This paper reports on the main findings from a cross- national collection of 19 papers reporting secondary analyses researchers from sixteen countries who utilized all data from TIMSS 1995 and/or 1999. There are many factors on school, class and student levels that have positive and negative effects on mathematics achievement (see Howie, 2002, for an extensive review of the literature). All papers address questions on what background and contextual factors are related to students' achievement in mathematics and sciences. In the analyses reported in the 19 papers two different approaches can be distinguished, which can be characterized by the metaphors of 'fishing' and 'hunting'. In the 'fishing' approach, researchers in principle begin with an open mind, considering all possible context variables as possible explanatory variables. Applying analysis techniques like regression analysis, Lisrel, PLS (partial least square analysis), HLM (hierarchical linear modelling), and MLn (multi-level modelling), they try to identify which factors within their countries or across a number of countries explain (to a certain extent) mathematics and/or science achievement. These chapters are labelled 'Background variables and achievement', with seven of them focusing on within country analysis and three focusing on across country analysis. In the 'hunting' approach, researchers have reasons to choose certain context variables upfront and investigate their relationships with achievement. Five authors discuss a number of curriculum related variables in relation to achievement, whereas in four papers the emphasis is on teaching aspects, classroom processes and/or school management variables and their effect upon achievement in several countries. All the secondary analyses utilise TIMSS 1995 and/or 1999 data. Whilst most of the chapters (viz ten) focus on mathematics achievement only, there are three studies on science achievement specifically; six of the chapters include both mathematics and science achievement.