the effects of classroom management on student behaviour research paper

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Health Research Alliance Author Manuscripts

Salient Classroom Management Skills: Finding the Most Effective Skills to Increase Student Engagement and Decrease Disruptions

Nicholas a. gage.

Assistant professor in the School of Special Education, School Psychology, and Early Childhood Studies at the University of Florida, in Gainesville

Ashley S. MacSuga-Gage

Clinical assistant professor of special education at the University of Florida and serves as the Special Education Program Area leader for the college’s School of Special Education, School Psychology, and Early Childhood Studies

Classroom Management and Student Achievement

Effective teaching requires a complex skill set. Teachers must deftly deliver academic instruction while maintaining efficiently managed classrooms to ensure student engagement and few disruptions. The bottom line is that students cannot learn if they are not engaged and paying attention to instruction. Therefore, successful classroom instruction is contingent upon effective classroom management to maintain appropriate student behavior, engagement, and, subsequently, academic achievement ( Evertson & Weinstein, 2006 ). In a recent study of elementary teacher effectiveness based on value-added models, classroom management was the only significant predictor of difference between the top-quartile and bottom-quartile teachers ( Stronge et al., 2011 ), supporting the contention that effective teachers are effective classroom managers.

A growing empirical research base supports the direct relationship between classroom management and reduction of disruptive behavior. Oliver, Wehby, and Reschly (2011) completed a meta-analysis on the impact of classroom management on disruptive and aggressive behavior for the Campbell Collaborative ( http://www.campbellcollaboration.org ). Their findings indicate that high-quality classroom management has an average effect of 0.80 ( p < 0.05), almost a full standard deviation reduction of classroom disruptive and aggressive behavior. An earlier meta-analysis by Marzano, Marzano, & Pickering (2003) also found a large average effect size for classroom management on the reduction of disruptive and aggressive behavior ( d = 0.90, p < 0.05) but also found a significant and positive effect size of 0.52 ( p < 0.05) for academic achievement. Taken together, it is clear that classroom management is a critical component of effective instruction ( Scott, 2017 ).

Although the evidence supports the impact of classroom management on student outcomes, research also indicates that many teachers struggle to implement successful classroom management. For example, teachers indicate that they consider classroom management to be the most challenging aspect of their job ( Barrett & Davis, 1993 ; Reinke et al., 2011 ), that they receive very little training in classroom management ( Freeman et al., 2014 ; Oliver & Reschly, 2010 ), and that many exiting the teaching profession within their first five years indicate that classroom management is their primary reason for leaving ( Wei et al., 2010 ). In addition, direct observation research has found, based on more than 3,000 teacher observations, that most teachers do not demonstrate the skills necessary to effectively manage their classrooms ( Scott et al., 2011 ).

Limited training and demonstration of evidence-based skills in classroom management is germane for all students, but particularly for students with, or at-risk for, emotional and/or behavioral disorders (EBDs). Research has established that students exhibiting elevated levels of behavioral problems in the classroom are regularly excluded from classroom instruction, either by being sent to the office ( Sugai et al., 2000 ) or being placed in restrictive settings ( McLeskey et al., 2012 ), and that they continue to fall further behind their peers academically. This issue has been noted as a major concern by the U.S. government. In July of 2015, the U.S. Departments of Education and Justice hosted superintendents, principals, and teachers from across the country to a day-long “Rethink Discipline” conference focusing on the reduction of the well-documented overuse of school suspension and expulsion ( http://www.ed.gov/news/press-releases/educators-gather-white-house-rethink-school-discipline ). Research suggests that the first step to reducing suspensions and increasing access to classroom instruction for students with EBDs is universal implementation of high-quality, evidence-based classroom management (Evans et al., 2013). Further, high-quality, effective classroom management has been noted as a core component for establishing a multitiered Interconnected Systems Framework (ISF; Barrett et al., 2013 ), a model for integrating positive behavior supports and mental health interventions to significantly improve outcomes for students with EBDs.

A handful of systematic reviews of the literature has identified a number of classroom management skills (CMS) that have sufficient evidence to support their effectiveness. These skills include antecedent-based, instruction-based, and consequence-based skills ( Conroy et al., 2013 ; Oliver et al., 2011 ; Scott & Anderson, 2011 ; Simonsen et al., 2008 ). Simonsen and colleagues (2008) identified 20 classroom management skills that have evidence of effectiveness and aggregated them into five domains that: (1) maximize structure and predictability; (2) post, teach, review, monitor, and reinforce expectations; (3) actively engage students in observable ways; (4) use a continuum of strategies to acknowledge appropriate behavior; and (5) use a continuum of strategies to respond to inappropriate behavior. Lewis and colleagues (2004) identified evidence-based classroom management skills that directly affect students with, or at-risk for, EBDs, including (1) teacher praise, (2) high rates of teacher-directed opportunities to respond during instruction, and (3) clear instructional strategies (i.e., direct instruction). Across all of these (see Conroy et al., 2013 for a review), three classroom management skills were consistently noted:

• Individual and group teacher-directed opportunities to respond (TD-OTR);
• Praise and behavior-specific praise (BSP); and
• Prompting for expectations, including pre-corrections.

Although these three skills do not encompass all classroom management skills, they have an ever-growing evidence base. It is also worth noting that these skills are typically incorporated into most evidence-based classroom management interventions and programs, including the Good Behavior Game ( Barrish et al., 1969 ), the Responsive Classroom ( https://www.responsiveclassroom.org ), BEST in CLASS ( Vo et al., 2012 ), and the Incredible Years Teacher Classroom Management ( Reinke et al., 2014 ).

