Musculoskeletal and neurologic injuries affecting shoulder strength are common in contact sports. Full-strength recovery is desired before resumption of competition. On-field assessment of shoulder strength is usually done by manual muscle testing, which lacks sensitivity and reliability. Our objective was to determine the reliability and validity of two field instruments capable of quantifying shoulder abduction strength. Twenty junior football players underwent bilateral isokinetic (60 degrees/s) and isometric shoulder abduction strength measurements using a Cybex 340 isokinetic dynamometer. Test-retest measurements of both shoulders of each player were made using strain gauge (SG) and handheld dynamometer (HHD) instruments. Players were tested during rested and competition conditions. Within and between session reliabilities were calculated using the intraclass coefficient, and validity was assessed using Pearson's correlation coefficient. Overall reliability for each device was calculated using Lisrel analysis. SG was found to be superior to HHD in overall reliability and validity. Within-session reliability in the rested and competition states was 0.75 and 0.78, respectively, for SG and 0.60 and 0.81, respectively, for HHD. Between-session reliability in the rested and competition states dropped to 0.51 and 0.63, respectively, for SG and 0.55 and 0.70, respectively, for HHD. Validity was 0.41 and 0.70 for SG when correlated with Cybex at 0 degree and 60 degrees/s respectively. Validity for HHD was 0.28 and 0.42 for Cybex speeds of 0 degree and 60 degrees/s, respectively. SG reliability and validity were similar when testing was done one shoulder at a time or both shoulders concurrently.(ABSTRACT TRUNCATED AT 250 WORDS)