Two internationally comparative studies of reading were conducted by the International Association for the Evaluation of Educational Achievement (IEA), namely the Reading Comprehension Study in 1970/71 and the Reading Literacy Study in 1990/91. Eight educational systems participated in both studies at the 14- year-old level, namely Belgium (French), Finland, Hungary, Italy, The Netherlands, New Zealand, Sweden, and the United States of America. The information on student, teacher, and school variables that were common to both studies provided the core evidence used for analysis in this thesis. The major results are described. A general factor, reasoning, could be identified as pertaining to the processes involved in reading comprehension at the 14-year- old level. The items that were administered on both occasions could be brought to a common scale using the one- parameter Rasch model. While the reading performance of students in Finland and Hungary improved between 1970/71 and 1990/91, students in Belgium (French), Italy, The Netherlands, New Zealand and the United States performed at a lower level in the Reading Literacy than in the Reading Comprehension Study. The performance of Swedish students remained stable over time. The two-level HLM analyses indicated that the average proportion of variance across countries at each level did not change greatly between 1970/71 and 1990/91 with approximately three-quarters of the variance occurring at the between-student level and one quarter of the variance at the between-group level. The three- level HLM analyses showed that 14-year-old students from Finland, Italy, and New Zealand achieved at a significantly above cross- country average while Hungary and the United States performed below average in 1970/71. In the three-level analysis of the Reading Literacy data, Finland emerged as the only country achieving above the grand mean while students in Belgium (French), The Netherlands, and Sweden performed below the cross- country average in 1990/91. It is important to note that these differences were observed after other variables in the three- level model such as socio- economic background of schools and reading materials in students' homes had been taken into consideration.