It is not uncommon for very young children to start music instruction on string instruments. Previous studies have examined the relationship between starting age of formal music instruction and years of study (Duke, Flowers & Wolfe, 1997; Hartley, 1996; Hartley & Porter, 2009). Duke et al. (1997) found that students who took more years of piano lessons had started instruction at an earlier age than those with fewer years of lessons. In contrast, an earlier study by Hartley (1996) comparing student retention in school music programs, showed no significant difference in the duration of band participation between those who had started lessons in 5th or 6th grade. In other words, an earlier beginning did not yield higher retention rates. Similarly, Kruth (1964) found that the starting grade level was not a significant predictor of dropout rates in secondary school instrumental programs. Contradicting the findings of these previous studies, Hartley and Porter (2009) found that a late start in string instruction was associated with higher retention rates in school orchestra programs. In summary, the results regarding the relationship between starting age of instrumental instruction are conflicting and inconclusive. The purpose of the present study is to investigate this relationship further. The authors completed the study with children attending an out-of-school string program associated with a large university. Approximately 250 students aged 4 to 18 enroll in this program and participate in private lessons, musicianship classes, and string ensembles. Two questions guided the study: (1) Is there a relationship between the starting age of music lessons and years of participation in the program?; and (2) Is there a relationship between selected student demographic characteristics and retention in string programs? The subjects included in this study were 99 students (female=52, male=47) aged 4 to 16 who registered in the string program at any time between 1998 and 2002 and dropped out within 10 years. Thirty-nine were violin students (female=23, male=16), 17 were viola students (female=10, male=7), 33 studied cello (female=16, male=17), and 10 studied bass (female=3, male=7). The authors calculated the starting age of music instruction for each student and established the correlation between this variable and duration of instruction (i.e., number of semester of participation in the program). The correlation between the starting age and years of study was low but significant r = -0.3498, p = 0.0004. Students who started earlier stayed in the program longer. The authors also calculated the correlation between starting age and years of study separately for female and male students. They found no significant correlation for female students but a moderate correlation for male students r = -0.4698, p = 0.0009 (see Figure 2). Finally, we established correlations between the same variables but for each of the four instruments separately. Only for cello students was the correlation significant r = -0.5064, p = 0.003. Students who started lessons on cello earlier, remained in the program longer than those who started later. Overall, the results of the study provide some evidence supporting the belief that an early start of music lessons may be advantageous for children; starting lessons at an early age may extend involvement in formal music instruction. [Paper presented at the Annual Meeting of the Texas Music Educators Association (San Antonio, TX, Feb 2013).]