M4 Cultural Reflection

Vol.:(0123456789)

Journal of Behavioral Education https://doi.org/10.1007/s10864-020-09419-w

1 3

ORIGINAL PAPER

Impact of Language on Behavior Assessment Outcomes

Leslie Neely1 · Amarie Carnett1,4 · S. Shanun Kunnavatana2 · Jordan Wimberley3 · Katherine Cantrell1,3

Accepted: 23 October 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract Conducting a functional analysis is the “gold standard” of behavior assessment with best practices recommending identification of idiosyncratic variables as essential to valid results. One potential variable that might impact assessment results is language of assessment. For individuals who operate in environments with multiple languages (e.g., English and Spanish), the language of assessment might differentially impact assessment results. Therefore, there is a need to evaluate if language of assessment affects identified function. The current study presents the results of 10 cases in which the experimenters conducted assessments (i.e., functional analysis) in both the Spanish and English language. Participants were nine children with autism who engaged in problem behavior and whose parents reported Spanish as the primary home language. Result indicates correspondence of function for eight of the ten cases. Discussion of results and recommendations for practice and future research are presented.

Keywords Functional analysis · Language · Spanish

Introduction

Functional analyses have been well established as a valid way to identify the maintaining variables for behavior (Iwata and Dozier 2008). Conducting a func- tional analysis is often regarded as the standard for functional behavioral assess- ments. In 2003, Hanley, Iwata, and McCord laid out initial best practices for a

* Leslie Neely leslie.neely@utsa.edu

1 Department of Educational Psychology, University of Texas At San Antonio, 501 West Cesar E Chavez, San Antonio, TX 78207, USA

2 Easterseals UCP of North Carolina & Virginia, Raleigh, USA 3 Autism Treatment Center of San Antonio, San Antonio, USA 4 Present Address: University of Victoria of Wellington, Wellington, New Zealand

Journal of Behavioral Education

1 3

functional analysis, including attending to establishing operations before and during the assessment. In addition, they discussed the potential need to evalu- ate idiosyncratic environmental variables when functional analysis data are undif- ferentiated. Schlichenmeyer et al. (2013) reviewed 42 studies published between 2001 and 2010 that implemented idiosyncratic procedural modifications and identified more than 30 idiosyncratic variables that had been tested. For example, when testing for social negative reinforcement, the difficulty, preference for, and/ or amount of a task may be relevant, whereas the specific type of attention or duration of access to a preferred activity may be relevant when testing for social positive reinforcement. Although there are numerous idiosyncratic variables that can be tested for any given individual, the concern remains identifying only those that are relevant to the occurrence of problem behavior.

Schlichenmeyer et al. (2013) presented strategies used to identify the relevant idiosyncratic variables and concluded that a more systematic pre-assessment approach was warranted. However, it is common practice to modify the standard procedures so that antecedents and reinforcement mirror what has been reported in indirect assessments or direct assessments (Hanley et  al. 2003). For exam- ple, tasks used to test for behavior maintained by social negative reinforcement are typically those that have been identified by a stakeholder as tasks that occa- sion the target behavior. Similarly, the attention delivered contingent on problem behavior during an attention condition reflects how others commonly respond to the behavior in the natural environment (e.g., statements of concern, brief repri- mands, etc.). These types of variables are typically manipulated during a func- tional analysis from the outset. Another variable that may be critical to consider when developing initial assessment procedures is the language in which the func- tional analysis is conducted.

There has been an increased focus on linguistic diversity in the field of applied behavior analysis (e.g., Brodhead et al. 2014; Fong et al. 2016; Lim et al. 2018). However, limited research has been conducted evaluating the effect language has on the outcome of a functional analysis. For example, Rispoli et al. (2011) evalu- ated the language of implementation on the functional analysis results for one child who lived in a Spanish-speaking home. The results indicated that higher levels of responding were observed during the English attention and escape con- ditions in comparison with conditions where Spanish was spoken only. A study by Lang et al. (2011) found that lower rates of problem behavior occurred during instruction delivered in Spanish than English for a 4-year-old girl diagnosed with autism who lived in a Spanish-speaking home. Similarly, Aguilar et  al. (2017) evaluated language preference during instruction for five children with autism and found that four of the participant had clear preferences specifically during difficult tasks. Collectively, these studies suggest that language can alter the evoc- ative effect of antecedents and the reinforcing effect of consequences during a functional analysis.

In the current study, the experimenters evaluated the role of language on func- tional analysis outcomes for children with an autism spectrum disorder (ASD). In particular, this study sought to evaluate if the language of assessment effects the

1 3

Journal of Behavioral Education

identified function, and if language of assessment effects the patterns of behavior observed.

Method

Participants

The study procedures were approved by the Institutional Review Board at the lead author’s university prior to initiation of the study. Participants were included in this study if they (a) presented with problem behavior warranting a functional analysis, (b) had a guardian report the participant lived in household where Spanish was the dominant language, (c) had a guardian report the participant received instruction in a school or clinical setting in English, (d) had a guardian consent to the research procedures, and (e) responded correctly to one-step instructions in both languages (a simple test of receptive language). A total of 9 participants met the inclusion cri- teria for this study, one participant participated in two separate functional analysis (Gabe). Gabe was referred to this study for one topography of behavior. However, during the first functional analysis, a distinct secondary response class emerged, and a second FA was conducted for him. All had a medical diagnosis of ASD prior to participation in the study. A summary of the participant demographic information and operational definitions of their target behavior are provided in Table 1.

