Table 4.

Terms involved in latent class models, for R = 3 laboratories

Model	Terms of interest	Other terms fitted
1	Test accuracy (A|X, B|X, C|X)	True exposure prevalence (X)
2	Test accuracy, differential by disease (A|DX, B|DX, C|DX)	True exposure prevalence (X)
3	Test accuracy, constant across labs (A|X = B|X = C|X)	True exposure prevalence (X)
4	Test accuracy, constant across labs, differential by disease (A|DX = B|DX = C|DX)	True exposure prevalence (X)
5	True exposure by disease (X|D)	Test accuracy (A|X, B|X, C|X)
6	True exposure by disease (X|D)	Test accuracy, constant across labs (A|X = B|X = C|X)
7	Empirical exposure by disease (A|D, B|D, C|D)
8	Empirical exposure by disease constant across labs (A|D = B|D = C|D)

= : indicates terms constrained to be equal