Check-In Check-Out + Social Skills

Participants and Setting

The study was conducted at a public elementary school serving Grades K–6 in a rural Mountain West town. During the 2012–2013 school year, approximately 600 students were enrolled in the school, 82% of which were White, and approximately 30% were eligible for free and reduced-price lunch. The school had been implementing Tier 1 PBIS for 3 years and met criteria (80/80) on the School-Wide Evaluation Tool (Sugai, Lewis-Palmer, Todd, & Horner, 2001) for the previous school year as completed by the district. The school had not implemented any Tier 2 behavior interventions prior to the study, and students exposed to the CICO and CICO+SS interventions continued to participate in all aspects of Tier 1 supports. The school was also implementing Response to Intervention (RtI), including Tier 2 academic interventions for students in need.

Four general education teachers at the school participated in the study, which included the one-time completion of a social skills assessment and the daily completion of a CICO DPR. All four teachers were White, and three of the four were female. Their ages ranged between 38 and 53, and all four teachers had been teaching at the school for at least 5 years.

In addition to the general education teachers, a student teacher enrolled at a nearby university delivered the basic CICO and CICO+SS intervention under the supervision of a cooperating teacher and a university supervisor. The student teacher (labeled the interventionist for the remainder of the article) was assigned to the special education resource room and had not interacted with nominated students prior to the intervention.

The school’s principal initially contacted researchers because of concerns related to increased problem behavior and indicated that several students in the school were in need of social skill training. Prior to the study, researchers met with all of the school’s teachers and the principal to discuss the project and request recommendations for students with social skill deficits. At that time, teachers and the principal were asked to nominate students who exhibited frequent problem behavior, had at least two ODRs by October, and lacked social skills in their interactions with other students. The Social Skills Rating System (SSRS; Gresham & Elliott, 1990) was shared with teachers at that time to provide examples of social skills that may be lacking. Any number of students could have been nominated, but the principal and her teachers indicated five students in particular who would most likely benefit from intervention: two fifth-grade girls from two separate classrooms, two third-grade girls from the same classroom, and one first-grade boy (participant information provided in Table 1). Once the parents and teachers of those five students provided active consent for involvement in the study and the students themselves provided verbal assent, the general education teachers of the five students completed the SSRS to confirm participant nomination. Only students who scored below the 15th percentile in overall social skills were included in the study.

OPEN IN VIEWER

Table 1.

Participant Information.

				SSRS
Name	Grade	Ethnicity	Problem behavior	percentile
Lucinda	3	Hispanic	Disruption	10th
			Disrespect
Sarah	3	White	Teasing	7th
			Physical aggression
Emily	5	Hispanic	Teasing	12th
			Gossip
Olivia	5	White	Teasing	10th
			Gossip
Tom	1	White	Non-compliance	12th
			Physical aggression

Note. SSRS = Social Skills Rating System.

Measurement

The primary dependent variable of interest in this study was the percentage of 10-s partial intervals during 5-min observations of lunch recess with positive and negative social engagement. Direct observations were conducted on a daily basis for each of the five participating students as well as non-target peers using the same recording method. This was done using a convenience sample of students for each grade chosen at random by the data collectors. Each day, a total of five non-target students were observed for 1 min each (total of 5 min). The observations were then combined to create a single “peer composite” score for comparison with engagement levels of target students.

Students were released to recess on an overlapping 20-min schedule: First and fourth grades were released first, followed by second and fifth grades, then third grade. To fit in all of the necessary observations, participating students and peer composites were observed daily in accordance with the recess schedule: Tom and a randomly selected first-grade peer were observed first; followed by Olivia, Emily, and two fifth-grade peers; followed by Lucinda, Sarah, and two third-grade peers. Observation of all target students occurred during the first 5 min of their recess or when they started recess following SST. If more than one participating student started recess at the same time (e.g., Olivia and Emily), they were observed in the order that they entered the recess playground.