Study Purpose

Although evidence-based classroom management skills have been delineated in the literature, the most salient among them have not been identified. Certainly all three should be in place to effectively manage classroom behavior, but identification of the most effective classroom management skills can inform professional development efforts targeting single classroom management skills to increase the likelihood that they are implemented at a priori determined levels with fidelity ( Gage et al., 2016 ). This study therefore examines direct observation data of teachers’ implementation of classroom management skills across 25 consecutive school days. Specific research questions asked were:

• What classroom management skills significantly predict student engagement during large group instruction?
• What classroom management skills significantly predict student disruptive behavior during large group instruction?

Study Setting and Sample

We recruited 12 elementary school teachers from two elementary schools in the southeastern United States. One school was a university laboratory school serving students in grades K-12 in which approximately 80% of those students performed at or above state benchmarks in reading and math. The second school was a Title I elementary school (K-5), in which 84% of the students received free or reduced lunch, 70% of the students were black, and fewer than 40% of the students were at or above state benchmarks for reading and math. The teachers at the university lab school requested classroom management assistance from the second author of this article, and the assistant principal at the Title I elementary school reached out to the first author for classroom management professional development.

Eight of the 12 teachers taught kindergarten or first grade; two taught third grade; one taught second grade; and one teacher taught fifth grade. All but one teacher was Caucasian and the average years of experience were 5.5 years (range 1:17 years). Eight of the 12 teachers had a master’s degree in education; one teacher was dual certified to teach special and general education. Most (54%) reported receiving classroom management training during their preservice coursework.

We randomly observed three different students during each observation to capture an estimate of overall class-wide performance. Data collectors were instructed to choose three students at random at the beginning of each observation and to exclude students who had been observed during the previous observation. The data collectors would observe the teacher and the first student for the first five minutes of the observation, followed by the second student and the third. No student-level characteristics were collected. Overall, we collected 195 observations of teachers and students.

Study Measures

Teacher behaviors.

We collected frequency data on teachers’ use of the three classroom management skills identified across the classroom management literature reviews. The operational definitions for individual and group TD-OTR, BSP, and prompting for expectations are provided in Table 1 . Operational definitions were congruent with those used in two large-scale direct observation studies of teachers’ class management behavior ( Kern et al., 2015 ; Scott et al., 2011 ).

Operational Definitions of Classroom Management Skills

Student Behaviors

In addition to collecting teacher data, we noted the duration of time that students were academically engaged and the frequency of disruptions during each observation. Academic engagement was defined as follows:

Target student is engaged with instructional content via choral response, raising hand, responding to teacher instruction, writing, reading, or otherwise actively completing an assigned task (e.g., typing on computer, manipulating assigned materials) or the student is passively attending to instruction by orientation to teacher, peer, or materials if appropriate but is not required to do anything other than listening or observing.

Disruptions were defined as follows:

Student displays behavior that does or potentially could interrupt the lesson in such a way that it distracts the teacher and/or other students (e.g., out of seat, makes noises, talks to peer, makes loud comments, and makes derogatory comments). Behaviors can range from low intensity (distracting another student by whispering something to him/her) to high intensity (making threatening statements or destroying property).

Study Procedures

Following institutional review board approval, we invited all kindergarten and first-grade teachers at the university lab school and five teachers requesting classroom management professional development at the Title I school to participate in the research study during a faculty meeting. All teachers invited consented to participate by completing and returning a written consent form after the meeting. The teachers were informed that a trained data collector would observe their instruction daily for up to three months in order to validate the direct observation system and that, based on their data and need, professional development would be provided in the fall.

Direct Observation Procedures

We collected 15-minute direct observations of each teacher during large group instruction, defined as the teacher leading direct instruction for all students in a class at the same time. Each teacher was asked to identify a 20-minute time period when she consistently provided large group instruction in either reading or mathematics. A trained graduate research assistant or hired hourly data collector (undergraduate or graduate student) would stand near the rear of the classroom and quietly observe the teacher without distracting from instruction. Data collectors used Dragon Touch I8 8” Quad Core Windows Tablet PCs loaded with the Lily data collection application, part of the Multi-Option Observation System for Experimental Studies (MOOSES; Tapp et al., 1995). MOOSES is a direct observation system for collecting real-time event recordings of teacher and student behaviors on either a frequency or duration scale. All data collectors received a two-hour group didactic training and conducted periodic observer drift checks to ensure the accuracy of the observations.

Inter-Observer Agreement Procedures

We collected inter-observer agreement (IOA) data for 52% of all observations. The two observers stood near each other but did not talk or interact during the observation except to start the observation at the exact same time. Inter-observer agreement was calculated in MOOSES using the point-by-point method with a three-second window. Across all classroom management skills and student behaviors, the average IOA was 90.1% (range from 82% to 97%).

Data Analysis

In analyzing the data, we first summed the individual and group TD-OTR frequencies together for each observation so that we had a total number of TD-OTRs per observation. Next, we divided all four classroom management skill values by the number of minutes the teacher was observed (i.e., 15 minutes) so that the scale of each variable was rate per minute. We followed the same procedure to calculate the rate of disruptive behavior per minute. To address the two research questions, we estimated a series of three-level random-effects models—also known as hierarchical linear models (HLMs; Raudenbush & Bryk, 2002)—to evaluate the relationship between teachers’ rates of classroom management skills and student behavior. We used three-level models to estimate student behavior nested in time (repeated observations) nested in teacher. First, we estimated a fully unconditional (null) model to calculate the intra-class correlation coefficient (ICC) for time and teacher. The ICC is the percentage of variance in student behavior attributable to time and to teacher. Next, we modeled full models with teachers’ classroom management skills predicting students’ percentage of time academically engaged and their rate of disruptive behavior. All analyses were conducted in the lmer4 package ( Bates et al., 2015 ) in R ( R Core Team, 2013 ) and estimated using restricted maximum likelihood (REML).

