Sorry, its definitely not what I was trying to state - just that absolute numbers matter as much as relative numbers matter - in there own way.
When choosing a surgeon - without the ability to get individual statistics - it would make more sense to go with the one graduating a higher proportion of highly skilled individuals; however, it is possible to get an individuals track record for many types of outcomes because they are part of a public record. In that case, it is easy to identify the 10% minority and may in fact be easier to get one of the surgeons from that school.
In the same sense, it is more likely that you will encounter well written papers by an Indian author IF you read all papers. With this example, I would read English Norwegian papers if I had limited time because they would be higher yield in terms of quality. If I were reading all papers, I would see many high quality papers coming out of India - in part by shear volume. Essentially stating that the good in India is being drowned out by a sea of mediocrity.
On a global scale, the majority of people from India are invisible as an internal colony of the wealthy. The implication of this is that the minority of Indians with a strong education and the ability to perform become representative of the nation and in this manner begin to matter. Its the same reason that China with its emerging middle class matters on a global scale. The much larger poverty stricken group in both countries allow for an exploitable underclass by the rich and middle class.
Being middle class in India means you can afford a staff of people to serve you much as slaves did In the US years ago or immigrants more recently. Its just that in the West, this population has remained low with the exception of nations that had extensive colonies (they just didn't count the large groups of exploited people as citizens). The fact that there are 9 uneducated people for every uneducated one doesn't make the 1/10 educated persons irrelevant.
With this, I think your analogy to a school would become a large university where only 10% are medical students to a medical college where 90% are medical students and perhaps 10% a health related occupation other than physician. From this, it becomes very easy to identify who you would go to for a surgery.