All observations were conducted on the recess playground, which included one basketball court, one soccer field, and two jungle gyms. First- and second-grade students primarily played on the smaller of the two jungle gyms, while the older students were spread out across the other three environments.

Positive social engagement was the primary dependent variable used to make phase change decisions in the study and was defined as appropriate play or positive communication with peers ranging from neutral to complimentary. It included behaviors such as statements of approval, laughter, neutral statements, negotiations, feedback to peers, and appropriate complaints. It also included involvement in games where the participant played with other students appropriately. Negative social engagement was defined as inappropriate play or negative communication with peers. This included physical behaviors such as harmful bodily contact, contact with objects, or physical pestering. It also included verbal behaviors such as name-calling, bossy comments, statements of rejection, accusations, and aggressive threats. Only engagement with other peers was coded in the direct observations, and independent play was not coded as either negative or positive engagement.

The operational definitions and direct observation strategies implemented in the current study are similar to those defined in the Systematic Screening for Behavior Disorders (SSBD; Walker & Severson, 1992). However, it is important to note one specific difference between the SSBD protocol and the observation protocol used in the current study. In the SSBD, a negative dominance rule is used to prioritize negative social engagement: When negative and positive interactive responses occur within any given 10-s interval, only the negative response is recorded. In the current study, no negative dominance rule was used. If a student engaged in both positive and negative social engagement during a given 10-s interval, that interval was scored with both positive and negative social engagement. This procedure resulted in many intervals with both negative and positive social engagement, allowing researchers to evaluate the impact of the intervention on both positive and negative social interactions.

Interobserver Agreement

Data were collected by student researchers using handheld computers loaded with MOOSES software, which allowed collection and analysis of data in a real-time running data stream (Tapp, 2004). Observers participated in training on behavior operational definitions and use of the handheld computers for 1 hr per day on the participating school’s lunch-recess playground for 3 weeks prior to the study, and all observers met an 85% interobserver agreement criterion prior to initiation of the study. As opposed to practicing and meeting criteria on contrived video clips of behavior, practicing interobserver agreement in the actual setting provided the added benefits of seeing behavior that would be recorded and reducing the reactivity of students being observed.

Throughout the study, coders were scored as in-agreement if each observer recorded the same code(s) within each interval of the observation. Observations were assessed for interobserver agreement on at least 30% of observations for each phase for each participant. We calculated overall agreement, occurrence agreement, and non-occurrence agreement for positive social engagement and negative social engagement for each of the five target students and peer composites. Agreement had to meet or exceed 80% for the observation to be included. Throughout baseline, CICO, and CICO+SS, interobserver agreement ranged from 88% to 100% (M = 94%), and no observations were excluded due to low interobserver agreement.

Design and Procedure

A concurrent multiple baseline across students (MBL) design was used to assess the effects of the CICO+SS intervention on positive and negative social interactions during lunch recess. The design involved the following four phases: baseline, basic CICO, CICO+SS, and maintenance.

CICO+SS

CICO+SS was implemented if basic CICO was not effective in increasing positive social engagement or decreasing negative social engagement as indicated by the level, trend, and variability of results. Once a stable data trend was verified, CICO+SS was implemented sequentially, one student at a time, following a documented change in positive and negative social engagement.

Implementation of the added social skills components of CICO+SS took approximately 15 min each day and involved (a) meeting with the target student(s) 5 to 7 min prior to the beginning of lunch recess (students were pulled out of the cafeteria), (b) initiating explicit social skills instruction in a school conference room or meeting area, (c) gradually moving instruction from modeling to guided practice of specific skills on the playground during the first part of lunch recess (d) involving other children or adults in the guided practice, and finally (e) providing an assignment based on the instructional objective for completion prior to the check-out at the end of the school day.

