109830072211265681.pdf

https://doi.org/10.1177/10983007221126568

Journal of Positive Behavior Interventions 1 –12© Hammill Institute on Disabilities 2022Article reuse guidelines: sagepub.com/journals-permissionsDOI: 10.1177/10983007221126568jpbi.sagepub.com

Empirical Research

Teacher praise has been examined for decades by research-ers as a form of positive reinforcement to improve student behaviors when attention maintains behavior (Becker et al., 1967; Brophy, 1981; Markelz, Taylor, et al., 2019; Sutherland et al., 2000; White, 1975). It is often stated that teacher praise is a low-intensity, effective, and efficient pro-active classroom management strategy for academic and social behaviors (Ennis et al., 2018; Lane et al., 2015). Although there is debate about whether sufficient high-quality research exists to categorically claim teacher praise as an “evidence-based strategy” (e.g., Moore et al., 2019; Royer et al., 2019), researchers have demonstrated the effectiveness of teacher praise in a variety of settings with a variety of populations (Ennis et al., 2020; Moore et al., 2019).

There is consensus among scholars that behavior-spe-cific praise (BSP) may be a more salient reinforcer than general praise (GP) due to the explicit linkage between teacher approval and a specific student behavior (Alberto et al., 2022; Gage & MacSuga-Gage, 2017). An example of GP is “Nice job!” An example of BSP is “I like the way you put your things away then immediately started working on your project, well done!” When students receive feedback about their behavior delivered as a specific positive affirma-tion, the student’s behavior is rewarded with attention, the student is told which specific behavior resulted in

that attention, and their classmates have been reminded of classroom expectations. Specific feedback is a potentially more effective reinforcer than nonspecific feedback because the recipient is oriented to the behavior that elicited rein-forcement and is thus able to replicate that behavior in the future (Cooper et al., 2020). Accordingly, BSP is recom-mended as a Tier 1 (classroom level), Tier 2 (small group level), or Tier 3 (individual student level) intervention to promote desirable behaviors (Floress et al., 2020).

Praise Variety

The past empirical literature has focused on the efficacy of BSP versus GP (Markelz, Taylor, et al., 2019). Yet, scholars have suggested another characteristic of praise may affect efficacy such as praise variety (Floress & Beschta, 2018; Hager, 2012; Markelz et al., 2020). Praise variety is a topo-graphical characteristic grounded in research related to the

1126568 PBIXXX10.1177/10983007221126568Journal of Positive Behavior InterventionsMarkelz et al.research-article2022

1Ball State University, Muncie, IN, USA2James Madison University, Harrisonburg, VA, USA3University of Virginia, Charlottesville, USA

Corresponding Author:Andrew M. Markelz, Ball State University, 749 Teachers College, 2000 W. University Ave., Muncie, IN 47306, USA. Email: ammarkelz@bsu.edu

The Effects of Varied and Non-Varied Praise on Student On-Task Behaviors

Andrew M. Markelz, PhD1, Benjamin S. Riden, PhD, BCBA-D2 , Stephanie Morano, PhD3, Alicia L. Hazelwood, MS1, and April M. Taylor, MA1

AbstractResearch has demonstrated behavior specificity as a salient characteristic of teacher praise efficacy. Praise variety may also be an important characteristic to reinforce desired student behavior based on research about the quality of reinforcers. In this study, we used an alternating treatments design to examine the effects of varied and non-varied behavior-specific praise (BSP) on two first-grade students’ on-task behaviors in general education classrooms. Visual and statistical analyses suggest both varied and non-varied BSP increased on-task behavior, with varied BSP resulting in marginally higher levels of on-task behavior. There was no functional relationship between varied and non-varied BSP conditions. Findings from this study contribute to teacher praise literature as the first to empirically investigate the effects of praise variety on student behavior. We discuss the implications of this preliminary research and encourage future inquiry into additional characteristics of praise.

Keywordsbehavior-specific praise, on-task behavior, praise variety, teacher training

2 Journal of Positive Behavior Interventions 00(0)

quality of reinforcer (Markelz et al., 2020). The stronger or more effective a reinforcer is, the greater likelihood of meaningful behavior change. Floress and Beschta (2018) suggested that teachers who praise in various ways (e.g., physical gestures, verbal praise, and written notes) may become a discriminative stimulus for the availability of praise. As students learn which teachers deliver more praise (i.e., provide more reinforcement), they are likely to per-form behaviors that in the past have led to teacher praise (Greer, 2002). However, stimulus habituation is a decrease in response to a stimulus after repeated delivery (Thompson & Spencer, 1966). For example, at a children’s birthday party, the first balloon that pops likely startles everyone, but over time, as more balloons pop, the sound becomes less startling because party-goers habituate to the sound. Similar to a balloon popping, a student who first hears a praise statement might take notice more readily, as the statement is novel to the environment. However, teacher praise, like any other stimulus (e.g., balloon popping), may become irrele-vant if students habituate to hearing praise.

Shriver and Allen (2008) contend that individuals learn to attend to relevant stimuli and ignore stimuli that have little meaning or importance. Student habituation to praise may diminish the reinforcer strength and the praise efficacy (Markelz et al., 2020). Given the potential for praise variety to impact the efficacy of this widely purported classroom management strategy, an examination of praise variety lit-erature is warranted.

Praise Variety Literature Review

Researchers have suggested that sincere praise will more likely capture students’ attention (Bayat, 2011; Brophy, 1981; Henderlong & Lepper, 2002; McKay, 1992). However, sincerity is a subjective and difficult to quantify term that has been neglected in the literature. Markelz, Riden, Floress, et al. (2022) suggest that praise variety may increase perceived sincerity as the praise statement novelty decreases the likelihood of habituation. Rather than habitu-ating to hearing “Good job,” a student may believe a teach-er’s praise statement is more genuine with statements like, “Thank you for putting your supplies away immediately after I asked the class.”

Few studies have empirically examined praise variety. Floress and Beschta (2018) measured “diverse praise” in 28 kindergarten through fifth-grade classrooms and defined it as “the use of verbal statements or gestures of approval that are delivered in a variety of distinguishable ways in response to desired student behavior” (p. 1191). The authors mea-sured the rate of teacher praise delivery method toward vari-ous student behaviors. Verbatim GP data were coded into eight categories: (a) work/job, (b) adjective, (c) effort, (d) compliance/appreciation, (e) gesture, (f) tangible, (g) physi-cal, and (h) miscellaneous. Verbatim BSP data were coded

into categories based on student behavior variety (e.g., sit-ting, working, hand-raising, following directions). For example, if the teacher said, “Nice job sitting in your seat” and “Good job sitting down when I asked,” within the same observation, one diverse BSP category was counted because both BSP statements targeted “sitting down.” On average, teachers used 3.7 total diverse praise categories per observa-tion (M = 18.2 min per observation) and more general diverse praise categories (2.2) compared with specific diverse praise categories (1.5; Floress & Beschta, 2018).

