Over the past fifty years, American public schools have faced much scrutiny and undergone multiple reform efforts in the search for the key ingredients that leads to increased student academic success.  In an attempt to best meet students’ needs and prepare them for future achievement, academic reform efforts and new policies generated a focus on holding all students accountable for high academic achievement, specifically in English and math, in the early 2000s. During the last two decades, researchers and policymakers alike have focused primarily on the development and measurement of students’ cognitive academic abilities (West et al., 2016).  However, recent increase in the research on how non-cognitive skills influence cognitive academic abilities has spurred another reform effort in which educational organizations focus on the development of students’ non-cognitive skills as well as cognitive academic competencies (Blazar & Kraft, 2017; Darragh, 2013; Gershenson, 2016). The Every Student Succeeds Act (ESSA), reauthorizing the Elementary and Secondary Education Act (ESEA) of 1965, has brought a new perspective for the definition and assessment of student learning.  One of the primary goals of the reauthorization of ESSA was to ensure states set high standards so that students successfully graduate from high school prepared for college or careers (ESSA Executive Summary, 2015).  The Act also empowered state and local decision-makers to develop their own accountability systems with a specific focus on both cognitive academic and non-cognitive indicators of success. This non-cognitive indicator of student success in ESSA incites schools to create an environment within the educational organization that takes a whole child approach to learning rather than a focus primarily placed on cognitive academic performance.

The growing body of research supporting the influence of non-cognitive skills has not gone un-noticed by policymakers within educational organizations (Maas, Jochim, & Gross, 2018).  Despite the abundance of research that expresses the importance of non-cognitive skills (Nicoll, 2014; Snipes et al., 2012), the research to date “does not directly translate into an inclusive perspective that captures the complex relationships between school, classroom and student level factors to support students cognitive and non-cognitive educational outcomes (Farrington et al., 2012).  The purpose of the current study was to explore the relationships that exist among teacher and school level factors (e.g., teachers’ instructional practices and school leaders’ instructional and distributed leadership practices), students’ academic mindsets and students’ academic achievement in mathematics.  A special emphasis was given to the analysis of whether any differences in these relationships might be attributed to students’ gender or socioeconomic status. 

             More specifically, the current study attempted answer the following research questions:

1. Do leadership practices, students’ academic mindsets, and academic achievement in mathematics differ by the students’ gender after controlling for students’ socioeconomic status?
2. Holding demographic attributes pertaining to participants and classroom practices constant, do school leaders’ leadership practices and students’ academic mindsets significantly influence students’ academic achievement in mathematics?
3. Holding demographic attributes pertaining to participants and classroom practices constant, do school leaders’ leadership practices significantly influence students’ academic mindsets?
4. To what extent, if any, do students’ academic mindsets mediate the relationships between leadership practices and students’ academic achievement in mathematics?

 Theoretical Framework

            Student learning is a complex phenomenon with multiple behavioral, psychological, and intellectual components.  The current study was grounded within the existing literature and research pertaining to a whole child approach to education (ASCD, 2007) and social cognitive theory (SCT) (Bandura, 1986).  Both of these theories provided a growing body of evidence that shows the increased use of high-stakes testing is restricting students’ conceptual learning and engagement in creative action innovation, both of which are “essential elements of contemporary schooling”

The current study employed a multilevel research design (MLM). Hierarchically structured data, such as the data that was utilized from the PISA dataset, is nested data where participants are clustered together in a systematic manner, such as students grouped within different schools. Data was drawn from the 2012 US PISA data sets. The U.S. PISA sample was stratified into eight explicit groups based on the school’s classification as public or private and region of the country (Northeast, Central, West, and Southeast). Within each category, the sampling was again sorted by five variables including grade range of the school, relativity to population (urban, suburban, etc.), combined minority population, gender, and state. The U.S. PISA 2012 school sample consisted of 240 schools, which was over the 150-minimum school participant requirement to allow for school nonresponse and to reduce design effects. Although 240 schools existed in the study, there were only 162 schools with valid school and student level data. Of those 162 schools, 143 were public schools, 16 were private institutions and three were not classified as either. Within the schools with valid data, there were 4,978 students who completed the background questionnaire and math assessment. Before any data cleaning procedures or analyses were conducted, provided weights were applied to all data. Weights were applied to the student data by employing the weight variable titled W_FSTUWT and to the school data by employing the weight variable titled W_FSCHWT. Descriptive statistics were conducted to determine the existence of missing data. Missing values at level-1 are allowed within HLM, but missing values at level-2 are not allowed (Raudenbush & Bryk, 2002). Therefore, the data set was analyzed for missing data at the school level (level-2) before conducting multilevel analyses. There were nine school groups missing data for level-2 variables and thus were removed from the data set. Once the data files were cleaned of missing data and outliers, composite scores were created for the study variables Instructional Practices, Classroom Environment, Growth Mindset, and Math Achievement. Before conducting reliability analyses, the assumption for unidimensionality was assessed by running factor analysis on each set of subscale items within the composite measures. The factor analysis was run using a principal components method of extraction with a varimax rotation. The rotated component matrix and initial eigenvalues from the analysis were used to assess the assumption of unidimensionality.

