The objective of the current investigation is to replicate the results of a previous study on subgrouping of poor readers. One important aim of the study to be replicated was to develop tools, not only for identifying poor readers, which is a rather straightforward task, but also for examining the pattern of poor reading based on a theoretical analysis of the reading process. Eight distinct subgroups of reading performance were identified. The statistical method used was Latent Profile Analysis. The literacy tasks were composed of three kinds of reading material: narrative prose, expository prose and documents, and the major emphasis was placed on comprehension. Low performance of reading comprehension can for some students be due to slow or inaccurate word decoding. Therefore, a speeded word recognition test was added to the battery. Background variables, such as economic and cultural factors were also examined in the different subgroups. The profiles were based on ten variables. The main aspects for the profile interpretation were word recognition, reading speed, comprehension of connected prose and interpretation of brief documents. The present study examines whether the subgroups among poor readers are stable over different samples. Participants in both samples were 9-yearold students from the Swedish part of the IEA 10-Year-Trend Study (the previous study from 2001 and the current study from 1991). Around 5000 students participated each time. The current study is not finalized. However, there are some indications that corresponding subgroups are identifiable.