Teacher judgments impact on students' academic development, yet studies have revealed only modest judgment accuracy. Identifying biases hampering judgment accuracy is crucial. The present study scrutinized one such bias, namely reference group effects on teacher judgments (i.e., systematic relations between class-average ability and teacher judgments of individual students' abilities). The study investigated both direction and size of reference group effects on teacher ratings of students' reading comprehension, mathematical ability, and cognitive ability; grades in German (mother tongue) and math; and tracking recommendation for secondary schooling. Teachers' professional experience and acquaintance with the class were examined as moderators of reference group effects. Data from N = 837 fourth graders (M[subscript age] = 9.14, SD = 0.54) and their N = 56 class teachers were analyzed. Multilevel structural equation models revealed negative reference group (contrast) effects on teacher ratings of mathematical ability ([beta] = -0.90), reading comprehension (-0.62), cognitive ability (-0.42), and on math grade (-0.61). Contrast effects were weak for grade in German (-0.13) and tracking recommendation (0.01 [less than or equal to] [beta] [less than or equal to] -0.28). For reading comprehension, cognitive ability, and grade in German, teacher's acquaintance with the class significantly moderated the reference group effects: Contrast effects emerged only in teachers who had taught their students for at least 3 years. For teachers who had taught their students for about 1.5 years or less, assimilation effects occurred (i.e., a positive relation between class-average ability and individual student assessment). Implications of the findings and approaches to intervention are discussed.