Post hoc evaluation of analytic rating scales for improved functioning in the assessment of interactive L2 speaking ability

Language Testing in Asia

Table 2 Summary of the category structure for the 9-point interaction rating scale

Category	Observed count	Average measure	Outfit MNSQ	Threshold calibration	Threshold change
1	5^a	−2.49	0.90	−	−
2	10	−0.99	3.40^a	−2.92	−
3	38	−1.34^a	1.20	−2.90	+ 0.02^a
4	27	−0.68	1.10	−0.67	+ 2.23
5	77	−0.16	1.50	−1.54	−0.87^a
6	77	0.52	1.10	0.10	+ 1.64
7	76	1.14	1.30	0.91	+ 0.81^a
8	47	2.98	0.90	2.69	+ 1.78
9	27	4.63	0.80	4.33	+ 1.64

^aSignifies a scoring category that failed to meet one of Linacre’s (2002a) guidelines