In a study examining a student-teacher’s use of effective teaching strategies in a moderate to severe self contained elementary classroom, Hager (2012) measured praise vari-ety by quantifying the adjective used in the praise state-ment. For example, if the student-teacher said, “Good job working quietly” and “Nice work working quietly,” praise variety was counted twice (i.e., once for “good” and once for “nice”). During baseline, the teacher delivered an aver-age of 7.6 varied praise statements. Following a video self-monitoring intervention, the teacher averaged 14.8 varied statements.

Markelz and colleagues (2020) developed and assessed the reliability of an observation tool to measure praise spec-ificity, contingency, and variety. Similar to Hager (2012) who identified the statement adjective as the key compo-nent to varied praise, Markelz et al. explained BSP variety as “the teacher praises a student (or group of students) for a specific behavior using a variety of descriptive language (e.g., ‘Sam, you did a super job raising your hand.’ At the next opportunity, the teacher praises Sam for raising his hand ‘Sam, wonderful job raising your hand’).” Unlike Floress and Beschta (2018) who included student behavior as a required component of diverse praise, if a teacher said, “Great job raising your hand” and “Great job sitting in your seat” those statements would be counted as non-varied since “great job” was used in both statements.

Markelz et al. (2020) determined that statement adjective was appropriate to measure praise variety since GP does not identify a specific behavior. Furthermore, teachers often need to single out a specific student behavior for frequent reinforcement. For example, a student’s behavior interven-tion plan (BIP) may target sharing with classmates. The teacher would want to identify instances of the student shar-ing with classmates and deliver a dense schedule of descrip-tive praise (“I love how you are sharing right now,” “You are doing an excellent job sharing with your friends,” “Awesome job, I love seeing you share with your classmates”).

Using the behavior-specific praise observation tool (BSP-OT), Markelz, Riden, Floress, and colleagues (2022) measured natural rates of teacher praise specificity, contin-gency, and variety. To calculate praise variety the authors (a) counted different adjectives used over a 15-min observation session and then (b) divided the number of different types of praise statements by total statements and multiplied by 100.

Markelz et al. 3

For example, if a teacher delivered 10 BSP statements and used “great,” “wonderful,” “good,” and “love,” then the teacher used four varied praise statements (4/10 × 100 = 40% praise variety).

Results from Markelz, Riden, Floress, et al. (2022) sug-gested that special education teachers (n = 12) delivered 44.9% BSP praise variety and 48.6% GP praise variety per 15-min observation. General education teachers (n = 13) delivered 44.5% BSP variety and 51.5% GP variety with no statistical differences between special education and gen-eral education teachers (BSP variety [U = 76.0, p = .934]; GP variety [U = 69.0, p = .643]). Preservice special educa-tion teachers (n = 17) delivered 51.5% BSP variety and 57.7% GP variety and were excluded from statistical differ-ence calculations due to limited observations. Across teacher groups, low natural rates of BSP and GP were observed per observation session (BSP M = 1.59; GP M = 7.2), which corresponded with prior research measuring natural rates of praise delivery (Floress et al., 2018; Jenkins et al., 2015).

Purpose of the Study

Given the identification of praise variety as a potentially salient characteristic of praise efficacy due to potential stimulus habituation, this study empirically examined the effects of varied and non-varied specific praise on student on-task behavior. Previous research has observed teacher’s natural rates of praise variety (e.g., Markelz, Riden, Floress, et al., 2022); however, no study has empirically examined whether praise variety affects student behavior. The follow-ing study adds to the literature on teacher praise by empiri-cally manipulating praise variety to expand researchers’ understanding of praise efficacy. In other words, does var-ied praise serve as a more salient reinforcer than non-varied praise on student on-task behaviors? The following research questions guided our analysis:

Research Questions

Research Question 1: Is there a functional relation between varied BSP and increased student on-task behavior?Research Question 2: Is there a functional relation between non-varied BSP and increased student on-task behavior?Research Question 3: Is there a difference in effective-ness between varied BSP and non-varied BSP on student on-task behavior?

Method

Participant and Setting InformationThe study took place in a public k-5 Title 1 elementary school in the United States with 490 students. Approximately

40% of the student population identified as a racial or ethnic minority. Prior to study initiation, institutional review board approval was obtained from university and school district review boards. A recruitment email was sent to three differ-ent principals within the school district to forward to their general and special education teachers. The email asked teacher volunteers to participate in a study that could help with classroom management practices. The inclusion crite-rion included no planned extended absences. Teacher par-ticipants were offered a US$400 incentive to participate. Two teachers from the same public elementary school vol-unteered to participate in the study. Each teacher selected an instructional period for observation to occur and identified a target student from their classroom, who they believed demonstrated frequent off-task behaviors, to complete a teacher–student dyad. Voluntary consents from teachers and students’ parents were obtained. All names have been changed to protect participant identity.

Two certified general education teachers and one of their students in an inclusive classroom participated in the study. Both participating teachers were white females with Bachelor’s in Elementary Education. Dyad A comprised a teacher and the target student in a classroom with 20 stu-dents total. Ms. Alisha was 41 years of age and had a total of 8 years of teaching experience, two of those years in her current position. Alli was a 6-year-old Black girl without a disability. Ms. Alisha reported that Alli frequently demon-strated off-task behaviors such as climbing on furniture, repeatedly seeking teacher attention, and not completing assigned activities. Alli was on a behavior intervention plan for 3 weeks consisting of a check-in/checkout (Klingbeil et al., 2019) when the study started. Ms. Alisha completed a Functional Analysis Screening Tool (FAST; Iwata et al., 2013), which indicated Alli’s off-task behavior was socially maintained (attention/preferred item). Dyad A observations occurred daily during the beginning of reading/writing class where students sat at their desks, clustered in small groups of 4 to 5 students, and participated in teacher-led activities, such as filling in blanks to complete story sentences, spell-ing practice, and sight word practice.

Dyad B comprised a teacher and the target student in a classroom with 22 students total. Ms. Brenda was 21 years of age and had 4 years of overall experience and 2 years in her current position. Brice was a 7-year-old Black boy who was medically diagnosed with attention deficit hyperactive disorder (unmedicated) and was not receiving special edu-cation services. Ms. Brenda identified Brice as a good can-didate for the intervention due to excessive off-task behaviors, such as not completing assignments and aggres-sion toward other students. Brice was also receiving a behavior intervention plan for 4 weeks with check-in/checkout. Ms. Brenda completed a FAST, which indicated Brice’s off-task behaviors were equally social and automat-ically maintained (attention/preferred item and sensory stimulation). Dyad B observations occurred daily during

4 Journal of Positive Behavior Interventions 00(0)

reading instruction where students sat at their individual desks (Brice’s desk was isolated in the front of the class-room) and participated in teacher-led activities, such as choral reading, or sat on the carpet in front of a smartboard and participated in the video-led reading activities.

