The papers were divided into short stories and non-fiction pieces, each 400 to 700 words long. Children were either given one of each type, or two of the same type.
Papers consisted of a mix of multiple-choice questions and questions which required written answers.
Many countries performed significantly better in one type of reading.
English-speaking countries scored more highly on reading stories, whereas France did better at factual reading.
Even in Canada, this was true, with the English-speaking province of Ontario doing significantly better reading fiction than fact, a pattern reversed in French-speaking Quebec.
But the highest performing nations, Sweden and Holland, did equally well at both.