Application Rasch Model and Extended Angoff Developing Golf Ranking System

Background/Purpose: Golf player rankings in Korean high schools are determined by sum of the frequency of successful driving (FSD) shot, frequency of successful approach (FSA) shot, and frequency of successful iron (FSI) shot. Players have to shot five times at each FSD, FSA, and FSI, so 15-score is the best score. However, evidence based on test content by golf experts suggests the driving, approach, and iron shot have different difficulty. Therefore, the purpose of this study is to develop a more fair and accurate ranking system for Korean high school golf players as employing different weights on the three shots.

Method: A total of 282 male high school golf players in 2013 were collected. The frequency of successful(S) and failed(F) at three shots (driving, approach, iron) was counted. To determine the difficulty of three shots, Rasch analysis with Winsteps 3.62 software (Linacre, 2009) was employed. The model fit criteria was set by Infit and Outfit statistics (≥0.70 and ≤1.30). Based on obtained difficulty score from Rasch analysis, weight scores on the three shots for S and F shots were assigned by five golf experts as employing three rounds of the extended Angoff method.

Analysis/Results: The difficulty of three shuts from Rasch analysis showed acceptable model-data fit. The most difficult shot was to success was approach (logit=1.02) compared to other shots (iron=0.75, & driving=-1.78). Based on the difficulty obtained from Rasch model, weight scores assigned from extended Angoff were as follows: FSA=1.50 vs. FFA (frequency of fail approach)=1.00, FSI & FFI (frequency of fail iron)=1.44, FSD=1.00 vs. FFD=1.50 (frequency of fail driving). Utilizing the weight of the shots, the derived equitation for the player rankings is = [(FSA×1.50)+(FSI×1.44)+FSD]-[FFA+(FFI×1.44)+(FFD×1.50)].

Conclusions: The newly developed high school golf male player ranking system should be validated to another sample (e.g., players in 2014) before applying to real evaluation. In addition, the derived equitation in this study may be able to apply for male players only, but not for female players.