Study Results

Descriptive statistics.

Prior to modeling, we examined the descriptive statistics for the full sample and for each teacher across all classroom management skills and student behaviors (see Table 1 ). The average rate of TD-OTR was approximately 2 per minute during large group instruction. Although the full sample average was close to recommended TD-OTR rates (i.e., about 3 per minute during large group direct instruction; MacSuga-Gage & Simonsen, 2015 ), there was considerable variability among teachers, with a range of average rates between 0.76 per minute and 5.12 per minute. The average rate of BSP per minute was 0.44, whereby teachers delivered approximately 6 BSP statements per 15-minute observation. Again, there was significant variability among teachers, evidenced by the standard deviation value greater than the sample average. Lastly, the average rate of prompting for expectations, including pre-corrections, was 0.22, or about 3 per 15-minute observation.

Across all observations and teachers, students were academically engaged 80% of the time. Similar to the teacher classroom management skills, there were large differences between teachers in the average percentage of time students were academically engaged. Two teachers’ students were academically engaged, on average, 68% of the time, while one teacher’s students were academically engaged only 48% of the time. Disruptive behavior was not frequent, with an average of just under two disruptions per observation per teacher. A few teachers had almost no disruptions, although one teacher had an average of almost six disruptions per observation.

Three-Level Random-Effects Models

We estimated four three-level random-effects models, two for each student behavior, to identify the most salient classroom management skills. The ICC results for the academic engagement model suggest that only 2% of the variance was attributable to time, indicating that there was very little variability across time. However, 30% of the variability was within teacher within time, suggesting that there was some variability by time and teacher, supporting the use of the three-level model. The ICC results for rates of disruptive behavior were the same for time, but much smaller for teacher (ICC = 0.13), indicating that the rate of disruptive behavior was consistent within time and within time by teacher.

Next, we estimated fully conditional models to identify the most salient of the three evidence-based classroom management skills. The average percentage of time a student was academically engaged, assuming the three classroom management skills were zero, was 76%. Of the three classroom management skills included in the models, only BSP was statistically significant and positive, suggesting that increased use of BSP had a corresponding positive impact on student engagement. Results for students’ rates of disruptive behavior were similar, with an average rate of 0.13 disruptions, assuming the three classroom management skills were at zero. Again, BSP was the only significant predictor, with a negative coefficient indicating that more BSP was predictive of fewer student disruptions.

Study Findings

Classroom management is a critical component of effective instruction and a prerequisite for classrooms hoping to successfully include students with, or at-risk for, EBDs. Classroom management is also a foundational and critical component of effective multitiered school behavior models, including school-wide positive behavior support and the ISF ( Barrett et al., 2013 ). Without classroom management, implementation of evidence-based behavioral and mental health interventions for students with EBDs is less likely to be successful or to generalize to their general education classrooms. Although myriad classroom management skills and practices have been developed, researched, and reviewed, three skills have been identified as evidence-based and are typically included in most classroom management interventions and programs: (1) teacher-directed opportunities to respond (TD-OTR); (2) behavior-specific praise (BSP); and (3) prompting for behavioral expectations, including pre-corrections. This study has sought to identify which of these three classroom management skills was most salient so as to inform both practice and professional development models about which skill to focus on first. Essentially, our goal was to determine which of the three is the most effective at increasing appropriate classroom behavior during large group instruction. Results from both the academic engagement and rate of disruptive behavior models suggest that BSP was the only classroom management skill that significantly predicted positive student behavior.

Based on the descriptive statistics, the sample of teachers in this study appeared to implement the three classroom management skills at rates greater than those in other studies. For example, Scott, Alter, and Hirn (2011) found that teachers delivered less than one TD-OTR per minute and that their rates of positive feedback were less than 0.1 per minute. In fact, the teachers in this study implemented both TD-OTR and BSP at rates near those recommended in the literature, i.e., 3 TD-OTR per minute during direct instruction ( MacSuga-Gage & Simonsen, 2015 ) and approximately 6 BSP statements per 15-minute observation ( Simonsen et al., 2016 ). However, there was significant variability across the teachers, particularly between the two schools. Teachers in the university lab school had an average TD-OTR rate of 2.4 per minute, and the teachers at the Title I schools had an average TD-OTR rate of 1.2 per minute. Results were similar for BSP, with an average of 7.5 BSP statements per 15-minute observation at the university lab school compared with 4.7 BSP statements per 15-minute observation at the Title I school. Yet, the average rates of classroom management skills in the Title I school were still much larger than those found by Scott and colleagues (2011) .

Results of this study indicate that BSP was the only significant predictor of student performance after controlling for the other classroom management skills. This finding does not indicate that increased TD-OTR and prompting for expectations, including pre-corrections, are not important. Other research has confirmed that each classroom management skill has a positive effect on student classroom behavior (see MacSuga-Gage & Gage, 2015 ). However, the results do suggest that, for the students in this study, BSP appeared to have a positive and statistically significant effect that was greater than that of the other classroom management skills. Therefore, when teachers are considering which classroom management skills they should focus on increasing, BSP is an ideal choice.

Similarly, a recent professional development model using a multitiered system of professional development (MTS-PD) has been developed, which focuses on teaching teachers to implement a single classroom management skill to an a priori level before teaching another classroom management skill. Prior research using the MTS-PD has focused on both TD-OTR ( MacSuga-Gage, 2013 ) and BSP ( Gage et al., 2016 ; Simonsen et al., 2016 ). The findings of this study suggest that starting with BSP may be the best approach to increase teacher buy-in because teachers may see greater increases in engagement and decreases in disruptive behavior as a result of increased BSP.

