classes in Soong Qing Ling. In one round, fathers need to pass the three-year-old students through their legs in a big line, like throwing children through a tunnel. In blatant violation of the rules, one teacher starts crawling into the leg-tunnel and pulling students through. Lenora watched in sheer incredulousness as this teacher and all the fathers triumphantly celebrated their victory, in an athletic competition for three-year-olds, at which they blatantly cheated.

4. And while we’re admiring the Dormin blog, do check out the post on the Opium War. The author discusses a book I’ve been meaning to read, Imperial Twilight.

   It’s a fantastic read, from the decline of the Qing dynasty to the trade policy to the Opium Wars. And as with everything in modern history so far - “According to Platt, the greatest beneficiaries of the war were… the Americans.” And an extremely depressing ending:

   According to Platt, the idea that the Opium War represents the start of China’s “Century of Humiliation,” where the once-mighty country was humiliated by Western dogs, was an invention of later Chinese nationalists. At the time, the Treaty of Nanking was undoubtedly a blow to Chinese dignity, but the Qing Dynasty knew China had been on the decline for centuries. Losing a war to a handful of British ships was just everybody admitting that the emperor had no clothes.

   Meanwhile, the opium epidemic got worse. Brits and Americans continued to pump it in at ever-higher rates until the emperor threw up his hands and legalized usage in the middle of the 19th century. According to Platt, this just threw fuel on the fire. Instead of importing opium, the Chinese began to grow it themselves. Supply ballooned and prices plummeted. By the end of the century, the Chinese were growing 10X more opium than they ever imported from foreigners.

5. Modelling the Human Trajectory. Very interesting read. Perhaps all throughout history (well, written history at least), mankind has thought “this time is different.” But maybe this time really is different. Or maybe this is just another attempt by yet another trigger happy aspiring applied mathematician into the world of time series prediction.