That doesn't come close to explaining it.
If the number of students killed remained constant and the population increased -- yes, students killed
per million would go down. You are absolutely right. It would also decrease if the population increased faster than shooting deaths.
But even if we assume the most conservative case, that the raw # of deaths/year remains constant while population doesn't, the population increase alone doesn't explain the drop. Here's why:
The population of children under age 18 in the United States in 1990 was 64.2 million and 74.1 million in 2010.
https://www.childtrends.org/indicators/number-of-children/
That website only gives projections after 2010 but let's be ultra conservative and use their 2050 estimate for the end date of the study, which is 2015, and their 1990 stat for 1992, which is the first year of the study.
Then .57 deaths per million when there were 64.2 million is .57*64.2 which rounds up to 37 deaths. 37 deaths when there are 80 million is 37/80 = .46 deaths per million. Now look at the chart with this number in mind.
The raw number of incidents is also down (see chart before deaths/million).