When CICO+SS began, a set of discrete social skills were prioritized for instruction according to the SSRS results (see Table 2 for top five prioritized skills for each student). Skills with the greatest deficiency and skills rated as most important by teachers were taught first. During the intervention, students were allowed to leave 7 min early from lunch to join the interventionist in a conference room for SST. Most of the time, the interventionists worked with one student at a time. However, some of the skills overlapped multiple students in the same grade, in which case they were taught together (up to three students per group).

OPEN IN VIEWER

Table 2.

Specific Social Skills Deficits Identified by the Social Skills Rating System.

Target student	Social skill deficits
Lucinda	Responding appropriately to teasing (indicated for all five participants)
	Responding appropriately to peers
	Controlling temper in conflict situations (also indicated for Emily)
	Receiving criticism (also indicated for Sarah)
	Making friends
Sarah	Responding appropriately to teasing
	Responding appropriately when pushed or hit (also indicated for Emily and Olivia)
	Receiving criticism
	Cooperation with peers
	Joining activities (also indicated for Olivia)
Emily	Responding appropriately to teasing
	Responding appropriately when pushed or hit
	Inviting other to join activities
	Controlling temper in conflict situations
	Appropriately telling adults when she thinks she has been treated unfairly
Olivia	Responding appropriately to teasing
	Responding appropriately when pushed or hit
	Volunteering to help peers
	Joining activities
	Responding appropriately to peer pressure

Once in the conference room or meeting area, the interventionist implemented a short 5-min lesson plan with an effective teaching cycle, sometimes referred to as the gradual release of responsibility (Fisher & Frey, 2008). First, a description of the target social skill was provided, including a task analysis of specific behaviors within the skill. This included modeling at least three examples and non-examples of each behavior. Next, the interventionist provided a series of guided practice opportunities, slowly removing scaffolding based on student performance. During initial opportunities, the interventionist demonstrated the skill with little student input. However, once the student could demonstrate basic understanding, practice opportunities gradually shifted responsibility over to the student until mastery of the skill was acquired (at least three independent demonstrations). For example, in a lesson on using “stop” language to respond to peer teasing, the interventionist first demonstrated the “stop” skill while the student watched. Next, both the interventionist and the student practiced the skill together, followed by student demonstration while the interventionist coached. Finally, the student had to demonstrate the skill at least 3 times without any help from the interventionist before they could move on to the next phase of the lesson.

Once basic mastery of the skill was acquired, the remaining 10 min of the lesson was conducted in applicable contexts throughout the school. Usually this was done on the recess playground, but sometimes it occurred in other settings where the targeted skill was needed (e.g., general education classroom, cafeteria, hallway). Once in the setting, the interventionist provided scaffolded opportunities to practice the new skill along with peers or adults important for that skill. For example, if the interventionist was teaching responses to peer criticism, other students on the recess playground were asked to help practice the new skill with the participating student until mastery of the skill had been achieved (again, at least three independent correct displays of the skill). This stage of instruction was absolutely essential to ensure that the student became fluent in the target skill in the environment where the skill was needed.

Finally, at the end of the instructional period, the interventionist provided participants with an assignment for the rest of the day, which involved the application of the new skill. For example, if the student(s) had been learning a skill about inviting other students to play, the assignment might be to invite others to play 3 times before the end of recess. Then, the students would report back to the interventionist at check-out regarding their success on the assignment and would receive an additional point for “responsible” on their check-out score for their effort.

It is important to note that the number of days each skill was taught was dependent on the performance of the student. If a student struggled to demonstrate a skill independently without the help of the interventionist, that skill would be practiced again the next day. Some skills attained independent mastery within a few days, whereas others took as much as 2 weeks. None of the students demonstrated mastery of all five indicated skills within the study timeframe. Lucinda demonstrated mastery of her first four skills: responding appropriately to teasing, receiving criticism, responding appropriately to peers, and controlling temper in conflict situations. Sarah demonstrated mastery of three skills: responding appropriately to teasing, responding appropriately when pushed or hit, and receiving criticism. Emily demonstrated mastery of two skills: responding appropriately to teasing and responding appropriately when pushed or hit. Finally, Olivia only had time to demonstrate mastery of responding appropriately to teasing.

