The issue that this study addresses is the need for secondary students to work on relevantrnnon-routine statistical problems to enhance their understanding of descriptive statistics.rnThe purpose of this study is to explain how secondary students can enhance theirrnunderstanding of descriptive statistics using Modeling Approach and to compare thernamount of learning achieved using Modeling and Non-modeling InstructionalrnApproaches. The Modeling Approach integrated the Contextual and Socio-criticalrnModeling Perspectives to teach a unit of descriptive statistics in secondary mathematicsrncurriculum. The leading research question set forth is, to what extent and how do highrnschool students enhance their understanding of descriptive statistics using a ModelingrnApproach?rnThis transformative embedded mixed method study looked into three facets of students’rnunderstanding of descriptive statistics: procedural, conceptual and critical. The study wasrnconducted at two high schools in urban district in West Shewa Zone, Ethiopia. Forrncomparisons, four Grade Nine sections with a total of 163 students were selected byrnpurposive sampling technique. An instrument called Teaching for Robust Understandingrnin Mathematics (TRU Math) was used for observing Model-eliciting Activities (MEAs)rnto capture the Modeling Approach classroom practice (Schoenfeld, 2013). QualityrnAssurance Guide instrument was used to assess students’ models on MEAs (Lesh et al.,rnivrn2000). Standard questions were used for achievement tests on procedural and conceptualrnunderstanding of descriptive statistics. An attitude survey instrument was administered tornsee students’ change of attitude towards statistics using Modeling Approach. Teams ofrnstudents did projects on their own themes for six weeks.rnThe quantitative data of the study was analyzed using descriptive statistics, t-tests andrnANOVA. The qualitative data of the study was analyzed using thematic and contentrnanalyses. A statistically significant difference was found on conceptual understandingrnachievement test with medium effect size using Modeling Approach, but no statisticalrnsignificant difference was found on procedural understanding achievement test exceptrnfemale comparison. Significant statistical effects were observed for attitude scale andrnsubscales. Students enhanced their critical understanding of descriptive statistics doingrnprojects through data modeling. The findings of this study suggested students more likelyrnenhanced their understanding of descriptive statistics using Modeling Approach