Question 1: Are the variables (columns) CLUSTER, YEAR and MEASUREMENT populated for every observation (row) in your data? The sample you have provided suggests not?? Question 2: Are you adding the relevant vars (HEIGHT, WEIGHT, INCOME) only for unique individuals at each crossing of CLUSTER, YEAR and MEASUREMENT? For example, in the sample output you've provided, do we only want individuals from cluster=1, year=2000 and measurement=1 that are greater than the specified individual? Question 3: Are you wanting to do a many-to-many analysis (e.g. all unique crossings compared to all other matched crossings), or a many-to-one analysis (e.g. all matched crossings compared to one selected unique observation)?
... View more