The "most recent cohorts" has earnings for cohorts at 6, 7, 8, 9, and 10 years after entry. Questions:
The documentation appears to indicate that these are earnings for different cohorts with different entry years, not earnings at different points in time for the same cohort. Is that correct?
According to the data dictionary, "count_ed" variable is "count of students in the earnings cohort." But since there are 6 earnings cohorts, which cohort is it for?
The "count_ed" from the raw data file merged_2005_PP appears to match the count_ed in the most recent cohort file. This makes it seem like the count_ed in the most recent cohorts data is for the cohort entering in 2005, and Ns for other cohorts might be obtained by looking in later raw data files. However, the count_ed field is NULL in all the merged_YYYY_PP where YYYY>2005. Is there another way to get cohort sizes for the earnings data for any cohort other than the 10 year?