The purpose of this study is to examine relationships among the such variables as socioeconomic status, math self-efficacy, anxiety, and mathematics achievement using structural equation modeling. A sample group of 8,806 students from England, Greece, Hong Kong, the Netherlands, Turkey, and the USA participated in the PISA 2012. To show how much variance on mathematics achievement can be attributed to the selected variables, separate structural equation modeling analyses were examined for each country. Next, multi-group structural equating modeling was determined to compare latent means. The results show that socioeconomic status has a significant effect on mathematics achievement. The relationship between socioeconomic status and mathematics achievement is highest in the Netherlands lowest in Hong Kong. For all six countries, the most important predictor of mathematics achievement is math self-efficacy. The relationship between math self-efficacy and mathematics achievement is highest for England. No statistically meaningful relationship between math anxiety and mathematics achievement is found for Hong Kong, England, or the Netherlands. A statistically meaningful relationship between math anxiety and mathematics achievement is observed in the countries that have low mathematics achievement; namely, Turkey, Greece, and the USA.