This survey design study was designed to test whether the graph skills and affective states of middle school students about graphs differ by their gender, grade level, and graph types (line, bar, and pie). The data collection instruments consisted of two scales developed by the authors and a Graph Skills Test, which consisted of graph questions from the previous TIMSS and PISA exams. Based on the findings, while middle school students were found to succeed at reading the data level graph questions, they were found to struggle in questions requiring higher graph skills, such as graph interpretation and graph construction. As for the affective states investigated, participants were found to hold high self-efficacy beliefs and positive attitudes toward graphs. No significant difference among the dependent variables (graph skills, self-efficacy beliefs about graphs, attitudes toward graphs, and graph literacy perceptions) was found by gender; however, grade level and graph type variables were found to impact students' graph skills, graph attitudes, and personal graph literacy perceptions. Middle school students with less school experience with graphs (seventh graders) were found to hold more positive attitudes toward graphs than the eighth graders. On the contrary, eighth graders were found to perform better at graph questions requiring interpretations of the graph data. Also, participants in all subgroups were found to hold significantly higher personal graph literacy perceptions for the bar graphs, than the line graphs and pie charts. Based on the findings of the study, while middle school students were found to hold positive affective states about graphs, they were found to lack advanced graph skills. In agreement with the previous literature, it is recommended that graph literacy should become a dedicated part of the school curriculum.