Fidelity of Implementation

We assessed the implementation fidelity of the basic CICO and the CICO+SS phases through a self-report checklist and a researcher direct observation checklist. First, a self-report checklist was completed daily by the interventionist in charge of the CICO and CICO+SS delivery. During the basic CICO phase, the interventionist indicated whether she (a) checked in with students at the beginning of school, (b) positively acknowledged students during check-in, (c) had each section of the DPR completed and signed by the general education teacher, (d) checked out with students at the end of the day, (e) recorded CICO points into an excel file, and (f) sent CICO data to the researchers. Once the CICO+SS phase was initiated, the interventionist also indicated whether she (g) provided a complete social skills lesson, (h) implemented modeling of at least three examples and non-examples, (i) implemented guided practice and independent practice in applicable settings with a demonstration of mastery, and (j) provided participants with a specific practice assignment for completion before check-out at the end of the day.

In addition to the interventionist daily self-report fidelity checklist, researchers also completed weekly observations of the interventionist to ensure accuracy of the self-report measure. Observations were conducted during check-in, check-out, and SST, with the researcher sitting or standing behind the CICO or SST session. During these observations, the researcher completed a checklist similar to that of the interventionist and scored agreement between the two measures.

Seven weekly fidelity observations were completed for each participant, but the type of observation differed as participants received CICO and CICO+SS for different amounts of time. Lucinda had two fidelity observations (29%) conducted during basic CICO and five (71%) during CICO+SS. Sarah had four CICO fidelity observations (57%) and three CICO+SS observations (43%). Emily had five CICO fidelity observations (71%) and two CICO+SS fidelity observations (29%). Finally, Olivia had six CICO fidelity observations (86%) and one CICO+SS fidelity observation (14%).

During the basic CICO phase, there was 100% agreement between the interventionist and the researchers. In addition, 88% (range = 84%–100%) of the CICO intervention components were completed correctly by the interventionist. For CICO+SS, 100% agreement between the interventionist and the researchers was maintained, and 94% (range = 90%–100%) of the CICO+SS intervention components were completed correctly by the interventionist. These results were consistent across participants, and the only components occasionally missed were in the areas of checking-in with the student at the beginning of school (sometimes the students came late) and ensuring the general education teacher completed the DPR (sometimes the classroom teacher forgot to score all areas of the card).

Social Validity

To assess social validity, researchers used a four-item questionnaire in paper format to both the teachers as well as the interventionist 2 months after implementation of the project. The forms were placed in the faculty member boxes, and they were asked to complete the form within a week and return them to an additional box the researchers provided. Questions on the form assessed the extent to which CICO+SS was perceived to (a) improve student behavior, (b) be worth the time and effort, (c) be worth recommending to others, and (d) be easy to implement. Scores on the questionnaire were recorded on a Likert-type scale from 1 to 6, with higher scores indicating a more favorable impression (strongly disagree, disagree, disagree somewhat, agree somewhat, agree, and strongly agree).

Baseline

The percentage of intervals of positive and negative social engagement during 5-min observations of lunch recess for each target student and their composite peers is presented in Figure 1. During baseline, when compared with peers, all five students displayed lower levels of positive social engagement and higher levels of negative social engagement than their peers. For positive social engagement, the five target students were positively engaged during 42.75% of intervals overall, compared with 85% for composite peers: 7.75% for Lucinda (range = 5%–12%), 28.75% for Sarah (range = 15%–45%), 69% for Emily (range = 61%–78%), 53.75% for Olivia (range = 45%–65%), and 54.5% for Tom (range = 50%–64%). For overall negative social engagement, the five target students were negatively engaged during 18.80% of baseline intervals overall, compared with 0.75% of baseline intervals for composite peers: 23.25% for Lucinda (range = 15%–31%), 17.25% for Sarah (range = 12%–25%), 14.5% for Emily (range = 8%–20%), 14.75% for Olivia (range = 10%–20%), and 24.25% for Tom (range = 15%–30%).

