I have data from a memory test. I created a table with the absolute number of hits, misses, false alarms, and correct rejections. Each participant had 10 targets and 10 non-targets at each test (there are three tests).
Now, I have observations with 10 out of 10 hits, and others with 0 out of 10 false alarms.
I read Hautus (1995) but I am still confused about how to carry out the log-linear method. First, it says I have to add 0.5 to all the cells, but I did not understand if I have to add to the absolute number (e.g., 10 + 0.5) or to the proportion (1 + .5).
Then it says that I need to add 1 to the number of signal trials and 1 to the number of noise trials. If each row has the number of hits, misses, false alarms, and correct rejections per participant, do I need to add 1 to all four observations? Or add 1 to the hits and 1 to the false alarms?
Wouldn't that be the same as just adding 1.5 to each cell of hits and 1.5 to each cell with false alarms?
Here is a bit of code to show you how my data frame looks:
structure(list(id = c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 4L,
4L, 4L, 5L, 5L, 5L, 6L, 6L, 6L, 7L, 7L, 7L, 8L, 8L, 8L, 9L, 9L,
9L), time = c("immediate", "hour", "day", "immediate", "hour",
"day", "immediate", "hour", "day", "immediate", "hour", "day",
"immediate", "hour", "day", "immediate", "hour", "day", "immediate",
"hour", "day", "immediate", "hour", "day", "immediate", "hour",
"day"), group = c("older", "older", "older", "younger", "younger",
"younger", "older", "older", "older", "older", "older", "older",
"older", "older", "older", "younger", "younger", "younger", "younger",
"younger", "younger", "younger", "younger", "younger", "younger",
"younger", "younger"), n_hit = c(9L, 8L, 8L, 8L, 7L, 6L, 9L,
9L, 6L, 9L, 7L, 8L, 10L, 5L, 6L, 7L, 5L, 5L, 6L, 5L, 5L, 7L,
7L, 7L, 7L, 5L, 6L), n_miss = c(1L, 2L, 2L, 2L, 3L, 4L, 1L, 1L,
4L, 1L, 3L, 2L, 0L, 5L, 4L, 3L, 5L, 5L, 4L, 5L, 5L, 3L, 3L, 3L,
3L, 5L, 4L), n_cr = c(9L, 9L, 10L, 9L, 8L, 6L, 8L, 8L, 6L, 9L,
9L, 10L, 8L, 7L, 8L, 9L, 5L, 5L, 7L, 8L, 6L, 5L, 8L, 6L, 8L,
8L, 7L), n_fa = c(1L, 1L, 0L, 1L, 2L, 4L, 2L, 2L, 4L, 1L, 1L,
0L, 2L, 3L, 2L, 1L, 5L, 5L, 3L, 2L, 4L, 5L, 2L, 4L, 2L, 2L, 3L
)), class = "data.frame", row.names = c(NA, -27L))
If it is of any use, I am calculating d prime using psycho::dprime.
