The purpose of this study was to investigate the impact of local item dependence (LID) in passage-based testlets on the test score reliability of an English as a Foreign Language (EFL) reading comprehension test from the perspective of generalizability (G) theory. Definitions and causes of LID in passage-based testlets are reviewed within the frameworks of both classical test theory (CTT) and item response theory (IRT) along with three reliability estimation procedures for testlet-based tests. Scored item responses of 1,072 Korean high school students to a 40-item, 10-passage EFL reading comprehension test were analyzed along with simulated data generated based on the real data by using the computer program GENOVA (G. Crick and R. Brennan, 1983). It was found that the item alpha coefficient from the p x I design was higher than the passage alpha coefficient from the p x T and the generalizability coefficient from the p x (I:H) designs in real data analyses. In simulated data analyses, however, the G-coefficients from the p x I, p x T, and P x (I:H) designed were very close because no LID was assumed within item blocks. Subsequent analyses of the fixed test length scenarios have also revealed that score reliability could be maximized by increasing the number of passages rather than the number of items nested within passages. (Contains 1 table, 3 figures, and 33 references.) (Author/SLD)