OPEN IN VIEWER

Figure 1.

Percent of 10-s intervals with positive and negative social engagement during 5-min observations of lunch recess.

Basic CICO

Once the teachers, staff, and target students had been trained on the basic CICO procedures, the study moved into the basic CICO phase with all five students. During the basic CICO phase, the mean percentage of intervals per observation for the first four students in basic CICO was 38.71% for positive social engagement (89.07% for peer composites) and 21.64% for negative social engagement (1.7% for peer composites). Unlike the other four target students, Tom responded very positively to the basic CICO intervention, with an overall 76.29% of intervals with positive social engagement and only 3.87% of intervals with negative social engagement. His positive social engagement increased immediately at the onset of basic CICO and ranged from 60% to 95% of intervals while maintaining an increasing trend. In addition, his negative social engagement dropped immediately once basic CICO was initiated and ranged from 0 to 12% of intervals for the remainder of the study. Because of these positive results, Tom was no longer indicated in need of social skills instruction and did not receive CICO+SS.

CICO+SS

Once the interventionist began the CICO+SS intervention, researchers immediately observed increases in positive social engagement (70.51% mean; 31.80 percentage increase) and decreases in negative social engagement (7.54% mean; 14.10 percentage decrease).

Mean positive social engagement during CICO+SS for Lucinda was 80.89% of intervals per observation (73.29 percentage increase), with a positively increasing trend. Mean positive social engagement during the CICO+SS phase for Sarah was 44% of intervals (24.05 percentage increase), again with a positively increasing trend for the remainder of the phase. For Emily, mean positive social engagement increased to 78.89% of intervals (12.35 percentage increase), and for Olivia, positive social engagement increased to an average of 78.25% (17.49 percentage increase).

Mean negative social engagement during the CICO+SS phase for Lucinda was 4.57% of intervals per observation (24.23 percentage decrease). Mean negative social engagement decreased to 4.46%, 8.63%, and 12.5% for Sarah, Emily, and Olivia, respectively. These levels represent reductions of 18.34, 11.18, and 2.64, respectively.

It is worth noting that the two fifth-grade participants (Emily and Olivia) displayed better behavior (higher positive social engagement and lower negative social engagement) during baseline than the other three participants. This may have been due to covert negative social interaction undetected by the direct observation data collection system, more positive relationships with peers, or both. Because of this higher percentage of appropriate behavior at baseline, positive and negative social engagement experienced a ceiling effect and reached similar levels of behavior as typically performing peers as a result of the intervention.

Maintenance

Last, once a functional effect had been demonstrated with each of the four target students in CICO+SS, all of the target students (including Tom) were moved into the maintenance phase for 1 week. During this phase, students no longer checked in with the interventionist, carried the DPR, received social skill instruction, or checked out at the end of the day. During the phase, the mean percentage of intervals per observation for the target students remained relatively stable with an overall of 73.83% for positive social engagement (3.33 percentage increase), and 6.62% for negative social engagement (0.92 percentage decrease).

Intervention Effect Size

In addition to the visual evidence demonstrating the efficacy of the CICO+SS intervention, intervention effect sizes were calculated using Tau-U (Parker, Vannest, Davis, & Sauber, 2011). Tau-U is an index for analysis of single-case research with more statistical power than any other non-overlap index currently known (Parker, Vannest, & Davis, 2011). It considers non-overlap between phases as well as trend within phases. Tau-U is based on all pairwise data comparisons made in a time-forward direction, the percentage of non-overlap minus overlap. To date, Tau-U has been primarily demonstrated with AB and ABA designs; however, the methods transfer easily to the MBL designs by combining the effects of each participant (Parker, Vannest, Davis, & Sauber, 2011).