Settings and Materials

Experimenters conducted all of the sessions at two outpatient behavior analysis clin- ics. The assessment room at site one contained an adult sized chair, two child-sized chairs, a child-sized table, and the relevant session materials. The assessment room at site two contained one child sized chair, an adult sized chair, a long table, and relevant session materials. Relevant session materials included materials necessary for the social-positive conditions (e.g., preferred stimuli such as tablets, bubbles, doll house, and printing material), social-negative conditions (e.g., materials such as picture cards and dry erase board and markers), and non-vocal play (e.g., preferred stimuli such as tablets, puzzles, coloring books and markers, doll house, cars, play- doh, and trains).

Sessions were video-taped and one experimenter implemented the session (“implementer”), while a second experimenter videoed the session. For Gabe and Spencer, consent was not provided for video, and two data collectors were present during their sessions in addition to the implementer. The implementer spoke in only the relevant language (e.g., Spanish during the Spanish FA). Sessions were con- ducted 1 to 2 days per week. The FAs were conducted at an outpatient clinic over the course of three years. Individual FAs were conducted over the course of four weeks. Session length was kept consistent at 5 min.

Journal of Behavioral Education

1 3

Ta bl

e 1

P ar

tic ip

an t i

nf or

m at

io n

Pa rti

ci pa

nt A

ge (y

rs )

G en

de r

Ta rg

et b

eh av

io r

Ja ck

ie 5

Fe m

al e

Ag gr

es si

on : h

itt in

g, p

us hi

ng , o

r h ai

r p ul

lin g

Sc re

am in

g: v

oc al

iz at

io ns

a bo

ve c

on ve

rs at

io n

le ve

l Sl

at er

5 M

al e

Ag gr

es si

on : h

itt in

g, k

ic ki

ng , o

r p us

hi ng

A le

x 5

M al

e Ag

gr es

si on

: h itt

in g,

k ic

ki ng

, o r b

iti ng

C ar

lo s

5 M

al e

Ag gr

es si

on : h

itt in

g or

k ic

ki ng

Fl op

pi ng

: m ov

in g

fro m

a si

tti ng

to a

la yi

ng p

os iti

on o

r a st

an di

ng to

a la

yi ng

p os

iti on

Pr op

er ty

d es

tr uc

tio n:

th ro

w in

g m

at er

ia ls

, h itt

in g

or k

ic ki

ng su

rfa ce

s Je

re m

y 6

M al

e Ag

gr es

si on

: h itt

in g

or k

ic ki

ng Fl

op pi

ng : m

ov in

g fro

m a

si tti

ng to

a la

yi ng

p os

iti on

o r a

st an

di ng

to a

la yi

ng p

os iti

on G

ab e

#1 11

M al

e Ag

gr es

si on

: h itt

in g,

k ic

ki ng

, p us

hi ng

, o r s

cr at

ch in

g G

ab e

#2 11

M al

e Sc

re am

in g:

v oc

al iz

at io

ns a

bo ve

c on

ve rs

at io

n le

ve l

El is

e 4

Fe m

al e

Sc re

am in

g: v

oc al

iz at

io ns

a bo

ve c

on ve

rs at

io n

le ve

l; ex

cl ud

es si

ng in

g an

d vo

ca liz

at io

ns a

cc om

pa ni

ed b

y la

ug hi

ng Pr

op er

ty d

es tr

uc tio

n: th

ro w

in g

m at

er ia

ls A

lfo nz

o 5

M al

e Ag

gr es

si on

: h itt

in g,

k ic

ki ng

, p us

hi ng

, o r p

ul lin

g ha

ir Sp

en ce

r 8

M al

e Sc

re am

in g:

v oc

al iz

at io

ns a

bo ve

c on

ve rs

at io

na l l

ev el

Pr op

er ty

d es

tr uc

tio n:

th ro

w in

g m

at er

ia ls

, p us

hi ng

ta bl

e or

c ha

irs , o

r h itt

in g

or k

ic ki

ng su

rfa ce

C ur

si ng

: s ay

in g

ph ra

se s t

ha t i

nc lu

de th

e w

or d(

s) , "

f* **

," "s

** *,

" " b*

** *

1 3

Journal of Behavioral Education

Response Measurement and Interobserver Agreement

Problem behavior included aggression, screaming, flopping, property destruction, and cursing. The operational definitions for each behavior are provided in Table 1. Frequency data for the problem behavior were collected live during the FA was col- lected using paper and pencil method with the exception of the frequency data for Jackie, Slater, and Alex (which were recoded from video-taped sessions). Following completion of the FAs, trained observers collected latency data from video-taped sessions using the Countee application. The Countee application is a mobile appli- cation that allows for real-time data collection and development of individual data collection templates. Each individual template contained the participants’ initials and operational definitions of the target behaviors. The use of the Countee applica- tion allowed for analysis of latency to behavior during each condition. Data during Gabe’s FAs were collected using the Countee application rather than the paper and pencil method (which facilitated the latency analysis). Since Spencer’s FA was not video-taped, latency data are not available for his FA. Before collecting primary or reliability data, the observers trained on the data collection procedures until they reached 100% reliability for one session.