Study Design

An alternating treatments design (Cooper et al., 2020) was used to assess if a functional relation between varied/non-varied praise and student on-task behaviors exists. This design was used to measure whether rapid changes in praise variety would affect student on-task behaviors. The present study aligns with the What Works Clearinghouse version 4.1 single-case design review standards (What Works Clearinghouse, 2020): (a) the independent variable was sys-tematically manipulated; (b) outcome variables were mea-sured systematically over time by more than one assessor, and interobserver agreement (IOA) was collected for at least 20% of all sessions across all phases; (c) the study included at least three attempts to demonstrate an interven-tion effect at three different points in time; (d) each phase had a minimum of five data points, and (e) no more than two consecutive treatments were scheduled.

Dependent Variable

The primary dependent variable was student on-task behav-ior. Student behaviors were measured as a percentage using whole interval recording (Ledford & Gast, 2018) with 20-s intervals. Whole interval recording was selected because it underestimates the duration of behavior (Ledford & Gast, 2018). On-task behavior was defined as participating in a lesson or activity or being focused and attending to a speaker or activity. Examples of student on-task behavior included looking at the teacher when they were talking, working on iPad activity (e.g., phonics practice), and rais-ing one’s hand to participate in classroom discussion. Students were also considered on-task if they were per-forming typical classroom activities (e.g., taking scrap paper to the trash can or handing out worksheets). Examples of off-task behavior included walking around the classroom without permission, talking with other students during instruction, and head down on desk and not completing assignments.

Secondary dependent variables were teacher praise char-acteristics delivered toward the target student at any time during observation sessions (i.e., specific/general praise, contingent/non-contingent, and varied/non-varied praise). Definitions of praise characteristics were used from the BSP-OT (Markelz et al., 2020). Specific praise was defined as a positive statement delivered by the teacher toward the target student that identified a behavior and incorporated an expression of approval (e.g., great job raising your hand,

Sarah). General praise was defined as a positive statement delivered by the teacher toward the target student that was unspecific to behavior (e.g., “Good work, Sarah”).

Specific and general praise can classify as contingent or noncontingent. Contingency is the relationship between two events, one being a consequence of temporal proximity to the other event (Markelz et al., 2020). Contingent-specific praise was defined as a positive statement provided by the teacher, when a desired behavior occurred (contin-gent), to inform the target student, specifically, what they did well (e.g., a teacher tells the class to quiet down; Sarah gets quiet, teacher responds, “Great job getting quiet, Sarah”). Contingent general praise was a positive state-ment provided by the teacher, when a desired behavior occurred to inform the student generally that they did well (e.g., Teacher tells the class to sit down; Sarah sits, teacher responds, “Good job, Sarah”).

Specific and general praise can also classify as varied or non-varied. Praise variety is about the quality of praise being different or diverse (Markelz et al., 2020). Varied spe-cific praise was defined as the teacher praising the student for a specific behavior using a variety of descriptive lan-guage (e.g., “Sarah, you did a super job raising your hand.” At the next opportunity, the teacher praises Sarah for raising her hand, “Sarah, wonderful job raising your hand”). Varied general praise was when the teacher praises a student with a variety of descriptive language without specifying a behavior (e.g., “Great work, Sarah!” “Nice job, Sarah”).

Independent Variables

Given the study’s purpose to examine differences between varied and non-varied praise on student on-task behaviors, teacher participants needed to deliver consistent specific and contingent characteristics of praise, prior to indepen-dently manipulating varied praise. We trained teachers to ensure high levels of specific and contingent praise were being delivered (i.e., a minimum of 10 statements toward the target student) before entering the treatment phases. Once participants entered the treatment phase, Treatment 1 (T1) was a high rate of varied praise. Treatment 2 (T2) was a high rate non-varied praise.

Data Collection

Dependent and independent variables were measured with an adapted BSP-OT. Previous research suggests the BSP-OT is a reliable tool to measure teacher praise characteristics with an interclass correlation of .80 and a κ score of .91 (Markelz et al., 2020). The adapted BSP-OT added a stu-dent on-task row to simultaneously collect teacher praise characteristic data (independent variable) as well as student on-task behavior data (dependent variable). The adapted BSP-OT is available in the supplemental file. The first

Markelz et al. 5

author trained the primary data collector (a graduate stu-dent). Following the 30-min training, the primary data col-lector watched a 15-min prerecorded video of a teacher/student dyad in an authentic classroom and used the adapted BSP-OT to data collect teacher praise characteristics and student on-task behavior. The primary data collector’s data was assessed for reliability against an accurate data sheet using kappa. A priori κ criterion of .80 was set. The primary data collector scored kappa reliability of .95.

Procedures

Baseline. Observation sessions were set to 15-min in length due to previous praise research on acceptable dependability and generalizability levels of teacher praise (Floress et al., 2021; Gage et al., 2014). A minimum of five baseline obser-vations were conducted, but the research team also analyzed data with a stability envelope criterion of 80% of data within 25% of the mean to ensure baseline data stability prior to intervention implementation (Lane & Gast, 2014).

Praise Training Condition. Once baseline conditions were met (i.e., student on-task behaviors were stable), the first author and primary data collector met with both teacher partici-pants and trained them using explicit instruction (Archer & Hughes, 2011) on the importance of specific and contingent praise (note: praise variety was not mentioned during this training). The first author trained teacher participants on the definition and use of BSP; research efficacies of BSP on student behaviors were highlighted, and the teacher partici-pants practiced delivering BSP with their target students’ name. The trainer delivered positive and corrective verbal feedback. Participants were then told that prior to each observation, they would be handed an Apple Watch™ (series 4) that had a Periodic Timer application (Kelin, 2014). The timer application was programmed to deliver a vibratory cue (i.e., tactile prompt) every 75 s. Previous research suggests tactile prompting effectively increases teacher praise rates (Markelz, Riden, & Hooks, 2021; Markelz, Taylor, et al., 2019).

Participants practiced putting the Apple Watch on and delivering BSP every tactile prompt. Participants were instructed to deliver specific praise to their target student following each tactile prompt, contingent on on-task behav-iors (e.g., following directions, participating in the activity). Participants were informed that they could deliver BSP at any time during observations; however, the Apple Watch prompted delivery at a minimum of 10 statements based on previous research on BSP in childhood settings (LaBrot et al., 2016), and previous recommendations of 6 to 10 praise statements per 15 min (Sutherland et al., 2000). The trainer instructed teacher participants that to reach mastery levels of praise delivery, they would need to meet the goal of at least 10 BSP statements per observation for 3 days

before entering the treatment condition. The training lasted approximately 30 min.