Study Limitations

Although all efforts were made to ensure the accuracy and reliability of study results, a number of limitations should be mentioned. First, the study does not include all potential classroom management skills identified in the literature. Based on prior reviews, we included TD-OTR, BSP, and prompting for expectations, because they appeared to be the most common and widely researched. However, other relevant skills include error correction, general praise, decreases in negative feedback, high structure, and posting behavioral expectations, as well as behavior intervention systems, including token economies and self-management systems. Therefore, future research should evaluate the relative influence of BSP when other classroom management skills and programs are present.

Second, the authors did not follow individual students across the observations or target students with EBDs. The measured student behaviors represent the classroom average using three five-minute observations of random students per observation. Future research should examine the influence of evidence-based classroom management skills on students with EBDs. We believe implementation of high-quality classroom management is a prerequisite to increase the likelihood that students with EBDs can remain in the general education classroom, but we also know that classroom management alone may not be enough and that additional function-based interventions will be necessary for those students to remain in the classroom. We believe that a continuum of classroom management, function-based interventions, and mental health services leveraging the ISF framework ( Barrett et al., 2013 ) may be the most effective approach to ensure that students with EBDs remain in the general education classrooms.

Last, our statistical models were limited by sample size and inclusion of student and teacher characteristics. Future research should leverage larger samples and include both teacher and student characteristics, including gender, ethnicity, years of experience, and other related variables to increase the precision and accuracy of the model parameters.

Implementation of high-quality, evidence-based classroom management is critical for the success of all students, but particularly for students with EBDs. We sought to identify the most salient single classroom management skill in order to inform practice and professional development models as to which classroom management skill to target first. Our results suggest that BSP may be the most effective classroom management skill to increase engagement and decrease disruptive behavior. That being said, there is no doubt that BSP alone cannot and will not change all students’ behavior in the classroom. Instead, BSP can be used as a first target for improving classroom management and for, ideally, increasing the likelihood that all students, and particularly students with EBDs, will be engaged with instruction.

Descriptive Statistics for the Full Sample and for Each Teacher

Three-Level Random Effects Model of Teacher Classroom Management Skills Predicting Student Behavior

Notes: Significant estimates are in boldface, with p < 0.05*, p < 0.01**, and p < 0.001***; 195 observations, 25 time points, 12 teachers.

Contributor Information

Nicholas A. Gage, Assistant professor in the School of Special Education, School Psychology, and Early Childhood Studies at the University of Florida, in Gainesville.

Ashley S. MacSuga-Gage, Clinical assistant professor of special education at the University of Florida and serves as the Special Education Program Area leader for the college’s School of Special Education, School Psychology, and Early Childhood Studies.

• Barrett ER, Davis S. Perceptions of beginning teachers’ needs in classroom management. Teacher Education & Practice. 1993; 11 :22–27. [ Google Scholar ]
• Barrett S, Eber L, Weist MD. Advancing Education Effectiveness: Interconnecting School Mental Health and School-Wide Positive Behavior Support. Eugene, OR: Center for Positive Behavioral Interventions and Supports (funded by the Office of Special Education Programs, U.S. Department of Education). University of Oregon Press; 2013. [ Google Scholar ]
• Barrish H, Saunders M, Wolf M. Good behavior game: Effects of individual contingencies for group consequences on disruptive behavior in a classroom. Journal of Applied Behavior Analysis. 1969; 2 :119–124. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Bates D, Maechler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. Journal of Statistical Software. 2015; 67 :1–48. [ Google Scholar ]
• Conroy MA, Alter PJ, Sutherland KS. Classroom-based research in the field of EBD: Future research directions. In: Garner P, editor. Handbook of Emotional and Behavioral Difficulties. London: Sage; 2013. pp. 465–477. [ Google Scholar ]
• Evertson CM, Weinstein CS. Handbook of Classroom Management: Research, Practice, and Contemporary Issues. New York: Lawrence Erlbaum; 2006. [ Google Scholar ]
• Freeman J, Simonson B, Briere DE, MacSuga-Gage AS. Pre-service teacher training in classroom management: A review of state accreditation policy and teacher preparation programs. Teacher Education and Special Education. 2014; 37 :106–120. [ Google Scholar ]
• Gage NA, MacSuga-Gage AS, Crews E. Increasing teachers’ use of behavior-specific praise using a multi-tiered system for professional development. 2016 Manuscript submitted for publication. [ Google Scholar ]
• Kern L, Evans SW, Lewis TJ, State TM, Weist MD, Wills H. CARS comprehensive intervention for secondary students with emotional and behavioral problems: Conceptualization and development. Journal of Emotional and Behavioral Disorders. 2015; 23 (4):195–205. [ Google Scholar ]
• MacSuga-Gage AS. PhD diss. University of Connecticut; 2013. Supporting teachers’ professional development: Investigating the impact of a targeted intervention on teachers’ presentation of opportunities to respond. Available at http://digitalcommons.uconn.edu/cgi/viewcontent.cgi?article=6267&context=dissertations . [ Google Scholar ]
• MacSuga-Gage AS, Gage NA. Student-level effects of increased teacher-directed opportunities to respond. Journal of Behavioral Education. 2015; 24 :273–288. [ Google Scholar ]
• MacSuga-Gage AS, Simonsen B. Examining the effects of teacher-directed opportunities to respond on student outcomes: A systematic review of the literature. Education and Treatment of Children. 2015; 38 :211–240. [ Google Scholar ]
• Marzano RJ, Marzano JS, Pickering DJ. Classroom Management That Works: Research-Based Strategies for Every Teacher. Alexandria, VA: ASCD; 2003. [ Google Scholar ]
• McLeskey J, Landers E, Williamson P, Hoppey D. Are we moving toward educating students with disabilities in less restrictive environments. Journal of Special Education. 2012; 46 :131–140. [ Google Scholar ]
• Oliver RM, Reschly DJ. Teacher preparation in classroom management: Implications for students with emotional and behavioral disorders. Behavioral Disorders. 2010; 35 :188–199. [ Google Scholar ]
• Oliver RM, Wehby J, Reschly DJ. Teacher Classroom Management Practices: Effects on Disruptive or Aggressive Student Behavior. Campbell Systematic Reviews. 2011; 2011 :4. [ Google Scholar ]
• R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013. Available at http://www.R-project.org/ [ Google Scholar ]
• Reinke WM, Stormont M, Herman KC, Puri R, Goel N. Supporting children’s mental health in schools: Teacher perceptions of needs, roles, and barriers. School Psychology Quarterly. 2011; 26 :1–13. [ Google Scholar ]
• Reinke WM, Stormont M, Herman KC, Wang Z, Newcomer L, King K. Use of coaching and behavior support planning for students with disruptive behavior within a universal classroom management program. Journal of Emotional and Behavioral Disorders. 2014; 22 :74–82. [ Google Scholar ]
• Scott TM. Teaching Behavior: Managing Classrooms through Effective Instruction. Thousand Oaks, CA: Corwin; 2017. [ Google Scholar ]
• Scott TM, Alter PJ, Hirn R. An examination of typical classroom context and instruction for students with and without behavioral disorders. Education and Treatment of Children. 2011; 34 :619–642. [ Google Scholar ]
• Scott TM, Anderson CM. Managing Classroom Behavior Using Positive Behavior Supports. Upper Saddle River, NJ: Pearson; 2011. [ Google Scholar ]
• Simonsen B, Fairbanks S, Briesch A, Myers D, Sugai G. Evidence-based practices in classroom management: Considerations for research to practice. Education and Treatment of Children. 2008; 31 :351–380. [ Google Scholar ]
• Simonsen B, Freeman J, Dooley K, Maddock E, Kern L, Meyers D. Effects of targeted professional development on teachers’ specific praise rates. Journal of Positive Behavior Interventions Published online March. 2016; 14 :2016. [ Google Scholar ]
• Stronge JH, Ward TJ, Grant LW. What makes good teachers good? A cross-case analysis of the connection between teacher effectiveness and student achievement. Journal of Teacher Education. 2011; 62 :339–355. [ Google Scholar ]
• Sugai G, Sprague JR, Horner RH, Walker HM. Preventing school violence: The use of office discipline referrals to assess and monitor school-wide discipline interventions. Journal of Emotional and Behavioral Disorders. 2000; 8 :94–102. [ Google Scholar ]
• Vo AK, Sutherland KS, Conroy MA. Best in class: A classroom-based model for ameliorating problem behavior in early childhood settings. Psychology in the Schools. 2012; 49 :402–415. [ Google Scholar ]
• Wei RC, Darling-Hammond L, Adamson F. Professional development in the United States: Trends and challenges. Dallas, TX: National Staff Development Council; 2010. [ Google Scholar ]

