Longitudinal measurement of growth in vocabulary size using Rasch-based test equating

Language Testing in Asia

Table 2 Descriptive and fit statistics for VST 1–3

Measures	VST 1	VST 2	VST 3
Mean (raw score)	69.2 (44.4%)	79.7 (51.1%)	88.9 (57.0%)
SD (raw score)	12.6	16.9	17.2
Mean Rasch person measure	− 0.27	0.18	0.71
Maximum infit MNSQ	1.13	1.26	1.20
Minimum Infit MNSQ	0.86	0.76	0.86
Maximum outfit MNSQ	2.53	2.45	3.06
Minimum outfit MNSQ	0.54	0.17	0.49
Item reliability	0.98	0.98	0.98
Item separation	7.04	7.26	6.81
Person reliability	0.87	0.93	0.93
Person separation	2.54	3.71	3.57

Note. N = 189; k (number of items) = 156; Rasch measures for each test form were separately calibrated to 0 logits = mean item difficulty