Treatment Condition. After each participant demonstrated mastery in delivering specific and contingent praise to their target student during observation sessions, an email was sent describing the difference between varied and non-var-ied praise. Teacher participants were informed that observa-tion sessions would now alternate between T1 and T2 conditions. When the data collector handed teacher partici-pants the Apple Watch prior to observation, they informed the teacher which treatment was in place for that session. A research team member randomly assigned T1 and T2 condi-tion orders without knowledge of which condition was var-ied or non-varied praise. To meet alternating treatments design standards without reservations (What Works Clear-inghouse, 2020), five sessions per treatment condition were scheduled with no more than two consecutive treatments.

Data Analysis

We used Barton’s (2021) visual analysis tool and Lanovaz et al.’s (2019) visual structured criterion (VSC) for alternat-ing treatment designs to determine the presence of a func-tional relation between experimental conditions and phases. The visual analysis tool (Barton, 2021) is a spreadsheet that guides users through the visual analysis process to help determine whether a functional relation is present and how confident the user can be in that determination. The tool prompts users to consider the overlap between conditions, differentiation between conditions, and the magnitude and trend of differentiation. We computed the nonoverlap of all pairs (NAP; Parker & Vannest, 2009) to assess overlap and judged differentiation between conditions by evaluating stability, trend, and effect size. We assessed stability by determining whether 80% of data points fell within 25% of the median of each phase (Lane & Gast, 2014) and used the split-middle method (White & Haring, 1980) to identify trend. To use the split-middle method, one finds the mid-rate, mid-date, and middle point of the mid-rate and mid-date for each phase; then draws a trend line between the middle point of the mid-rate and mid-date (Ledford & Gast, 2018).

We also used Lanovaz et al.’s (2019) visual structured criterion (VSC) for alternating treatments designs to sup-plement visual analysis. In contrast to nonoverlap methods used in Barton’s (2021) tool (i.e., NAP; Parker & Vannest, 2009), which compare overlap on a point-by-point basis and do not account for trend, the VSC compares the relative position of points and paths (Lanovaz et al., 2019). To use the VSC, one counts the number of times that a data path and/or points for one treatment condition fall above the path and points for a second treatment condition at each session. Then, that number is compared with a predetermined cutoff

6 Journal of Positive Behavior Interventions 00(0)

value based on the number of data points and comparisons (Lanovaz et al., 2019).

To report a quantitative index of effectiveness (Ledford & Gast, 2018) that captures the magnitude of behavior change between T1 and T2 conditions, we calculated the weighted average difference between success observations (ADISO; Manolov & Onghena, 2018). To do so, T1 and T2 data were entered into https://manolov.shinyapps.io/ATDesign/ where an ADISO score was calculated. The ADISO score is expressed in the same measurement units as the dependent variable. Thus, the ADISO score repre-sents a weighted change in the percentage of on-task behav-ior between T1 and T2 conditions.

Visual and Statistical Analysis Agreement

The second and third authors conducted independent visual analyses on the graphed data collected for each participant to determine if there were functional relations between baseline and treatment conditions and T1 versus T2 conditions. Barton’s (2021) visual analysis tool was used to examine level changes, trend, variability, overlap, and immediacy of change. The second author earned his PhD in special education and is a Doctoral Level Board Certified Behavior Analyst. The third author earned her Ph.D. in special education and completed the Institute of Education Sciences/The National Center for Special Education Research (IES/NCSER) Summer Research Training Institute on Single-Case Intervention Research Design and Analysis. After the independent visual analyses were completed, agreement across 66 coding opportunities within the visual analysis tool was 97%.

IOA

To ensure observers remained reliable using the BSP-OT, we gathered IOA data for student on-task behavior and praise characteristics across 27.3% (n = 6) of total observa-tions (n = 22) for Dyad A, and 36.8% (n = 7) of total obser-vations (n = 19) for Dyad B. Similar to observer training, IOA scores were calculated with a priori criterion of accept-able kappa reliability at .80. For Dyad A, baseline κ reli-ability was .86, .89, and .93 for 33% (n = 3) of observations. During the mastery training phase, overall κ was .91 for 33% (n = 1) of observations. T1 reliability for Dyad A was .93 for 25% (n = 1) of observations. T2 reliability for Dyad A was .92 for 25% (n = 1) of observations.

For Dyad B, baseline κ reliability was .95, .85, and .97 for 60% (n = 3) of observations. During the mastery train-ing phase, kappa reliability was .92 for 25% (n = 1) of observations. T1 reliability was .87 for 25% (n=1) and T2 reliability was .96 for 25% (n= 1) of observations.

Treatment Fidelity

During intervention conditions, data were analyzed daily to confirm the only praise characteristic manipulated was

varied and non-varied BSP. Based on previous research on low rate BSP (Jenkins et al., 2015; Markelz, Riden, Floress, et al., 2022) and previous praise rate recommendations (Floress et al., 2020), a minimum of 10 BSP statements toward the target student per 15-min session was set so that sufficient data could be observed. The number of GP state-ments was recorded to document if anomalous rates of GP occurred during any particular session. A post hoc paired samples t test was conducted on the number of GP state-ments between T1 and T2 for each Dyad. There was no sta-tistical difference between the two treatments for Dyad A, t(4) = −1.500, p = .208, and Dyad B, t(4) = −0.767, p = .486.

Contingent BSP was measured using the BSP-OT to confirm that consistent BSP contingency was being deliv-ered. Due to the assumption of independence of groups not being met, Wilcoxon signed-rank tests were used to exam-ine post hoc statistical differences in praise characteristics. Results indicated no statistical difference for BSP contin-gency between T1 and T2 for both Dyads (z = −.074, p = .941). In addition, the post hoc Wilcoxon signed-rank test indicated no significant differences for the number of BSP statements per T1 and T2 sessions for both Dyads (z = −.291, p = .771).

A post hoc paired samples t test did confirm a statistical difference of BSP variety between T1 and T2 for Dyad A, t(4) = −5.48, p = .005, and Dyad B, t(4) = −7.64, p = .002. The Wilcoxon signed-rank analysis confirmed this signifi-cant difference in praise variety between T1 and T2 (z = −2.316, p = .021). Treatment fidelity analyses suggest the only praise characteristic manipulated between T1 and T2 was BSP variety.

Results

Visual analysis indicates a positive effect of praise (both var-ied and non-varied) on participants’ on-task behavior, but fails to indicate a clear difference in effectiveness between T1 and T2 for either participant. Alli’s data are presented in Figure 1 and Brice’s data are presented in Figure 2. Teachers’ praise data (number of BSP statements and variety percent-age) are summarized in Tables 1 and 2.

