Hi all
I have a question re: nesting structures...
I have a data set containing the number of tackles (TAC) made by rugby
players in each match (MATCH) played over a season. I have data for each
player (PID) from each team (TID) in a league. (players have not
transferred between teams, so are nested within a single team). There
are 8 teams, and each team plays each other once, giving 28 distinct
matches (dyads?) numbered 1 to 28 e.g. TID=1 vs TID= 2 is match 1, TID =
1 vs TID = 3 is match 2.... TID = 1 vs TID = 8 is match 7, TID = 2 vs
TID = 3 is match 8,... etc
I want to examine how the variation in 'tackles made' breaks down in
terms of player, team and match effects. The data is set up as follows:
TAC | PID | TID | MATCH
15 101 1 1
23 101 1 2
18 101 1 3
35 101 1 4
...
19 102 1 1
13 102 1 2
26 102 1 3
17 102 1 4
...
5 201 2 1
7 201 2 8
My problem is understanding the nesting structure - tackles made are
nested within that player, players are nested within teams, teams within
matches, yet matches are nested within players!
How do I define this type of set up in terms of which level is nested
within/cross-classified with which, so as to partition up the
variability into within player, between player, between team and between
match components
Thanks in advance for any help provided