Class size, student behaviors and educational outcomes

Organization Management Journal

ISSN : 2753-8567

Article publication date: 17 February 2022

Issue publication date: 2 August 2022

While many business schools use large classes for the sake of efficiency, faculty and students tend to perceive large classes as an impediment to learning. Although class size is a contested issue, research on its impact is inconclusive, mainly focusing on academic performance outcomes such as test scores and does not address classroom dynamics. This study aims to expand the focus of class size research to include classroom dynamics and subjective educational outcomes (e.g. student learning outcomes and satisfaction).

Design/methodology/approach

Using Finn et al.’s (2003) theoretical framework and research conducted in introductory business classes, this study investigates how student academic and social engagement influence educational outcomes in different class sizes.

Results highlight the critical role that student involvement and teacher interaction play on student success and student satisfaction regardless of class sizes. In addition, the results indicate that students perceive lower levels of teacher interaction and satisfaction in larger classes.

Originality/value

This study applies Finn’s framework of student engagement in the classroom to understand the dynamics of class size in business education. The results reveal the influential roles of academic and social engagements on educational outcomes. Practical strategies are offered to improve learning outcomes and student satisfaction in large classes.

• Large class teaching
• Student satisfaction
• Assessment of learning
• Business education outcomes
• Student learning outcomes

Wang, L. and Calvano, L. (2022), "Class size, student behaviors and educational outcomes", Organization Management Journal , Vol. 19 No. 4, pp. 126-142. https://doi.org/10.1108/OMJ-01-2021-1139

Emerald Publishing Limited

Copyright © 2022, Liz Wang and Lisa Calvano.

Published in Organization Management Journal . Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence maybe seen at http://creativecommons.org/licences/by/4.0/legalcode

Introduction

According to the 2017 Inside Higher Ed Survey, 71% of 409 chief business officers agreed that higher education institutions were facing significant financial difficulties ( Jaschik, 2017 ). Many business schools use large classes to address the challenges of shrinking resources. Large classes may enable institutions to deploy faculty more efficiently and accommodate more students, especially when it is not feasible to expand facilities or increase hiring ( Guseman, 1985 ). Nevertheless, large class size is a contested issue for students and instructors because it is thought to affect student success ( Blatchford et al., 2009 ; Maringe & Sing, 2014 ).