Interobserver Agreement

Two raters coded data using the Countee application for a minimum of 25% of ses- sions within each language for each participant (e.g., 25% of English sessions for Slater and 25% of Spanish sessions for Slater). The raw data were transposed into 10-s intervals to allow for calculation of interobserver agreement (IOA). The fourth author calculated the IOA using percent agreement. An agreement was scored if both raters coded the same frequency of behavior during a 10-s interval. A disa- greement was scored if the raters did not score the same frequency of behavior in the 10-s interval. The fourth author then divided the number of agreements by the sum of the agreements and disagreements and multiplied by 100 to obtain a meas- ure of IOA. The minimum resulting IOA across the participants was 96.3% (range 90–100%; Jackie’s Spanish FA), and the maximum resulting IOA was 100% (e.g., Slater’s English FA).

Experimenter Training and Procedural Fidelity

All of the research sessions were implemented by students enrolled in a master’s program who served as therapists at the outpatient clinics. The FAs were overseen by a Board Certified Behavior Analyst (BCBA) or Board Certified Behavior Ana- lysts—Doctoral (BCBA-D). All of the experimenters were trained prior to imple- menting study sessions using behavioral skills training (i.e., verbal and written instruction, modeling, role-play, and performance feedback). All of the experiment- ers were fluent in both English and Spanish.

Journal of Behavioral Education

1 3

In addition to training, two raters collected procedural fidelity for a minimum of 85% of participant sessions for each assessment (e.g., 85% of conditions for Jackie’s English FA and 85% of the conditions for Jackie’s Spanish FA). Raters collected procedural fidelity using procedural task lists developed by the lead author. Task lists included the implementers’ adherence to study procedures including mainte- nance of establishing operations, presentation of discriminative stimuli, prompting procedures, and presentation of consequent variables.

The task lists contained between four and six expected implementer behaviors (depending on the FA condition). Raters coded a “1” if the implementer behavior occurred and a “0” if it was absent. The total was summed, divided by the total number of expected behaviors, and multiplied by 100 to obtain a percentage. The minimum resulting fidelity was 96.3% (range 75–100%; Elise’s English FA), and the maximum resulting IOA was 100% (e.g., Spencer’s English FA).

Pre‑Assessment of Receptive Language

To assess for fluency of receptive language, the experimenters administered the Woodcock Johnson Tests of Oral Language IV, the Picture Vocabulary subtest (Eng- lish), and the Vocabulario sobre dibujos (Spanish) subtest to participants (Schrank, McGrew, and Mather 2014). Unfortunately, engagement in problem behavior resulted in termination of the test for all participants.

Procedure

Experimental Design The experimenters conducted two sequential FAs (one in each language). The experimenters conducted the assessments using an A-B design (“A” representing the assessment in the first language and “B” representing the assess- ment in the second language). Each assessment was conducted using an alternating treatment designs embedded within each phase. The presentation of language was randomized with nine of the participants randomized to receive the assessment in the English language first and one participant randomized to receive the assessment in the Spanish language (i.e., Alex). The lead author randomized the presentation of the conditions within each assessment, and the sequence was held consistent across the two languages.

Functional analysis The experimenters designed the FAs according to the pro- cedures described by Neely et  al. 2019 and adapted from Rispoli and colleagues (2011). Each FA included a test for social-positive reinforcement (e.g., access to attention), social-negative reinforcement (e.g., escape from demands or social atten- tion), and a control condition (e.g., non-vocal play condition). The FAs for some of the participants (i.e., Jackie, Alex, Carlos, Jeremy, Gabe, and Alfonzo) also included a second test for social-positive reinforcement in the form of access to preferred toys (e.g., access to tangibles) as this was indicated as a possible function by their clinical records. We did not include any test for automatically maintained behavior as it was

1 3

Journal of Behavioral Education

not hypothesized for any of the target behaviors. The experimenters wore colored shirts during the assessments that corresponded to the condition (e.g., blue for the play control condition; Conners et  al. 2000) to facilitate discrimination across the conditions. All vocal–verbal utterances made by the experimenters were matched for phonetic length (within one syllable) across the languages to control for response effort. The content of the vocal–verbal utterances was also held constant. For exam- ple, the experimenter might say “Do this” in the English language (two syllabus) and “Has esto” in the Spanish language (three syllables). The experimenter was held constant (i.e., the same experimenter implemented both the English and the Spanish FA for each participant).

Prior to initiating the assessments, the lead experimenter trained the rest of the research team on the FA protection procedures. For the child, the following was included or available during each session: (1) padding in the form of mov- able pads, (2) oversight by a doctoral level BCBA or onsite BCBA, (3) on-site physicians or nurses to respond to potential injury, and (4) skin checks following each session. For the experimenter, the following was included or available dur- ing each session: (1) protective clothing including arm and skin guards, (2) train- ing physical and crises management, (3) skin checks following session, and (4) weekly discussion with the lead experimenter to evaluate assessment risk.

Attention condition The experimenter initiated an attention condition by directing the participant to preferred toys. The experimenter then stated they needed to work (i.e., “Tu juega, yo trabajaré” or “You play, I will be doing work”), and removed attention by looking down at their work and turning their body slightly away from the participant. The experimenter did not respond to any non-target behavior. Con- tingent on the target behavior, the experimenter turned toward the child and pro- vided brief vocal–verbal attention (e.g., brief physical touch and a statement for a fixed interval of 10 s).

Non-vocal play During the non-vocal play condition, the participant had preferred stimuli freely available. The experimenter engaged with the participant following their lead, responding to all appropriate social interactions with non-vocal responses (e.g., smiles, head nodding, tickles, or high-fives) and provided attention on a fixed interval schedule (e.g., 10 s) of the same quality and topography (e.g., smiles, head nodding, tickles, or high-fives). The experimenter did not provide any consequences for target behavior.