To calculate Tau-U for the current study, all of the data points for the four participants who received CICO+SS were entered into an online Tau-U calculator (http://www.singlecaseresearch.org/calculators/tau-u; Vannest, Parker, & Gonen, 2011; see Table 3 for results). Baselines for both positive and negative social engagement were first evaluated for significant baseline trend. Only Participant 1’s baseline for negative social engagement had a Tau score greater than .40 (Tau = .44) and was corrected for baseline trend. Next, comparisons were calculated between baseline and intervention phases for each behavior and for each participant. Results indicated significant Tau-U scores on positive social engagement for all four participants and significant Tau-U scores on negative social engagement for all but one. Only Emily’s Tau-U score for negative social engagement failed to meet statistical significance (p = .10). Finally, the comparisons between phases were combined across all of the participants to result in an overall Tau-U score for the study for both positive social engagement (Tau = .90, p < .001) and negative social engagement (Tau = −.72, p < .001).

OPEN IN VIEWER

Table 3.

Pre–Post Means and Tau-U Results for Positive and Negative Social Engagement.

	CICO	CICO+SS
Participant	M (SD)	M (SD)	Pairs	SD	Tau	p
Positive social engagement
Lucinda	7.60% (2.60)	80.89% (15.68)	288	.22	1.00	<.001
Sarah	19.95% (9.73)	44.00% (10.78)	432	.18	.90	<.001
Emily	66.54% (10.18)	78.89% (3.72)	360	.20	.81	<.001
Olivia	60.76% (10.33)	78.25% (6.23)	297	.22	.89	<.001
Combined	38.71% (8.21)	70.51% (9.10)	1,377	.10	.90	<.001
Negative social engagement
Lucinda	28.80% (4.76)	4.57% (4.53)	288	.22	−1.05	<.001
Sarah	22.80% (11.04)	4.46% (4.42)	432	.18	−.89	<.001
Emily	19.81% (10.94)	8.63% (4.96)	360	.20	−.55	.006
Olivia	15.14% (6.74)	12.50% (6.80)	297	.22	−.36	.100
Combined	21.64% (8.37)	7.54% (5.18)	1,377	.10	−.72	<.001

Note. CICO = Check-In Check-Out; CICO+SS = Check-In Check-Out + Social Skills.

Social Validity

The four teachers along with the interventionist (five staff altogether) completed the social validity questionnaire on a 6-point scale, with higher scores indicating a higher satisfaction with the intervention components. The mean rating for “CICO+SS improved student behavior” was 5.2 (range = 4–6). The mean rating for “CICO+SS was worth the time and effort” was 5.6 (range = 5–6). The mean rating for “would recommend CICO+SS to others” was 5.6 (range = 5–6), and the mean rating for “ease of implementation” was 5.0 (range = 4–6).

Implications for Practice

The current study focused on the addition of social skill training to Tier 2 CICO. CICO provided frequent opportunities for pre-correction and student feedback. It also provided a DPR that was used to monitor performance. Although future research should evaluate SST without CICO, we found it essential for ongoing adjustments and sustained implementation. Also, because basic CICO was in place, the addition of CICO+SS remained low cost in terms of student time, staff time, and necessary resources.

When educators determine that a student is not responding to basic CICO, they should not only consider brief functional assessment to address alternate maintaining functions but also consider assessment to determine a lack of specific social skills. Once the needed skills have been identified, short (15 min), frequent (daily) SST sessions can be implemented, including (a) specific instructional objectives based on assessment results; (b) short lesson plans using conspicuous teaching strategies, mediated scaffolding, and frequent review; (c) planned generalization strategies including practice in the applied setting; and (d) ongoing progress monitoring of student performance. By maintaining a focus on specific objectives, daily assignments, and guided practice, the current study demonstrated that SST can increase student positive social engagement and decrease negative social engagement.