Most studies on class size focus on student academic performance, but the results are inconclusive. At the elementary and secondary levels, some studies suggest that smaller classes positively impact academic performance ( Glass & Smith, 1978 ; Robinson, 1990). Others indicate that class size has limited or no impact on performance ( Hanushek, 1986 ; Hoxby, 2000 ). Similarly, in higher education, some studies indicate no difference on course grades between large and small classes ( Guseman, 1985 ; Raimondo et al., 1990 ; Karakaya et al., 2001 ) and others report negative effects on academic performance ( Paola et al., 2013 ; Maringe & Sing, 2014 ). Moreover, there is a dearth of research explaining how and why class size influences student behaviors and educational outcomes. Additional research is needed to better understand classroom dynamics related to class size ( Anderson, 2000 ; Finn et al., 2003 ); Blatchford et al., 2009 ).

Another issue in the class size literature is that most studies focus on grades or standardized test scores as the primary measure of student success. Thus, research that examines the relationship between class size and educational outcomes beyond academic performance should be included in learning assessment. For example, the Association to Advance Collegiate Schools of Business ( AACSB, 2021 ) endorses the use of “well-documented assurance of learning (AoL) processes that include direct and indirect measures for ensuring the quality of all degree programs that are deemed in scope for accreditation purposes.” In addition, the shift from teacher-directed to student-centered pedagogy means that student perception of learning has become an important educational outcome ( Maher, 2004 ; Adam, 2004 ). Today, student satisfaction is recognized as critical factor in attracting and retaining students ( Santini et al., 2017 ).

This study aims to fill the aforementioned gaps in the literature by applying Finn et al.’s (2003) theoretical framework of student engagement in the classroom. They suggest that student academic and social engagement with peers and teachers may influence academic achievement. This study uses Finn’s framework to investigate how student learning and social behaviors influence relevant educational outcomes in different class sizes. The purpose of the study is twofold: to better understand the dynamics of class size in business education and to provide practical strategies to improve educational outcomes and student satisfaction in large and small classes.

Theoretical background

Although business schools typically consider class size a factor in determining teaching loads, there is no accepted definition of a large class. Mateo and Fernandez (1996) propose a numerical taxonomy. For example, a large class contains between 60 and 149 students. Maringe and Sing (2014) define large class size qualitatively as “any class where the number of students poses both perceived and real challenges in the delivery of quality and equal learning opportunities to all students in the classroom” (p. 763). In practice, class size norms vary greatly across institutions and disciplines, with some business schools considering sections of 25–35 students to be small and between 200 and 350 to be large ( Raimondo et al., 1990 ).

Conceptual model

Finn et al. (2003) suggest student academic achievement is influenced by a combination of academic and social engagement in the learning process. Academic engagement refers to student learning behaviors related directly to the learning process, such as class participation. Social engagement is student social interactions with classmates and the instructor. Using group theory, Finn et al. (2003) argue that students in small classes are more visible and more likely to engage in learning and social behaviors during class. Conversely, large classes permit students to reduce their visibility. Also, smaller classes encourage participation or interaction as students may receive more support from classmates. Because social and academic interactions are the focal point of the higher education, these classroom dynamics are critical to positive learning outcomes ( Demaris and Kritsonis, 2008 ).

Blatchford et al. (2009) suggest that a negative relationship exists between class size and classroom processes. Class size differences may impact classroom processes, which in turn influence student attentiveness and active involvement with teachers and peers. Teachers in small classes are more likely to give individual attention to students, effectively control and manage the classroom and build better relationships with students. Similarly, students in small classes may be more engaged in classroom and more likely to interact with teachers and peers ( Blatchford et al., 2009 ).

Consistent with Finn et al. (2003) and Blatchford et al. (2009) , this study proposes a research model in Figure 1 suggesting how class size affects student learning and social behaviors, as well as learning outcomes. It includes perceived learning outcomes and satisfaction as additional educational outcomes.

Research hypotheses

Student attentiveness level will be higher in small classes than in large classes.

Student involvement level with the course will be higher in small classes than in large classes.

Student class participation level will be higher in small classes than in large classes.

Student interaction with classmates will be higher in small classes than in large classes.

Student perception of teacher encouragement will be higher in small classes than in large classes.

Student perception of teacher supportiveness will be higher in small classes than in large classes.

The factors affecting student academic performance may include attentiveness, involvement, class participation, interaction with classmates, perception of teacher encouragement, perception of teacher supportiveness and class size.

The factors affecting the student perceived learning outcome of business knowledge may include attentiveness, involvement, class participation, interaction with classmates, perception of teacher encouragement, perception of teacher supportiveness and class size.

The factors affecting the student perceived learning outcome of communication skills may include attentiveness, involvement, class participation, interaction with classmates, perception of teacher encouragement, perception of teacher supportiveness and class size

The factors affecting student satisfaction may include attentiveness, involvement, class participation, interaction with classmates, perception of teacher encouragement, perception of teacher supportiveness and class size.

This study was conducted in the business school of a medium-sized public university in the northeastern USA. Students were recruited from three introductory business courses (management, marketing and business law) that were offered in both small and large sections. The typical size of most classes in this business school is 25–40 students and a class with more than 50 students is considered large. In general, students are able to select their own classes and all majors were represented in all classes.

Survey participants were invited by their instructors to take an online survey voluntarily at the end of the semester. Participating students granted permission to the researchers to access to their academic records, including course grade. A total of 280 student participated with 52 respondents from small classes and 228 from large classes. The overall response rate was 43% (37% for small classes vs 45% for large classes). Approximately 47.9% of respondents were male and 51.4% were female. Nearly all (98.6%) respondents attended school full-time. The average number of credits completed was 65.