Escape During the escape condition, the experimenter provided continuous pres- entation of the aversive stimuli (academic instruction or social attention as identi- fied during the clinical intake process). Experimenter vocalizations were matched in response efforts and phonetic length (within one syllable) across languages. If a vocal–verbal response was indicated, participants were required to respond in the relevant language. Contingent on problem behavior, the experimenter removed all relevant stimuli and turned away for a fixed interval (e.g. 10 s). Presentation of the aversive stimuli was reintroduced following a fixed interval without problem behav- ior. The experimenter restarted the interval if problem behavior occurred before the

Journal of Behavioral Education

1 3

interval elapsed. The procedures for escape from academic instruction and social attention were identical except, during presentation of academic demands, a least- to-most prompt hierarchy with an inter-prompt interval of 3  s was utilized for the presentation of the academic instruction.

Tangible During the tangible condition, the experimenter first provided 30  s of access to the preferred tangible. Preferred tangibles were identified using either a paired choice or multiple-stimulus with/without replacement preference assessment during the clinical intake process. After the brief access period, the experimenter retrieved the item and placed the item in sight but out of reach. The experimenter did not respond to any non-target behaviors (including attempts to get the item). Contingent upon the target behavior, the experimenter provided access to the item for 30 s. The tangible was retrieved after a fixed interval (e.g., 10 s).

Data Analysis

Following completion of the ten FAs, the lead experimenter recruited six BCBA- Ds with experience implementing functional analyses to analyze the resulting FA graphs using visual analysis. The six BCBA-Ds independently reviewed the graphs and provided their conclusion regarding identified function. Using the functions identified by the BCBA-Ds, the first and fourth author analyzed the data using descriptive analysis and effect size analysis. The descriptive analysis included cal- culation of the average latency to behavior and average rate of behavior during each condition conducted within the respective FAs (e.g., Spanish and English). The experimenters also conducted an effect size analysis for the identified functions using Tau-U effect size (Parker et al. 2011). Tau-U is a robust effect size that allows for greater precision as compared to other nonparametric effect sizes. Tau-U is also consistent with visual analysis with demonstrated convergent validity (Ninci et al. 2015). Data for each identified function were contrasted with the play-control condi- tion to calculate the corresponding effect size.

Results

The functional analysis graphs for Jackie, Alex, and Slater were originally published in Neely et al. 2019. However, the experimenters recoded the videos to obtain rate of behavior (with the exception of Jackie English session 2 and Slater English session 6 as the files were corrupted). Therefore, the graphs for the all ten cases are presented in Fig. 1 (Jackie, Slater, Alex, Carlos, Jeremy, and Gabe #1) and Fig. 2 (Gabe #2, Elise, Alfonzo, and Spencer) Fig. 3.

Identification of Function and Correspondence of Function

The functions identified by the BCBA-Ds are presented in Table  2. The BCBA- Ds indicated that eight of the ten FAs demonstrated correspondence of functions

1 3

Journal of Behavioral Education

between the languages. For multiple-function FAs, there was some discrepancies of identified functions with three of the BCBA-Ds indicating correspondence of one function but not the second function (e.g., BCBA-D#1 indicated a tangible function for Alfonzo in the English language and a tangible and escape function in the Span- ish section). All BCBA-Ds indicated that Spencer’s FAs results did not correspond across functions. For the purpose of analysis, the experimenters adopted the func- tions identified by the majority of the BCBA-Ds (i.e., 50% or higher agreement).

Fig. 1 Functional analysis results in English and Spanish languages for Jackie, Slater, and Alex

Journal of Behavioral Education

1 3

Fig. 2 Functional analysis results in English and Spanish languages for Carlos, Jeremy, and Gabe

1 3

Journal of Behavioral Education

Descriptive Results

The lead experimenter and fourth author conducted the descriptive analysis for each graph. The data were analyzed for average rate of behavior for each identified func- tion, average latency to behavior for each identified function, and effect size. The results for the descriptive analysis are presented in Table 3.

For the first graph, Jackie engaged in elevated levels of behaviors during the English FA in the tangible condition (M = 4.2 behaviors/min; range 3.2–4.6 behaviors/min) and escape condition (M = 5.0 behaviors/min; range 3.6–6.2 behaviors/min), with behavior not observed in the attention and non-vocal play conditions. During the Spanish FA, Jackie engaged in elevated levels of behav- ior during the tangible condition (M = 5.1 behaviors/min; range 4.6–6 behaviors/ min) and escape condition (M = 5.2 behaviors/min; range 3.6–6 behaviors/min).

Fig. 3 Functional analysis results in English and Spanish languages for Elise, Alfonzo, and Spencer

Journal of Behavioral Education

1 3

Ta bl

e 2

R es

ul ts

o f v

is ua

l i ns

pe ct

io n

an d

id en

tifi ca

tio n

of fu

nc tio

n

B C

BA -D

# 1

B C

BA -D

# 2

B C

BA -D

# 3

B C

BA -D

# 4

B C

BA -D

# 5

B C

BA -D

# 6

N am

e Fu

nc tio

n En

gl is

h Fu

nc tio

n Sp

an is

h Fu

nc tio

n En

gl is

h Fu

nc tio

n Sp

an is

h Fu

nc tio

n En

gl is

h Fu

nc tio

n Sp

an is

h Fu

nc tio

n En

gl is

h Fu

nc tio

n Sp

an is

h Fu

nc tio

n En

gl is

h Fu

nc tio

n Sp

an is

h Fu

nc tio

n En

gl is

h Fu

nc tio

n Sp

an is

h

Ja ck

ie Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Sl

at er

Es ca

pe Es

ca pe

Es ca

pe Es

ca pe

Es ca

pe Es

ca pe

Es ca

pe Es

ca pe

Es ca

pe Es

ca pe

Es ca

pe Es

ca pe

A le

x Ta

ng ib

le Ta

ng ib

le Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le Ta

ng ib

le Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe C

ar lo

s Ta

ng ib

le Ta

ng ib

le Ta

ng ib

le

es ca

pe Ta

ng ib

le Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe U

nc le

ar Ta

ng ib

le Ta

ng ib

le Ta

ng ib

le Ta

ng ib

le

es ca

pe Ta

ng ib

le

G ab

e( FA

1)

