BACKGROUND: Despite increased awareness of factors related to athletic performance and injury prevention, youth and adolescent baseball players continue to report injuries at alarming rates. Upper extremity muscle strength is an integral part of physical assessment and injury prevention in baseball players, however minimal data exists in youth populations. Changes in anthropometric measures, inherent in physically developing athletes, have been shown to impact strength measures, however normalization methodology is rarely reported. PURPOSE: The purposes of this study were to 1) compare the measurement properties of five potential methods for normalizing isometric shoulder strength in a cohort of 9-12 year old male baseball players and 2) examine the relationship between normalized isometric shoulder strength and ball velocity in a cohort of 9-12 year old male baseball players. STUDY DESIGN: Prospective cohort study (n=159). METHODS: Baseline and follow up height, weight and bilateral ulnar length measurements were assessed followed by isometric strength in both the dominant and non-dominant shoulders. Strength measures included scapular plane abduction (scaption), external rotation (ER) at 0°, ER and internal rotation (IR) at 90°. Ball velocity was assessed as a measure of throwing performance. Intraclass correlation coefficients (ICC2,1), standard errors of measurement (SEM) and minimal detectable change (MDC95) were calculated for all strength measures. Repeated measures ANOVA were conducted comparing changes in normalized strength using five separate anthropometric measures: weight, height, body mass index, ulnar length and % of non-dominant shoulder strength. Linear regression models were used to examine the relationships between normalized isometric shoulder strength and ball velocity. Statistical significance was set a priori at α=0.05. RESULTS: Shoulder strength normalized using ulnar length was the only method that demonstrated excellent reliability (ICC2,1 0.98-0.99) and detected significant changes between strength in each of the four measures tested (SEM 0.39-0.69 Nm). Modest but significant correlations were observed between scaption and ball velocity (r2 = 0.27, p < 0.001) and ER at 0° and ball velocity (r2 = 0.23, p < 0.001). CONCLUSION: Ulnar length was the most stable and reliable normalization method for assessing isometric shoulder strength in youth baseball players. In addition, normalized scaption strength was the most significant predictor of ball velocity, followed by ER at 0° strength in this population. LEVEL OF EVIDENCE: Level 2b (etiology).