I am working with the results of a test with 14 items, all of which are dichotomous (1 - right answer, 0 - wrong). Is it possible to use omega to estimate reliability if I have dichotomous items? Why does psych::alpha give a different coefficient of Alpha than psych::omega? In psych::alpha, it is Cronbach's alpha, and in psych::omega, it should be too, but there are different coefficients (0.64 and 0.75).
psych::omega(var1, poly=TRUE)
Alpha: 0.75
G.6: 0.77
Omega Hierarchical: 0.49
Omega H asymptotic: 0.63
Omega Total 0.78
psych::alpha(x = var1)
raw_alpha std.alpha G6(smc)
0.64 0.64 0.63