The purposes of this study were to assess the usefulness of a variety of readability formulas in predicting the relative difficulty of passages, and to explore the contribution of pupils' background to text difficulty. Subjects were 285 special education students in grades 1-9, 117 of whom were based in rural and suburban Minnesota (MN) and 168 of whom lived in New York City (NYC). Each student was tested twice on three passages of a Passage Reading Test (PRT). Analyses of variance applied to the data revealed an interaction between passages and student site (MN vs NYC), suggesting that text difficulty was affected by students' background. Additionally, six readability formulas were applied to the three PRT passages. There was no agreement between passage difficulty rankings derived from the formulas and those based on students' actual performance. Implications for creating readability formulas are discussed. (Author)