Multivariate and multilevel data analysis techniques were employed to examine the direct and mediating relationships between the variables. Multivariate analyses showed that students’ gender was a differentiating factor for most teacher classroom practices, principal leadership practices, students’ academic mind sets, and their math achievements. Multilevel analyses suggested that school leaders and teachers should focus their efforts on developing a school environment that fosters teachers’ orderly classroom management practices, leaders’ exhibition of distributed leadership practices, and students’ self-efficacy beliefs for both genders. Analyses also suggested that efforts to increase female students’ math achievement require more attention as they are currently influenced more negatively compared to their male peers. Results from each of the research questions in the current study have implications for policymakers and practitioners within any educational organization. In terms of leadership practices, the results show that although instructional leadership had a positive influence on student achievement, this relationship was not significant in the full model. Yet, distributed leadership practices did have a significant positive influence on student achievement. Based on the findings from the current study and relevant literature, school leaders should purposefully provide more opportunities for teachers to be involved in the decision making processes within the school by considering each teachers’ strengths and creating a culture of trust and collaboration among teacher teams and committees. Although distributed leadership shows the most significance on math achievement overall, school leaders who are specifically trying to close the gender gap in mathematics and STEM should implement instructional leadership practices such as promoting teaching practices based on research, praising teachers for student participation, and drawing teachers’ attention to the importance of students social and academic development.

Allensworth, E. M., Gwynne, J. A., Pareja, A. S., Sebastian, J., & Stevens, W. D. (2014). Free to fail or on-track to college: Setting the stage for academic challenge: Classroom control and student support. Chicago: University of Chicago Consortium on Chicago School Research. Association for Supervision and Curriculum Development. (2007). The learning compact redefined: A call to action. A report of the commission on the whole child. Alexandria, VA: Carter, G. R. Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological Review, 84, 191–215. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1995). Self-efficacy in changing societies. New York: Cambridge University Press. Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman & Company. Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review Psychology, 52, 1-26. Birch, S. H., & Ladd, G. W. (1998). Children’s interpersonal behaviors and the teacher- child relationship. Developmental Psychology, 34, 934–946. Black, L., Mendick H., & Solomon, Y. (2009). Introduction. In L. Black, H. Mendick, & Y. Solomon (Eds.), Mathematical relationships in education: Identities and participation (pp. 1-4). New York: Routledge. Blazar, D. & Kraft, M. A. (2017). Teacher and teaching effects on students’ attitudes and behaviors. Educational Evaluation and Policy Analysis, 39(1), 146-170. Bowlby, J. (1982). Attachment and loss: Vol. I. Attachment. New York: Basic Books. Brandell, G. & Staberg, E. (2008). Mathematics: a female, male or gender-neutral domain? A study of attitudes among students at secondary level. Gender and Education, 20(5), 495-509. Chetty, R., Friedman, J. N., Hilger, N., Saez, E., Schanzenbach, D., & Yagan, D. (2011). How does your kindergarten classroom affect your earnings? Evidence from project STAR. Quarterly Journal of Economics, 126(4), 1593-1660. Chohlis, D. M. (2014). Engaging girls in mathematical problem solving: A comparative study of a girls-only and co-educational out-of-school time math environment. (Doctoral Dissertation). Retrieved from ProQuest Dissertations Darragh, L. (2013). Constructing confidence in identities of belonging in mathematics at the transition to secondary school. Research in Mathematics Education, 15(3), 215-229. DuFour, R., & Marzano, R. (2009). High-leverage strategies for principal leadership. Educational Leadership, 66(5), 62-68.