Student academic performance was measured using a student’s course grade (A = 4.0, A− = 3.670, B+ = 3.330, B = 3.000, B− = 2.670, C+ = 2.330, C = 2.000, C− = 1.670, D+= 1.330, D = 1.000, D− = 0.670, F = 0.00). The mean and the standard deviation for course grades were 3.223 and 0.497. Most measures of this study were adapted from previous research. Student attentiveness was assessed by items from Leufer’s (2007) research on the factors affecting the learning environment. Student involvement with the course was measured by using a personal involvement inventory and the questions were adapted to fit the context of the survey ( Zaichkowsky, 1994 ). Bai and Chang’s (2016) measures were adapted to assess class participation, student interaction with classmates, student perception of teacher encouragement and student perception of teacher supportiveness. Measures of the SLOs of business knowledge and communication skills were developed by the researchers. These two SLOs were selected because they are assessed as part of the institution’s AACSB AOL process. Student satisfaction was adapted from Eastman et al. (2017) . All questions used seven-point scales.

SPSS was used to run Cronbach’s coefficient alpha of each measure. The results ranged from 0.929 to 0.756, higher than the minimum requirement of 0.7 ( Nunnally, 1978 ). The scores of composite reliabilities for all constructs were higher than 0.6 and demonstrated the reliability of the measures. Harman’s single factor test was used to evaluate common method variance (CMV). The CFA results indicate 41% total variance explained and then no CMV issues ( Podsakoff et al., 2003 ). Table 1 displays the means, standard deviations, and correlation matrix of all constructs.

A one-way ANOVA was run to test H1 with class size as the independent variable and student attentiveness as the dependent variable. On Table 2 , there was no significant difference in student attentiveness between small and large class sizes (F = 1.00, p < 0.32).

Following the same logic, a series of ANOVA were run to test H2 – H6 . On Table 2 , the results indicate no significant differences between class sizes on student involvement ( H2 ), class participation ( H3 ), interaction with classmates ( H4 ) and student perceived teacher encouragement ( H5 ). But, the results support H6 (F = 6.25, p < 0.013), suggesting that class size differences impact perceived teacher supportiveness. Students in small classes perceived higher level of teacher supportiveness than those in large class (Mean = 6.14 > Mean = 5.68).

Multiple regression analysis was conducted to test H7 – H10 . A dummy variable was created to represent two categories of class size: small class (coded as 0); and large class (coded as 1). As discussed below, each educational outcome was a dependent variable and all factors involving student learning and social behaviors and the class-size dummy variable were independent variables. Table 3 displays the statistical results for H7 – H10 .

The regression model for H7 was not statistically significant (F (7, 272) = 1.424, p < 0.195) and the results did not support H7 . The regression analysis for H8 was statistically significant (F (7, 272) = 101.912, P < 0.000). The results indicate that student involvement ( β 2 = 0.321, p < 0.000), student interaction with classmates ( β 4 = 0.085, p < 0.048), student perception of teacher encouragement ( β 5 = 0.276, p < 0.000) and teacher supportiveness ( β 6 = 0.349, p < 0.000) impact student perceived learning outcome for business knowledge.

The regression model for H9 was statistically significant (F (7, 272) = 56.75, P < 0.000). The results reveal that student involvement ( β 2 = 0.176, p < 0.000), student class participation ( β 3 = 0.254, p < 0.000), student interaction with classmates ( β 4 = 0.262, p < 0.000), student perception of teacher encouragement ( β 5 = 0.173, p < 0.013) and teacher supportiveness ( β 6 = 0.131, p < 0.042) also influenced perceived learning outcome of communication skills.

The regression model for H10 was statistically significant (F (7, 272) = 67.174, P < 0.000). Student attentiveness ( β 1 = 0.11, p < 0.004), student involvement ( β 2 = 0.144, p < 0.000), student class participation ( β 3 = 0.099, p < 0.042), student perception of teacher encouragement ( β 5 = 0.239, p < 0.000), teacher supportiveness ( β 6 = 0.347, p < 0.000) and class size ( β 7 = −0.127, p < 0.001) were determinants of student satisfaction. Importantly, student satisfaction level was significantly lower in large classes than those in small classes by 12.7%.

Structural equation modeling (SEM) was conducted to understand how student learning behaviors, peer interaction, and teacher interactions influence the three significant educational outcomes. The partial least square SEM analysis (PLS-SEM) works efficiently with small sample sizes, multi-item measures and complex structural models. It makes no distributional assumptions for data ( Hair et al., 2017 ). We use SmartPLS 3 software to run PLS-SEM analysis, as it is an appropriate approach for our data set to assess the key drivers for educational outcomes.

Given the concern of discriminant validity for teacher encouragement and teacher supportiveness, these two factors are combined into one construct to represent teacher interaction. Figure 2 displays the structural model. The sample size for small classes in this study exceeds the minimum sample size for PLS SEM analysis based on the 10 times rules ( Hair et al., 2017 ).

The measurement model on convergent validity, reliability and the discriminant validity were assessed. On Table 4 , the values of Cronbach’s alpha and composite reliability of all the constructs exceeding the standard level of 0.70. The average variance extracted (AVE) for all the constructs exceeds the lower acceptable limit of 0.50 ( Hair et al., 2017 ) except for attentiveness. Discriminant validity is assessed with Fornell and Larcker criterion and the heterotrait–monotrait ratio of correlations (HTMT) ( Hair et al., 2017 ). Table 5 results support discriminant validity as the variance shared between constructs is lower than the variance shared by a construct with its indicators ( Fornell & Larcker, 1981 ). The bootstrapping results also support discriminant validity as all the HTMT values are significantly from 1 ( Hair et al., 2017 ). Except for attentiveness with a slightly lower AVE of 0.46 (< 0.5), the other constructs demonstrate good reliability, convergent and discriminant validity for the measurement model ( Hair et al., 2017 ).

