Objective Selecting and optimizing items of Stoke Syndrome Factor Evaluation Scale based on item response theory. Methods Clinical information of 410 cases of stroke was collected. Logistic item response model of two parameters was adopted to calculate item information function of each item respectively

based on the item pool formed in the previous study. Items were selected based on the standard of deleting items with item information function value smaller than 0. 1. Results There were too little amount of items with item information function value greater than or equal to 0. 1 in sub-scale of endogenous wind syndrome. Consequently

its items were deleted on the standard of item information function value smaller than 0. 01. Totally 7 items were deleted

including items of aggravation or unstable condition within the current 48 h

abnormal pupil god（ myosis

mydriasis or anisocoria）

lockjaw

stiff neck

rigidity of the limbs

ankyloglossia

and string-like pulse; One item（ slippery pulse） was deleted in subscale of endogenous fire syndrome according to the selecting standard; Four items（ nausea and vomiting

enlarged tongue or tooth-printed tongue

thick fur

and greasy fur） were deleted in sub-scale of phlegm-dampness syndrome;Two items of numbness of skin and muscle

and uneven pulse were deleted in sub-scale of blood stasis syndrome;Three items of incontinence of urine or feces

moderate pulse

weak pulse or thready pulse were deleted in sub-scale of Qi deficiency syndrome; And four items of tinnitus

soreness and weakness of waist and knees

string-like pulse and thready pulse were deleted in sub-scale of Yin deficiency syndrome. Conclusion The Stroke Syndrome Factor Evaluation Scale（ a preliminary scale） is formed based on item response theory. It contains 76 items

among which there are8 items of endogenous wind syndrome

24 items of endogenous fire syndrome

12 items of phlegm-dampness syndrome

8 items of blood stasis syndrome

15 items of Qi deficiency syndrome

and 9 items of Yin deficiency syndrome.