The present study investigated whether two summated rating scales used in the Trends in International Mathematics and Science Study (TIMSS 2003), Mathematics Self-Concept (MSC), and Students' Valuing of Mathematics (VoM) seemed appropriate, meaningful, and useful in a Swedish context. The 4,256 Swedish eighth-grade students involved in TIMSS 2003 participated in the study. Items in the TIMSS Student Background Questionnaire asking for motivational beliefs were subjected to item analyses, principal components analyses, and confirmatory factor analyses. Factor scores were saved and regressed on mathematics score. The results indicated that the MSC scale was internally consistent, showed a simple factor structure that fitted the data rather well, and was positively related to mathematics achievement. The VoM scale was unrelated to mathematics score, showed a poor model-data fit, and the items in the scale were rather heterogeneous. The findings are discussed from an expectancy-value perspective and suggestions for researchers performing secondary analyses on TIMSS data are presented. (Contains 2 tables.)