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

Ta ng

ib le

G ab

e( FA

2)

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe

fro m

at

te nt

io n

Es ca

pe fr

om

at te

nt io

n

El is

e A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

cc es

s t o

R R

B A

lfo nz

o Ta

ng ib

le Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Je

re m

y Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Ta

ng ib

le

es ca

pe Sp

en ce

r In

co nc

lu -

si ve

In co

nc lu

- si

ve In

co nc

lu -

si ve

A tte

nt io

n es

ca pe

In co

nc lu

- si

ve A

tte nt

io n

es ca

pe In

co nc

lu -

si ve

A tte

nt io

n es

ca pe

In co

nc lu

- si

ve A

tte nt

io n

A tte

nt io

n es

ca pe

A tte

nt io

n es

ca pe

1 3

Journal of Behavioral Education

Ta bl

e 3

D es

cr ip

tiv e

an al

ys is

N am

e Fu

nc tio

n En

gl is

h Fu

nc tio

n Sp

an is

h A

ve ra

ge R

at e

En gl

is h

(b eh

av io

rs /m

in )

A ve

ra ge

R at

e Sp

an is

h (b

eh av

io rs

/m in

)

A ve

ra ge

La

te nc

y En

gl is

h (s

)

A ve

ra ge

L at

en cy

Sp an

is h

(s )

Ta u-

U En

gl is

h [C

I 9 0]

Ta u-

U Sp

an is

h [C

I 9 0]

Ja ck

ie Ta

ng ib

le Es

ca pe

Ta ng

ib le

Es ca

pe Ta

ng ib

le : 4

.2 Es

ca pe

: 5 .0

Ta ng

ib le

: 5 .1

Es ca

pe : 5

.2 Ta

ng ib

le : 3

.7 Es

ca pe

: 5 .4

Ta ng

ib le

: 2 .0

Es ca

pe : 2

.7 Ta

ng ib

le 1

.0 [0

.3 , 1

] Es

ca pe

:1 .0

[0 .4

, 1 ]

Ta ng

ib le

:1 .0

[0 .4

, 1 ]

Es ca

pe :1

.0 [0

.4 , 1

] Sl

at er

Es ca

pe Es

ca pe

5. 5

3. 6

15 .4

12 .8

1. 0

[0 .3

, 1 ]

1. 0

[0 .4

, 1 ]

A le

x Ta

ng ib

le Es

ca pe

Ta ng

ib le

Es ca

pe Ta

ng ib

le : 4

.6 Es

ca pe

: 1 .4

Ta ng

ib le

: 2 .7

Es ca

pe : 0

.9 Ta

ng ib

le : 0

.8 Es

ca pe

: 3 7.

0 Ta

ng ib

le : 4

.6 Es

ca pe

: 7 4.

9 Ta

ng ib

le : 1

.0 [0

.3 , 1

] Es

ca pe

: 1 .0

[0 .3

, 1 ]

Ta ng

ib le

: 1 .0

[0 .4

, 1 ]

Es ca

pe : 1

.0 [0

.4 , 1

] C

ar lo

s Ta

ng ib

le Es

ca pe

Ta ng

ib le

Ta ng

ib le

: 1 .8

Es ca

pe : 1

.0 Ta

ng ib

le : 1

.0 Ta

ng ib

le : 1

68 .3

Es ca

pe : 1

99 .5

Ta ng

ib le

: 8 6.

2 Ta

ng ib

le : 0

.5 [-

0. 2,

1 ]

Es ca

pe : 0

.6 [0

.0 , 1

] Ta

ng ib

le : 0

.8 [0

.2 , 1

]

Je re

m y

Ta ng

ib le

Es ca

pe Ta

ng ib

le Es

ca pe

Ta ng

ib le

: 3 .4

Es ca

pe : 1

.6 Ta

ng ib

le : 4

.7 Es

ca pe

: 1 .8

Ta ng

ib le

: 4 .6

Es ca

pe : 2

9. 3

Ta ng

ib le

: 6 .0

Es ca

pe : 5

4. 3

Ta ng

ib le

:1 .0

[0 .4

, 1 ]

Es ca

pe : 1

.0 [0

.4 , 1

] Ta

ng ib

le :1

.0 [0

.4 , 1

] Es

ca pe

: 0 .9

[0 .3

, 1 ]

G ab

e (F

A 1

) Ta

ng ib

le Ta

ng ib

le 3.

0 4.

9 3.

1 2.

7 1.

0 [0

.2 , 1

] 1.