Dyad A

During baseline, Ms. Alisha provided an average of less than 1 BSP statement to Alli per observation (range 0–2), and Alli was on-task an average of 49% of the time (range 27%–76%). Alli’s baseline data were variable, and a split middle trend analysis indicated a decelerating trend. During training, Ms. Alisha provided Alli an average of 13 BSP statements per observation (range 12–14) and her BSP statements averaged 41% variety (range: 36%–46%). Alli’s percentage of time-on-task increased to an average of 70% (range: 60%–82%).

Markelz et al. 7

Figure 1. Percentage of On-Task Behavior Across Conditions for Alli.

Figure 2. Percentage of On-Task Behavior Across Conditions for Brice.

During T1 sessions (i.e., varied praise), Ms. Alisha pro-vided an average of 14 BSP statements per observation (range: 13–15) with an average of 39% variety (range 33–46%). During T2 sessions (i.e., non-varied praise), she pro-vided an average of 13 BSP statements per observation (range 10–14) with an average of 9% variety (range: 0–15%). Alli’s average percentage of time on task rose to 80% (range 64–93%) during T1 sessions and 74% (range 62–80%) during T2 sessions. Split middle trend analyses indicated that Alli’s on-task behavior showed an accelerat-ing trend in both T1 and T2 sessions.

We used visual analysis guided by Barton’s (2021) visual analysis tool to determine that a functional relation was present between both treatment conditions and on-task behavior. For both T1 and T2, there was low overlap between data in the baseline and praise conditions

(NAP=93% for both comparisons), and changes in level and trend provided clear differentiation between baseline and T1 and T2. The level of on-task behavior increased in T1 and T2 and the trend changed from decelerating during baseline to accelerating.

Visual analysis and the VSC (Lanovaz et al., 2019) fail to indicate a functional relation between T1 and T2 condi-tions. There was a high degree of overlap between T1 and T2 data for Alli (NAP = 30%) and moderate overlap between T2 and T1 data (NAP = 70%). Given undifferenti-ated intervention conditions, the visual analysis tool (Barton, 2021) indicated no functional relation between T1 and T2 for Alli. The VSC (Lanovaz et al., 2019) supported this determination. The T1 data points and/or data path were higher than T2 in 7 out of 8 comparisons, which falls below the cutoff of 8 out of 8 comparisons (Lanovaz et al., 2019).

8 Journal of Positive Behavior Interventions 00(0)

Although a functional relation between treatments was unestablished, the ADISO calculation demonstrated that T1 had a +8.5% difference in on-task behavior compared with T2.

Dyad B

During baseline, Ms. Brenda provided an average of 1 BSP statement for Brice per observation (range 0–2). Brice was on-task an average of 28% of sessions (range 22–36%) and his baseline data were stable. During training, Ms. Brenda provided an average of 12 BSP per observation (range 8–13) and her BSP statements averaged 46% variety (range 37–58%). Brice’s percentage of time on task increased to an average of 41% during training (range 33–53%).

During T1 sessions (i.e., varied praise), Ms. Brenda pro-vided an average of 11 BSP statements per observation (range 9–13) with an average of 58% variety (range 45–66%). During T2 sessions (i.e., non-varied praise), she pro-vided an average of 12 BSP statements per observation (range 11–13) with an average of 9% variety (range 0–16%). Brice’s average percentage of time on task was 66% (range 42–84%) during T1 sessions and 64% (range 44–80%) dur-ing T2 sessions. Split middle trend analyses indicated that

Brice’s on-task behavior showed an accelerating trend in both T1 and T2 sessions.

Barton’s (2021) visual analysis tool suggested a func-tional relation between the intervention and on-task behav-ior. For both T1 and T2, there was no overlap between data in baseline and treatment conditions (NAP = 100% for both comparisons); and differences in level and trend provide clear differentiation between baseline and each treatment condition.

There was substantial overlap in Brice’s data in the two treatment conditions (NAP=46% for T1 vs. T2 and NAP=54% for T2 vs. T1) and undifferentiated conditions. As a result, the visual analysis tool (Barton, 2021) indicated no functional relation between T1 and T2. We were unable to confirm this determination using the VSC because there were 7 comparison points between conditions and 8 are required for VSC (Lanovaz et al., 2019). The AIDSO calcu-lation demonstrated that T1 had a +1.6% difference in on-task behavior compared with T2.

Discussion

Results from this study are the first in teacher praise research to examine the efficacy of praise variety on student on-task

Table 1. Praise Statement Data for Dyad 1: Ms. Alisha and Alli.

Session Phase TreatmentBSP

frequencyPercentage of variety

1 Baseline 0 02 Baseline 0 03 Baseline 1 04 Baseline 1 05 Baseline 0 06 Baseline 0 07 Baseline 0 08 Baseline 1 09 Baseline 0 010 Training 14 3611 Training 12 4112 Training 13 4613 Intervention T2 14 1414 Intervention T1 13 3815 Intervention T1 15 3316 Intervention T2 13 1517 Intervention T2 14 1418 Intervention T1 14 3619 Intervention T2 10 020 Intervention T1 15 4021 Intervention T2 12 022 Intervention T1 13 46

Note. BSP = behavior-specific praise; T1 = varied praise condition; T2 = non-varied praise condition.

Table 2. Praise Statement Data for Dyad 2: Ms. Brenda and Brice.

Session Phase TreatmentBSP

frequencyPercentage of variety

1 Baseline 0 02 Baseline 1 03 Baseline 2 04 Baseline 2 05 Baseline 1 06 Training 12 417 Training 8a 378 Training 12 589 Training 13 4610 Intervention T1 11 6311 Intervention T1 9a 6612 Intervention T2 11 013 Intervention T2 12 1614 Intervention T1 11 4515 Intervention T2 13 1516 Intervention T1 10 6017 Intervention T2 11 018 Intervention T1 13 5419 Intervention T2 12 16

Note. BSP = behavior-specific praise; T1 = varied praise condition; T2 = non-varied praise condition.aParticipant did not meet minimum BSP criterion and was retrained prior to subsequent session.

Markelz et al. 9

behaviors. Increases in student on-task behaviors between baseline and both T1 and T2 conditions support consensus in the research community that BSP may positively affect a variety of student behaviors (Ennis et al., 2020). Although praise research has been conducted for decades, few meth-odologically sound studies exist to classify BSP as an evi-dence-based practice according to the Council for Exceptional Children (CEC) and the What Works Clearinghouse quality indicators and standards (Moore et al., 2019). Results from this study contribute to the grow-ing body of rigorous research which supports BSP as a potential evidence-based practice according to the CEC guidelines (Royer et al., 2019).

