This study evaluated students' representational choices while they solved linear function problems. Eighty-six secondary-school students solved problems under one choice condition, where they chose a table, a formula, or both to solve each problem, and two no-choice conditions, where one of these representations was forced upon them. Two conceptualisations of representational flexibility were used: a groupwise conceptualisation, where group-based measures of students' fluency with the different representations were used to determine which representational choices could be considered flexible, and an individualised conceptualisation, where each individual student's fluency with each of the representations to solve each problem type was taken into account when determining which choices could be considered flexible for that particular student. A strong correlation between groupwise flexibility and choice condition accuracy, and an even stronger correlation between individualised flexibility and choice condition accuracy were found. The implications for research and instruction are discussed.