0 [0

.3 , 1

]

G ab

e (F

A 2

) Es

ca pe

fr om

So ci

al A

tte nt

io n

Es ca

pe fr

om

So ci

al A

tte n-

tio n

1. 9

2. 0

6. 6

2. 6

1. 0

[0 .4

, 1 ]

1. 0

[0 .4

, 1 ]

El is

e A

cc es

s t o

R R

B A

cc es

s t o

R R

B 2.

5 2.

7 4.

8 2.

9 1.

0 [0

.2 , 1

] 1.

0 [0

.2 , 1

] A

lfo nz

o Ta

ng ib

le Es

ca pe

Ta ng

ib le

Es ca

pe Ta

ng ib

le : 5

.1 Es

ca pe

: 1 .7

Ta ng

ib le

: 6 .8

Es ca

pe : 4

.8 Ta

ng ib

le : 4

.4 Es

ca pe

: 1 32

Ta ng

ib le

: 3 .9

Es ca

pe : 5

7. 2

Ta ng

ib le

: 1 .0

[0 .4

, 1 ]

Es ca

pe : 0

.6 [0

.0 , 1

] Ta

ng ib

le : 1

.0 [0

.4 , 1

] Es

ca pe

: 1 .0

[0 .4

, 1 ]

Sp en

ce r

In co

nc lu

si ve

A tte

nt io

n Es

ca pe

N /A

A tte

nt io

n: 1

.5 Es

ca pe

: 1 .2

N /A

N ot

A va

ila bl

e N

/A Es

ca pe

: 0 .9

[0 .2

, 1 ]

A tte

nt io

n: 0

.8 [0

.1 , 1

]

Journal of Behavioral Education

1 3

Behaviors were not observed in the attention or non-vocal play conditions. Both the English FA and Spanish FA indicated a social-positive and social-negative function with behaviors occurring at a higher rate in the Spanish FA. The average latency to behavior was less in the Spanish FA (2.0 s during the tangible condi- tion and 2.7 s during the escape condition) than the English FA (3.7 s during the tangible condition and 5.4 s during the escape condition). The Tau-U effect sizes were all 1.0 indicating clear differentiation between the identified function and play-control condition.

Slater engaged in elevated levels of behaviors during the English FA in the escape condition (M = 5.5 behaviors/min; range 3.0–7.8 behaviors/min), with behavior not observed in the attention and non-vocal play conditions. During the Spanish FA, Slater engaged in elevated levels of behavior during the escape condition (M = 3.6 behaviors/min; range 2.4–4.6 behaviors/min), with behaviors not observed in the attention or non-vocal play conditions. While both the English FA and Spanish FA indicated a social-negative function, Slater had lower rates of behavior in the Span- ish FA. The average latency to behavior was less in the Spanish FA (12.8 s during the escape condition) than the English FA (15.4 s during the escape condition). The Tau-U effect sizes were all 1.0 indicating clear differentiation between the identified function and play-control condition.

During the English FA, Alex engaged in elevated levels of behavior during the tangible condition (M = 4.6 behaviors/min; range 3.4–5.6 behaviors/min) and escape condition (M = 1.4 behaviors/min; range 0.6–3.0 behaviors/min), with behaviors not observed in the attention or non-vocal play conditions. Alex engaged in elevated lev- els of behaviors during the Spanish FA in the tangible condition (M = 2.7 behaviors/ min; range 1.2–4.2 behaviors/min) and escape condition (M = 0.9 behaviors/min; range 0.2–1.4 behaviors/min), with behavior not observed in the attention and non- vocal play conditions. While both the English FA and Spanish FA indicated a social- positive and social-negative function, the Spanish FA had lower levels of behav- ior across both the tangible and escape condition. The average latency to behavior was less in the English FA (0.8 s during the tangible condition and 37.0 s during the escape condition) than the Spanish FA (4.6 s during the tangible condition and 74.9 s during the escape condition). The Tau-U effect sizes were all 1.0 indicating clear differentiation between the identified functions and play-control condition.

Carlos engaged in elevated language during the English FA in the tangible condi- tion (M = 1.8 behaviors/min; range 0.0–4 behaviors/min) and the escape condition (M = 1.0 behaviors/min; range 0.0–3.2 behaviors/min), with low levels of behav- ior in the attention and non-vocal play conditions. During the Spanish FA, Carlos engaged in elevated levels of behavior during the tangible condition (M = 1.0 behav- iors/min; range 0.0–1.8 behaviors/min). Zero levels of behavior were observed in the attention, escape, and non-vocal play conditions. While the English FA indi- cated a social-positive and social-negative function, the Spanish FA indicated only a social-positive function. The average latency to behavior was less in the Spanish FA (86.2 s during the tangible condition) than the English FA (168.3 s during the tan- gible condition and 199.5 s during the escape condition). The resulting Tau-U effect sizes were 0.5 CI90 [− 0.2, 1] for the English tangible condition and 0.6 CI90 [0.0, 1] for the English escape condition indicating some overlap between the identified

1 3

Journal of Behavioral Education

functions and the play-control condition. The resulting Tau-U effect size for the Spanish tangible condition was 0.8 CI90 [0.2, 1] indicating clear discrimination between the identified function and the play-control condition.