The visual and statistical examination of varied versus non-varied praise on student on-task behaviors indicates no functional relation in efficacy. Varied praise did produce slightly higher on-task behavior percentages for both stu-dent participants; however, insufficient change to claim dif-ferences in efficacy. Given the novelty of praise variety research, we recommend further examination between var-ied and non-varied praise to build a larger literature base. Because challenges with classroom management contribute to teacher stress and decreased self-efficacy (Bottiani et al., 2019) and BSP is an effective, efficient, low-intensity strat-egy (Ennis et al., 2018), further research into BSP charac-teristics is warranted.

Future research should acknowledge a potential ceiling effect for varied praise percentages. Participants in this study used a total of 13 different praise statements across all sessions (i.e., good, great, awesome, thank you, appreciate, like, wonderful, love, nice, perfect, fantastic, excellent, super). On average, teacher participants used only five var-ied statements during T1 conditions. With a ceiling effect on the number of unique statements, fluctuations in total statements may misrepresent desired levels of percent var-ied. For example, if a teacher uses two unique BSP state-ments out of 4 total BSP statements; the percent varied for that session is 50%. However, if the teacher during the next observation uses five unique BSP statements, but 15 state-ments in total, the percent varied for that session is 33%. The use of more BSP statements is encouraged and aligned with best practices; however, the teacher is “penalized” with a lower percent variety. In other words, the more BSP statements a teacher delivers, the more difficult it is to have higher percentages of praise variety.

In addition to the ceiling effect on the number of unique statements, previous research suggests cognitive load may also play a factor in limiting the number of unique state-ments (Markelz, Scheeler, et al., 2019). In a classroom with competing stimuli, teachers are often focused on their peda-gogy, curricular content, and reacting to management issues (Maag, 2001). One explanation for consistently higher GP rates over BSP rates is that it takes less cognitive effort to deliver “Good job” as opposed to identifying a specific

behavior to praise (Markelz, Riden, Floress, et al., 2022). The same may be true for delivering unique BSP state-ments. Teachers may deliver repeated BSP statements because it is easier than producing a unique BSP statement every time. Based on these realities of delivering praise in a busy classroom, it is unrealistic to expect teachers to use proportionally higher numbers of unique statements as total statements increase.

In a study that examined teachers’ natural rates of praise characteristics, praise variety was around 50% on average (Markelz, Riden, Floress, et al., 2022). Participants within that study, however, delivered meager total BSP statements (M = 1.6 per 15-min observation). Thus, simply delivering two unique statements out of three total BSP statements would result in a 66% variety. Teacher participants in the current study averaged 39% and 58% during varied praise treatment conditions; yet, also averaged 14 and 11 BSP statements, respectively. The question arises, which is more important, praise variety or the number of BSP statements? Preliminary evidence from this study suggests the number of total BSP statements supersedes higher percentages of praise variety. However, we caution any definitive declara-tions on the importance of praise variety through descrip-tive words as a salient characteristic on student behaviors until more research is conducted.

Future researchers of praise variety should consider that although the BSP-OT measures praise variety by differen-tiation in descriptive words, Floress and Beschta (2018) identified various student behaviors as means of varying praise statements. For explicit operationalized behaviors (e.g., hand raising), descriptive word variation may be a more appropriate method of varying praise. For example, “Great job raising your hand before speaking,” and “I really appreciate you raising your hand and not calling out.” Yet for more ambiguous behaviors, like “on-task,” descriptive word differentiation and various student behaviors could contribute to praise variety. For example, a teacher could say “I love how focused you are on your assignment right now,” or “Great job working hard today.” Both praise state-ments target on-task behavior and use different descriptive words (i.e., love and great). The statements also praise two different behaviors (i.e., focused and working hard). Although this study lacked control for the praise of varied student behaviors (as long as they targeted on-task), future inquiry should examine possible interactions between these two modes of praise variety and their possible contribution to praise efficacy.

Implications

Current evidence that teachers may use low-rate BSP (Floress et al., 2022) suggests teachers might not connect best practice with actual practice. Multiple interventions have been documented to successfully increase teachers’

10 Journal of Positive Behavior Interventions 00(0)

BSP rates with components like didactic training, self-mon-itoring, performance feedback, and tactile prompting (Markelz et al., 2018; Zoder-Martell et al., 2019). Preservice teacher preparation programs or professional development for in-service teachers must provide opportunities for teach-ers to learn about and practice BSP to counter persistent low-level use. Future research should continue to examine the maintenance and generalized use of BSP following intervention to provide guidance on best practices in sus-tainable and efficient teacher training.

Markelz, Riden, Floress, and colleagues (2022) dis-cussed future research to determine normative guides of praise characteristics to provide recommendations to teach-ers and teacher trainers. Regarding frequency of BSP deliv-ery, ranges have been recommended to offer flexibility such as 6 to 10 BSP statements class wide per 15-min (Sutherland et al., 2000). Other researchers have recommended similar ranges like 18 to 30 BSP class-wide per hour (Floress et al., 2020). At the secondary level, researchers have recently identified using BSP once per 2 min as superior to once per 4 min in increasing academic engaged behavior and reduc-ing disruptive behavior (O’Handley et al., 2022). In line with previous research, we also recommend at least 1 BSP statement class-wide every 2 min. However, a denser sched-ule of reinforcement may be appropriate for a particular stu-dent given the frequency of student behavior and reinforcement goals.

When discussing a normative guide dictating optimal praise variety, either in professional development or in future research, we suggest using a ratio (unique BSP state-ments to total BSP statements) as opposed to a percentage. Unsimplified ratios are more descriptive than percentages. An unsimplified ratio provides the number of unique BSP statements as well as the number of total BSP statements, while percentages obscure actual counts and provide only a proportion of the total. Using un-simplified ratios under-scores the importance of the rate of praise as well as the proportion of unique praise statements. We acknowledge, however, that more research is needed to examine varying ratio effects on student behaviors before definitive guide-lines are established.

Limitations

There are several limitations to this study that require dis-cussion. First, the BSP-OT developed and assessed for reli-ability (Markelz et al., 2020) was amended for this study. To measure teacher praise characteristics and their effect on student on-task behavior, we added an additional row to the tool that allowed data recorders to capture student on-task data with whole interval recording. Tool amendment may have impacted reliability; however, acceptable IOA data suggest the tool was reliable and recorded teacher praise characteristics as well as student on-task behavior.

Second, we collected data using a whole interval recording approach which can underestimate the true occurrence of behavior and thus is likely to underestimate the level of the behavior (Fiske & Delmolino, 2012). A behavior is marked as occurring only if it occurs for the duration of the entire interval. For example, a student may have been on-task for 19-s within an interval, however, the 1-s non-response (i.e., off-task) is coded for the entire interval as a non-occurrence. In other words, on-task behaviors that occurred for a fraction of the interval escape capture. Although we considered an underestima-tion of behavior an acceptable limitation, future research should consider other data collection methods (e.g., par-tial interval recording).