Limitations

Several limitations of the current study warrant discussion. First, we hypothesized that the target students suffered from acquisition deficits rather than (or in addition to) performance deficits. Because the focus of the study was on a novel approach to SST, no FBA was conducted. CICO+SS was designed as an efficient addition to basic CICO within a school-wide framework of PBIS; therefore, assessment and intervention components were added only when students failed to respond. All five students in the study were initially placed in basic CICO to determine whether it alone would result in a desired change in behavior (as it did for Tom). Then, when the remaining four students failed to respond, they were placed in the CICO+SS intervention. Even though all four responded positively to the added intervention, it is possible that they and other students may also benefit from an FBA and related functional modifications. Future research should compare the two approaches for students not responding to CICO. Efforts should also be made to efficiently combine the two strategies for assessment and intervention. For example, it may be efficient to begin with the basic FBA interview along with simple social skill assessment. Then, simple functional modifications and/or simple social skill training can be indicated for non-responding students. Finally, it is important to note that some students will fail to respond to both CICO+SS and functional modifications and should be considered for even more individualized supports, falling into the third tier of behavioral support with individualized assessment and intervention planning.

A second limitation of the study was the use of a student teacher as the interventionist. Because the social skill lessons of CICO+SS are based on specific student deficits and contexts, it is ideal to have the lessons developed by the interventionists rather than prescribed by a curriculum. To create an environment where this is feasible for all school personnel, we decided to use a student teacher with very little experience developing lesson plans or teaching social skills lessons. We trained her in very simple, explicit, and data-based strategies for the development of daily 15-min lessons. Future research should evaluate the implementation of CICO+SS when implemented by typical school staff, because it is not clear whether a high degree of implementation would still occur. However, the current study successfully demonstrated that CICO+SS can be effectively delivered by practitioners with minimal experience. The results of the social validity survey provide additional support for this feasibility.

The use of both a baseline phase and a basic CICO phase prior to the CICO+SS intervention phase can also be considered a design limitation. Because the CICO+SS phase is not immediately adjacent to the baseline phase, direct comparisons between the two phases cannot be made. Instead, results of the study can only demonstrate the functional effect of CICO+SS compared with basic CICO. Although future research should evaluate CICO+SS compared directly with baseline, there are two critical reasons we decided to compare CICO+SS with basic CICO instead. First, if behavior had only been measured in baseline and CICO+SS, it would be impossible to understand whether the CICO or SS component of CICO+SS had resulted in behavior change. Second, it was important to demonstrate the process of implementing Tier 2 supports and CICO+SS. When a student fails to respond to Tier 1 supports, schools should add the most efficient strategies that result in improved behavior (i.e., basic CICO). Then, if students are not successful, additional supports like CICO+SS can be added easily and systematically.

A fourth limitation of the study was its approach to measurement. Positive social engagement and negative social engagement are not common direct observation behaviors of interest. More typically, research on CICO and related interventions consider the number of incidents of problem behavior or student academic engagement. For the current study, we wanted to determine not only the quantity of student engagement but also the quality of those engagements. In addition, a partial interval recording system was used to measure social engagement. Partial interval recording systems tend to overestimate the occurrence of behavior and are most valuable for behaviors where the goal is behavioral reduction. This was particularly effective when measuring negative social engagement. Whole interval recording of positive social engagement would have been a more conservative measure, but it would have been near impossible to execute in the applied setting. It is recommended that future research address incidents of problem behavior and academic engagement in addition to positive and negative social engagement to allow for more comparison with previous research efforts.

Finally, direct observation in the current study was limited to 5-min observations of each participating student and composite peers. This was done so that observations of all five students could be completed within the allocated lunch-recess time period. Although the demonstration of a functional effect using such a limited sample of time reveals the potency of the intervention on playground behavior, the data collected may not be representative of behavior in other structured or unstructured settings in the school or outside the school environment. Future efforts should emphasize data collection on the generalization of skills to other environments, including direct observations of classroom behavior, cafeteria behavior, and so forth.