Science, technology, engineering, and mathematics (STEM) programs serving students from low socioeconomic areas are understudied in the literature. More research studies need to be conducted to make informed instructional decisions for students who may be at a disadvantage compared to their peers from higher socioeconomic areas. The purpose of this research study was to determine the effects of traditional science instruction and blended learning on STEM achievement of elementary school students from low socioeconomic areas. Third, fourth and fifth grade students (N = 129) from a low-socioeconomic school were randomly assigned to receive traditional science instruction or a blended learning science curriculum approach. The science, technology, engineering, and mathematics (STEM) achievement scores were analyzed by conducting a one-way two-group Multiple Analysis of Variance (MANOVA) implemented in R statistical computing platform (R Core Team, 2018). The results indicated that the teaching method had a statistically significant effect on the linear combination of the science, technology, mathematics and engineering scores (F(4,124) = 80.27, p < .0001, Pillai’s Trace = .721, partial η2 = .721), in favor of the blended learning approach.