Third, we used the VSC for alternating treatments design (Lanovaz et al., 2019) as a supplement to visual analysis in interpreting Alli’s data, but we were unable to use the pro-cedure for Bryce’s data. We followed single case design quality guidelines and used random assignment of treat-ment condition to session and collected data across five ses-sions for each intervention (T1 and T2). Based on random assignment, Bryce received T1 for his first two intervention sessions, but this session order in combination with the lim-ited number of total sessions resulted in only seven possible comparisons between T1 and T2. We were unable to use the VSC as a supplement for Bryce’s data as the VSC requires a minimum of 8 comparisons between conditions to make a determination about functional relation. We recommend researchers adhere to the comparison requirement or extend the total number of sessions.

Finally, we did not collect social validity data from teacher participants on their perceived differences between varied and non-varied praise conditions. Measuring teach-er’s acceptability of treatment conditions could have pro-vided a more nuanced examination of willingness to deliver higher percentages of praise variety. Future research should include social validity measures when conducting praise variety research to explore potential barriers like cognitive overload.

Conclusion

This study’s findings add to the evidence base supporting BSP as an effective practice to change student behavior. Participating teachers significantly increased their rate of BSP delivery following training, and visual analysis of results indicated that increased praise was correlated to increased on-task behavior for student participants. In addition, characteristics of praise, such as variety, may increase praise efficacy. Although a functional relation between varied and non-varied praise was not identified, preliminary results from this study warrant further inves-tigation into praise variety as a salient characteristic of BSP.

Markelz et al. 11

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Benjamin S. Riden https://orcid.org/0000-0002-6733-1942

Supplemental Material

Supplemental material for this article is available on the Journal of Positive Behavior Interventions website with the online version of this article.

References

Alberto, P. A., Troutman, A. C., & Axe, J. B. (2022). Applied behavior analysis for teachers (10th ed.). Pearson.

Archer, A., & Hughes, C. A. (2011). Explicit instruction: Efficient and effective teaching. Guilford Press.

Barton, E. E. (2021, March 19). Visual analysis tool. Vanderbilt University. https://lab.vanderbilt.edu/barton-lab/visual-anal-ysis-framework/

Bayat, M. (2011). Clarifying issues regarding the use of praise with young children. Topics in Early Childhood Special Education, 31(2), 121–128. https://doi.org/10.1177/0271121410389339

Becker, W. C., Madsen, C. H., Arnold, C. R., & Thomas, D. R. (1967). The contingent use of teacher attention and praise in reducing classroom behavior problems. The Journal of Special Education, 1(3), 287–307. https://doi.org/10.1177%2F002246696700100307

Bottiani, J. H., Duran, C. A., Pas, E. T., & Bradshaw, C. P. (2019). Teacher stress and burnout in urban middle schools: Associations with job demands, resources, and effective classroom practices. Journal of School Psychology, 77(1), 36–51. https://doi.org/10.1016/j.jsp.2019.10.002

Brophy, J. (1981). Teacher praise: A functional analysis. Review of Educational Research, 51(1), 532–532. https://doi.org/10.3102/00346543051001005

Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.). Pearson/Merrill-Prentice Hall.

Ennis, R. P., Royer, D. J., Lane, K. L., & Dunlap, K. D. (2020). Behavior-specific praise in PK-12 settings: Mapping the 50-year knowledge base. Behavioral Disorders, 45(3), 1–17. https://doi.org/10.1177%2F0198742919843075

Ennis, R. P., Royer, D. J., Lane, K. L., Menzies, H. M., Oakes, W. P., & Schellman, L. E. (2018). Behavior-specific praise: An effective, efficient, low-intensity strategy to support stu-dent success. Beyond Behavior, 27(3), 134–139. https://doi.org/10.1177/1074295618798587

Fiske, K., & Delmolino, L. (2012). Use of discontinuous meth-ods of data collection in behavioral intervention: Guidelines for practitioners. Behavior Analysis in Practice, 5(2), 77–81. https://doi.org/10.1007/BF03391826

Floress, M. T., Beaudoin, M. M., & Bernas, R. S. (2022). Exploring secondary teachers’ actual and perceived praise and repri-mand use. Journal of Positive Behavior Interventions, 24(1), 46–57. https://doi.org/10.1177%2F10983007211000381

Floress, M. T., & Beschta, S. L. (2018). An analysis of general education teachers’ use of diverse praise. Psychology in the Schools, 55(10), 1188–1204. https://doi.org/10.1002/pits.22187

Floress, M. T., Briesch, A. M., Jenkins, L. N., & Hampton, K. A. (2021). Teacher praise and reprimand: Examining the gen-eralizability and dependability of observational estimates. Behavioral Disorders, 47(3), 196–206. https://doi.org/10.1177%2F01987429211012020

Floress, M. T., Cates, G. L., Poirot, K. E., & Estrada, N. J. (2020). Conceptualizing fixed-interval praise delivery. Intervention in School and Clinic, 56(2), 84–91. https://doi.org/10.1177%2F1053451220914889

Floress, M. T., Jenkins, L. N., Reinke, W. M., & McKown, L. (2018). General education teachers’ natural rates of praise: A preliminary investigation. Behavioral Disorders, 43(4), 411–422. https://doi.org/10.1177%2F0198742917709472

Gage, N. A., & MacSuga-Gage, A. S. (2017). Salient classroom management skills: Finding the most effective skills to increase student engagement and decrease disruptions. Report on Emotional & Behavioral Disorders in Youth, 17(1), 19–24. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6345407/

Gage, N. A., Prykanowski, D., & Hirn, R. (2014). Increasing reliability of direct observation measurement approaches in emotional and/or behavioral disorders research using gen-eralizability theory. Behavioral Disorders, 39(4), 228–244. https://doi.org/10.1177%2F019874291303900407

Greer, R. D. (2002). Designing teaching strategies: An applied behavior analysis systems approach. Academic Press.