On the model results, the value of SRMR is 0.063 (< 0.08) which is considered a good fit. No VIF issue was found as each predictor construct’s VIF value was between 0.2 and 5 ( Hair et al., 2017 ). Then, we ran bootstrapping to assess the significance of path efficient. Table 6 displays each path coefficient, t -values and p -value in the structural model. Most of the paths were statistically significant. However, attentiveness has no significant path coefficients with knowledge and communication outcomes. Student class participation and peer interactions have no significant coefficients with knowledge outcome and satisfaction. The coefficients of determination ( R 2 value) for communication outcome, knowledge outcome and satisfaction are 0.60, 0.74 and 0.68 respectively, which indicate moderate predictive power ( Hair et al., 2017 ). Among all the paths, teacher interaction has large effects on knowledge outcome (coefficient = 0.601) and satisfaction (coefficient = 0.638). Involvement has a medium effect on knowledge outcome (coefficient = 0.315).

Multiple group analysis was tested for differences on path coefficients between small and large classes in the same structural model. The results only indicate a significant difference on the path coefficient from student peer interaction to communication outcome between the two groups. While the path coefficient for large class is 0.22, the value for small class is 0.54. It suggests that peer interaction in a small class exerts a stronger positive effect on student perceived communication outcome than that in a large class.

Discussion and implications

Large classes are unlikely to disappear given the financial pressures that most institutions face in the USA ( Maringe & Sing, 2014 ). However, large classes are associated with challenges in delivering high-quality and equitable learning opportunities ( Bligh, 2002 ). For example, students in large classes do not have the same opportunities to interact with the teacher compared to students in small classes ( Maringe & Sing, 2014 ). To offer large classes without sacrificing quality of education, educators must understand how and why class sizes influence student engagement behaviors and educational outcomes.

For all class sizes, this study found student involvement as the most influential academic engagement behavior and teacher interaction as the most influential social engagement behavior for positive educational outcomes. In addition, students only perceived teacher supportiveness more positively in small classes with no differences for other engagement behaviors. In terms of educational outcomes, this research found negative effects on student satisfaction in large classes, but no differences on course grade, knowledge and communication learning outcomes in large and small classes. Despite of the mixed effects of class sizes, the study reveals that in the large classes, students may perceive a lower level of teacher interactions and satisfaction. It raises critical concerns as teacher interaction was found as the most influential driver for all subjective educational outcomes. Given that student satisfaction is a key component of student and institutional success ( Santini et al., 2017 ), educators must develop strategies to enhance teacher interactions and satisfaction in large classes to maintain and enhance the quality of education and student success.

Practical implications

Because teacher interaction is the most influential factor for student satisfaction, business schools may consider allocating technology resources to support a more interactive learning environment in large classes. For example, the use of a student response systems (SRS), such as clickers or Poll Everywhere, has become more popular. Heaslip et al. (2014) found that students in large classes became more engaged and involved when clickers were in use. In addition, SRS enable students to have more equal opportunities to interact with the teacher easily and efficiently.

Choosing educational outcomes that accurately measures learning objectives is critical to monitor and improve education quality. Educational outcomes should reflect what the program wants the students to know and be able to do. For example, this study included a communication student learning outcome because it was assessed as part of AACSB AOL process. The results found that the communication learning outcome is associated with student participation and peer interaction. Thus, when a course focuses on communication goals, faculty should create more opportunities for student participation and peer interaction in course design. For example, in large classes, SRS allow students to participate in class discussion and also to see other students’ responses. Additionally, as peer interaction in a small class is stronger on communication outcome than that in a large class, a small class is a better choice for a course focusing on communication skills.

Different educational outcomes may involve different student learning and social behaviors in the classroom, whereas class size may not influence all these behaviors. What educational outcomes do schools expect for students? If, for example, student satisfaction is the key educational outcome, then our study suggests that large class sizes should not be used. On the other hand, if course grade is the key educational outcome, both large and small classes will work as grade differences are not related to class size.

Research implications and future research

While the extant literature focuses on academic performance as the primary learning outcome, this study shifted to a “student-centered” perspective and added three measures of subjective educational outcomes – student satisfaction and the perceived SLOs for knowledge and communication skills – to transcend the usual academic performance outcomes. As schools use AOL results for continuous improvement, educators should consider a broader set of educational outcomes.

There have been tremendous changes in the modality of course delivery in higher education since the covid pandemic. Drea (2021) notes that higher education is unlikely to fully return to pre-COVID-19 course delivery models, as students have now experienced the intensive integration of technology into their courses, and this has likely reset their expectations for the future. There are a variety of course delivery formats emerging since the pandemic. Educators must understand whether a change of modality of content delivery has an impact on quality. The current study recognized the importance of both academic and social engagements in student learning. It may offer a groundwork to investigate how student engagements influent their learning outcomes in different delivery formats, such as online, hybrid, synchronous online or asynchronous online.

Limitations

Because this study was conducted in a face-to-face classroom setting, the results cannot be generalized to different learning environments, such as online or hybrid. Similarly, the results are not generalizable across all types of higher education institutions because it was conducted at one university. In addition, the results are limited to introductory business courses and do not include advanced courses that require higher-order thinking and analytical skills. Finally, the sample included only undergraduate students and does not consider age as a salient factor when considering effects on classroom processes ( Blatchford et al., 2009 ).

Conclusions

Many schools use large classes to respond to shrinking resources. This study contributes to the existing literature by showing how and why class sizes influence student engagement behaviors and educational outcomes other than academic performance. Student involvement and teacher interaction are found as influential factors on student learning outcomes and satisfaction regardless of class sizes. However, the study results indicate students perceive lower levels of teacher interaction and satisfaction in larger classes. In conclusion, to offer large classes without sacrificing quality of education, we suggest faculty creating more opportunities to encourage more student–teacher interactions, such as using SRS technologies.

Research model

PLS-SEM structural model

Descriptive statistics and correlation matrix