Prior research demonstrates the existence of gender achievement gaps and the variation in the magnitude of these gaps across states. This paper characterizes the extent to which the variation of gender achievement gaps on standardized tests across the United States can be explained by differing state accountability test formats. A comprehensive analysis of the interplay between state standardized test formats and differences in gender achievement on those tests is important for informing policies and practices that aim for greater equity in education. This study performs both a state-level and district-level analysis, using student test score results in grades 4 and 8 in 2009 from three different tests: (1) state accountability tests (using data from all 50 states and roughly 9,400 school districts); (2) the Measures of Academic Progress (MAP) assessment administered by the Northwest Evaluation Association (NWEA) (using data from roughly 3,700 school districts); and (3) NAEP tests administered (all 50 states). State accountability tests vary in item format among states; the NWEA and NAEP tests have a common item structure across states. For the state-level analyses, the population of study is the set of 48 U.S. states in the 2008-09 school year. For the district-level analyses, the population of study is the set of districts in both the NWEA and EdFacts data, which is approximately 700 districts. In order to understand the effects of test item format on gendered achievement, researchers leverage the variation in the item format across three assessments to model how within-state gender achievement gaps vary with differences in test item format, specifically the proportion of multiple-choice, short constructed-response, and extended-response questions. States vary substantially in the proportion of multiple-choice items on their tests in mathematics and English Language Arts (ELA) (ranging from 50-100% multiple-choice). Findings reveal that boys do better on multiple-choice tests than girls of the same academic skill. Specifically, estimates imply that gender gaps are, on average 0.22 standard deviations (SD) greater (favoring boys more) on multiple-choice tests than on constructed response item tests. These results appear to be driven primarily by gender-by-item format interactions affecting performance on ELA tests: in ELA, gender gaps on multiple-choice tests are roughly 0.30 to 0.40 SD larger (favoring girls less and boys more) than on constructed-response tests. On mathematics tests, the difference in performance is roughly 0.10 SD smaller, but still favoring girls less and boys more on multiple-choice tests than on constructed-response tests. These patterns are consistent regardless of whether NAEP or NWEA tests are used as the audit test. A table and figures are appended.