Hager, K. D. (2012). Self-monitoring as a strategy to increase student teachers’ use of effective teaching practices. Rural Special Education Quarterly, 31(4), 9–17. https://doi.org/10.1177%2F875687051203100403

Henderlong, J., & Lepper, M. R. (2002). The effects of praise on children’s intrinsic motivation: A review and synthe-sis. Psychological Bulletin, 128(5), 774–795. https://doi.org/10.1037/0033-2909.128.5.774

Iwata, B. A., DeLeon, I. G., & Roscoe, E. M. (2013). Reliability and validity of the functional analysis screening tool. Journal of Applied Behavior Analysis, 46(1), 271–284. https://doi.org/10.1002/jaba.31

Jenkins, L. N., Floress, M. T., & Reinke, W. (2015). Rates and types of teacher praise: A review and future directions. Psychology in the Schools, 52(5), 463–476. https://doi.org/10.1002/pits.21835

Kelin, V. (2014). Periodic timer (Version 4.1) [Mobile application software]. https://itunes.apple.com/us/app/periodic-timer/id933241656?mt=8

Klingbeil, D. A., Dart, E. H., & Schramm, A. L. (2019). A system-atic review of function-modified check-in/check-out. Journal of Positive Behavior Interventions, 21(2), 77–92. https://doi.org/10.1177%2F1098300718778032

LaBrot, Z. C., Pasqua, J. L., Dufrene, B. A., Brewer, E. A., & Goff, B. (2016). In situ training for increasing Head Start after-care

12 Journal of Positive Behavior Interventions 00(0)

teachers’ use of praise. Journal of Behavioral Education, 25(1), 32–48. http://doi.org/10.1007/s10864-015-9233-0

Lane, J. D., & Gast, D. L. (2014). Visual analysis in single case experimental design studies: Brief review and guidelines. Neuropsychological Rehabilitation, 24(3–4), 445–463. http://doi.org/10.1080/09602011.2013.815636

Lane, K. L., Menzies, H. M., Ennis, R. P., & Oakes, W. P. (2015). Supporting behavior for school success: A step-by-step guide to key strategies. Guilford Press.

Lanovaz, M. J., Cardinal, P., & Francis, M. (2019). Using a visual structured criterion for the analysis of alternating-treatments designs. Behavior Modification, 43(1), 115–131. https://doi.org/10.1177/0145445517739278

Ledford, J. R., & Gast, D. L. (Eds.). (2018). Single case research methodology: Applications in special education and behav-ioral sciences. Routledge.

Maag, J. W. (2001). Rewarded by punishment: Reflections on the disuse of positive reinforcement in schools. Exceptional Children, 67(2), 173–186. https://doi.org/10.1177%2F001440290106700203

Manolov, R., & Onghena, P. (2018). Analyzing data from single-case alternating treatments designs. Psychological Methods, 23(3), 480–504. https://doi.org/10.1037/met0000133

Markelz, A. M., Riden, B. S., Floress, M., Balint-Langel, K. B., Heath, J. A., & Pavelka, S. (2022). Teachers’ use of specific, contingent, and varied praise. Journal of Positive Behavior Interventions, 24(2), 110–121. https://doi.org/10.1177/1098300720988250

Markelz, A. M., Riden, B. S., & Hooks, S. D. (2021). Component analysis of training and goal setting, self-monitoring, and tactile prompting on early childhood educators’ behavior specific praise. Journal of Early Intervention, 43(2), 99–116. https://doi.org/10.1177/1053815120927091

Markelz, A. M., Riden, B. S., Zoder-Martell, K., Miller, J. E., & Bolinger, S. J. (2020). Reliability assessment of an observa-tion tool to measure praise characteristics. Journal of Positive Behavior Interventions, 23(1), 12–29. https://doi.org/10.1177%2F1098300720907988

Markelz, A. M., Scheeler, M. C., Riccomini, P. J., & Taylor, J. C. (2019). A systematic review of tactile prompting in teacher education. Teacher Education and Special Education, 43(4), 296–313. https://doi.org/10.1177%2F0888406419877500

Markelz, A. M., Scheeler, M. C., Taylor, J. C., & Riccomini, P. J. (2018). A review of interventions to increase behavior spe-cific praise. Journal of Evidence Based Practice for Schools, 17(1), 67–87.

Markelz, A. M., Taylor, J. C., Kitchen, T., Riccomini, P. J., Scheeler, M. C., & McNaughton, D. B. (2019). Effects of tactile prompting and self-monitoring on teachers’ use of behavior specific praise. Exceptional Children, 85(4), 471–489. https://doi.org/10.1177%2F0014402919846500

McKay, J. (1992). Building self-esteem in children. In M. McKay & P. Fanning (Eds.), Self-esteem (2nd ed., pp. 239–271). New Harbinger.

Moore, T. C., Maggin, D. M., Thompson, K. M., Gordon, J. R., Daniels, S., & Lang, L. E. (2019). Evidence review for teacher praise to improve students’ classroom behavior. Journal of Positive Behavior Interventions, 21(1), 3–18. https://doi.org/10.1177%2F1098300718766657

O’Handley, R. D., Olmi, D. J., Dufrene, B. A., Radley, K. C., & Tingstrom, D. H. (2022). The effects of different rates of behavior-specific praise in secondary classrooms. Journal of Positive Behavior Interventions. Advance online publica-tion. https://doi.org/10.1177%2F10983007221091330

Parker, R. I., & Vannest, K. (2009). An improved effect size for single-case research: Nonoverlap of all pairs. Behavior Therapy, 40(4), 357–367. https://doi.org/10.1016/j.beth.2008.10.006

Royer, D. J., Lane, K. L., Dunlap, K. D., & Ennis, R. P. (2019). A systematic review of teacher-delivered behavior-spe-cific praise on K–12 student performance. Remedial and Special Education, 40(2), 112–128. https://doi.org/10.1177%2F0741932517751054

Shriver, M. D., & Allen, K. D. (2008). Working with parents of noncompliant children: A guide to evidence-based par-ent training for practitioners and students. American Psychological Association.

Sutherland, K. S., Wehby, J. H., & Copeland, S. (2000). Effects of varying rates of behavior-specific praise on the on-task behavior of students with EBD. Journal of Emotional and Behavioral Disorders, 8(1), 2–8. https://doi.org/10.1177%2F106342660000800101

Thompson, R. F., & Spencer, W. A. (1966). Habituation: A model phenomenon for the study of neuronal substrates of behavior. Psychological Review, 73(1), 16–43. https://doi.org/10.1037/h0022681

What Works Clearinghouse. (2020). What works clearing-house standards handbook (Version 4.1). National Center for Education Evaluation and Regional Assistance, U. S. Department of Education, Institute of Education Sciences. https://ies.ed.gov/ncee/wwc/handbooks

White, M. A. (1975). Natural rates of teacher approval and disapproval in the classroom. Journal of Applied Behavior Analysis, 8(4), 367–372. https://doi.org/10.1901/jaba.1975.8-367

White, O. R., & Haring, N. G. (1980). Exceptional teaching (2nd ed.). Merrill.

Zoder-Martell, K. A., Floress, M. T., Bernas, R. S., Dufrene, B. A., & Foulks, S. L. (2019). Training teachers to increase behavior-specific praise: A meta-analysis. Journal of Applied School Psychology, 35(4), 309–338. https://doi.org/10.1080/15377903.2019.1587802