Jeremy engaged in elevated levels of behaviors during the English FA in the tan- gible condition (M = 3.4 behaviors/min; range 3.0–3.8 behaviors/min) and escape condition (M = 1.6 behaviors/min; range 1.0–2.6 behaviors/min), with very low lev- els in one session of attention and zero levels in non-vocal play conditions. During the Spanish FA, Jeremy engaged in elevated levels of behavior during the tangible condition (M = 4.7 behaviors/min; range 4.0–5.4 behaviors/min) and escape condi- tion (M = 1.8 behaviors/min; range 0.2–4.6 behaviors/min), with minimal behavior in the attention and non-verbal play condition. Both the English FA and Spanish FA indicated a social-positive and social-negative function. The average latency to behavior was less in the English FA (4.6 s during the tangible condition and 29.3 s during the escape condition) than the Spanish FA (6.0 s during the tangible condi- tion and 54.3 s during the escape condition). The Tau-U effect sizes between 0.9 and 1.0 indicating clear differentiation between the identified functions and play-control condition.

Gabe engaged in elevated levels of behaviors during the English FA in the tan- gible condition (M = 3.0 behaviors/min; range 2.0–4.0 behaviors/min) with zero or near zero levels of behavior in the attention, escape, and non-vocal play conditions. During the Spanish FA, Gabe engaged in elevated levels of behavior during the tan- gible condition (M = 4.9 behaviors/min; range 4.0–6.6 behaviors/min) with near zero to near zero levels in the attention, escape, and non-vocal play. Both the English and Spanish FA indicate a social-positive function with higher rates of behavior in the Spanish FA. The average latency to behavior was less in the Spanish FA (2.7 s dur- ing the tangible condition) than the English FA (3.1 s during the tangible condition). The Tau-U effect sizes were 1.0 indicating clear differentiation between the identi- fied functions and play-control condition.

In his second FA, Gabe engaged in elevated levels of behaviors during the Eng- lish FA in the escape from social attention condition (M = 1.9 behaviors/min; range 1.4–2.2 behaviors/min), with near zero levels of behavior in the non-vocal play con- ditions. During the Spanish FA, Gabe engaged in elevated levels of behavior dur- ing the escape from social attention condition (M = 2.0 behaviors/min; range 1.6–2.2 behaviors/min) with near zero levels in the non-vocal play. Similar to his first FA, in both the English and Spanish FA indicate a social-negative function with similar rates of behavior in both the English and Spanish FA. The average latency to behav- ior was less in the Spanish FA (2.6 s during the escape from social attention condi- tion) than the English FA (6.6 s during the escape from social attention condition). The Tau-U effect sizes were 1.0 indicating clear differentiation between the identi- fied functions and play-control condition.

Elise engaged in elevated levels of behaviors during the English FA in the access to restrictive and repetitive behavior condition (M = 2.5 behaviors/min; range 2.0–2.8 behaviors/min) with zero levels of behavior during non-vocal play, escape, and attention conditions. During the Spanish FA, Elise engaged in elevated levels of behavior during the access to restrictive and repetitive behavior condition (M = 2.7 behaviors/min; range 2.6–3.0 behaviors/min) with zero levels in the attention,

Journal of Behavioral Education

1 3

escape, and non-vocal play. In both, the English and Spanish FA indicate a social- positive function with similar rates of behavior in both the English and Spanish FA. The average latency to behavior was less in the Spanish FA (2.9 s during the escape from social attention condition) than the English FA (4.8 s during the escape from social attention condition). The Tau-U effect sizes were 1.0 indicating clear differen- tiation between the identified functions and play-control condition.

Alfonzo engaged in elevated levels of behaviors during the English FA in the tan- gible condition (M = 5.1 behaviors/min; range 3.2–6.4 behaviors/min) and escape condition (M = 1.7 behaviors/min; range 0.0–3.8 behaviors/min), with behavior not observed in the attention and non-vocal play conditions. During the Spanish FA, Alfonzo engaged in elevated levels of behavior during the tangible condition (M = 6.8 behaviors/min; range 2.4–8.8 behaviors/min) and escape condition (M = 4.8 behaviors/min; range 0.8–8.8 behaviors/min) with behaviors not observed in the attention or non-vocal play conditions. While both the English FA and Spanish FA indicated a social-positive and social-negative function, the Spanish FA had higher levels of behavior with more variability in the social negative condition. The average latency to behavior was less in the Spanish FA (3.9 s during the tangible condition and 57.2 s during the escape condition) than the English FA (4.4 s during the tangi- ble condition and 132 s during the escape condition). The Tau-U effect sizes were 1.0 for the English tangible function, Spanish tangible function, and Spanish escape function indicating clear differentiation between the identified functions and play- control condition. The Tau-U effect size was 0.6 CI90 [− 0.03, 1.0] for the English escape condition, indicating some data overlap between the identified function and the play-control condition.

The majority of the BCBA-Ds indicated that the results for Spencer’s English FA were inconclusive. During the Spanish FA, Spencer engaged in elevated levels of behavior during the attention condition (M = 1.5 behaviors/min; range 0.4–3.2 behaviors/min) and escape condition (M = 1.2 behaviors/min; range 0.6–1.6 behav- iors/min) with some behavior observed in the non-vocal play condition (M = 0.3 behaviors/min; range 0.0–1.2 behaviors/min). The latency analysis was not per- formed for Spencer’s FA, as the FA was not videoed, and data were originally col- lected via paper and pencil (rather than the Countee application). While the English FA was inconclusive, the Spanish FA indicated a social-positive and social-negative function. The Tau-U effect sizes were 0.9 CI90 [0.18, 1] for the Spanish escape and 0.8 CI90 [0.13, 1] for the Spanish attention indicating clear differentiation between the identified functions and play-control condition.

Discussion

The current study extends the FA literature by evaluating the effects of language (e.g., Spanish vs. English) for dual language learners (Lang et al. 2011; Rispoli et al. 2011; Schlichenmeyer et al. 2013). Specifically, eight of the ten FAs conducted in this study showed correspondence of function across languages. Results also indi- cate differential rates of behavior and latencies to behavior across languages. Taken as a whole, these findings highlight the need to assess specific cultural variables,

1 3

Journal of Behavioral Education

such as language, to ensure assessment of the relevant EOs. Further, these results extend previous research evaluating English and Spanish conditions during an FA.

Of the ten FAs conducted, eight demonstrated correspondence across languages. These findings indicate that behavior may serve the same function across languages for some individuals from dual-language environments. However, for two of the FAs conducted in this study, the behavior function did not correspond across language. The differential findings are consistent with the findings of the Rispoli et al. (2011) and Lang et al. (2011) where different functions were observed in the English lan- guage as compared to the Spanish language. Differential results may correlate with participant fluency with each language and history of reinforcement within each lan- guage. For example, for those with corresponding results, participant fluency and history of reinforcement may have been equivalent in both languages. Unfortunately, that information was not available for the study participants. Therefore, it is equally possible that, for the participants which demonstrated correspondence across lan- guages, the participants may have attended to alternative discriminative stimuli (e.g., color of shirts) and the language of implementation may not have influenced the assessment. These limitations might be considered for future research.

When evaluating the behavior within each assessment, various trends were noted in terms of the different rates of problem behaviors observed. For example, six of the nine participants engaged in higher rates of behavior in the Spanish FA as com- pared to the English FA (e.g., Jackie, Jeremy, Gabe, Elise, Alfonzo, & Spencer). The latency to behavior was also faster in the Spanish than the English FA for seven of the ten FAs. One possible explanation could be that the participants had a longer reinforcement history in the Spanish versus the English language. A second possible explanation is that the participant learned the contingency and responded accord- ingly in the second FA (i.e., sequence effects). This would be supported by Alex’s assessment in which the English FA was conducted after the Spanish FA. To note, although Slater’s FA results indicate lower rate of behavior during the Spanish as compared to the English FA, Slater engaged in his topography of behavior (bit- ing) at a lower intensity but longer duration as the assessments progressed (i.e., he held the bite on the experimenters clothing). This may also indicate he learned the contingency.

The results regarding higher rates of behavior and shorter latencies in the Span- ish versus English language did not hold for all participants. For example, Carlos’ FA results present an interesting case in that the behaviors occurred at a lower rate yet quicker latency in the Spanish condition (second FA) than the English condition. Anecdotally, Carlos did have a more advanced vocal–verbal repertoire as compared to the other participants. As Spanish was his primary language, these results could indicate a longer learning history in the Spanish language or a preference for the Spanish language (Aguilar et al. 2017). Alfonzo’s FA results also present an interest- ing case as his behavior during the escape condition in the Spanish language was variable and higher than the English FA. Perhaps Alfonzo had a higher level of flu- ency with the English language, and the demands were not as difficult or aversive. Similarly, Carlos had lower rates of behavior in the Spanish escape condition, per- haps signaling he was more fluent with the Spanish language and demands were not as difficult or aversive in that language. These subtle differences highlight the

Journal of Behavioral Education

1 3

need for careful analysis in behavioral variations across conditions. Given the dif- ferent trends observed across the participants in the present study, future research is needed to evaluate larger Ns of participants to determine what participant character- istics and reinforcement histories might account for these differences.

Although these findings provide an extension to the literature by evaluating lan- guage diversity, they are not without limitations. First, the participant fluency with each language and history of reinforcement within each language was not assessed prior to conducting the FAs. As these two factors could explain the results, future researchers might explicitly attempt to evaluate these factors prior to conducting the FAs. Second, the generalizability of these findings as they relate to other languages highlights the need for further extensions to this literature. In the current study, we had a fairly homogenous group of learners who were learning Spanish and English language. Thus, research evaluating other languages (e.g., Arabic, Chinese, Hindi, etc.) of dual-language learners would be beneficial a greater understanding of the impacts of language diversity. Further, given the children included in the current study were not fully bilingual, research is needed to evaluate the differences that might be observed in learners who are more proficient in their language repertoires. Third, we chose to implement an A-B design, which does not demonstrate experi- mental control. In addition, the resulting randomization led to only one participant randomized to the Spanish language first. These design constraints limit the strength of the conclusions that can be drawn from this study. Future research might fur- ther this line of research and consider using a reversal design to further evaluate the impact of language on assessment results. However, a cost–benefit analysis might be considered to evaluate potential impacts on participants. Finally, there is a lim- ited research evaluating the effect of language on the treatment of problem behav- ior (Neely et al. 2019). Given the preliminary nature of this study, future research is needed that includes the effects of the treatment plans developed to help fur- ther evaluate the impacts of language on the assessment and treatment of problem behavior.

For practitioners, there are a few considerations that may impact the efficacy of our practice as it relates to language. This study highlights the need to evaluate vari- ations in rates of behavior identified during an FA and the considerations that should follow when developing treatment plans. For example, children who show increased rates of problem behavior in one language condition may also show a preference of language. Thus, it may be important for practitioners to accounting for preference of language and the behavioral impacts (i.e., increased rates or intensity) to yield better treatment efficacy (Aguilar et al. 2017).

As a field, we highlight the need for addressing socially valid behaviors (Baer et al. 1968; Brodhead et al. 2014). In recent years, several researchers have high- lighted the need for inclusion of inclusive practices (i.e., Brodhead et al. 2014; Fong et al. 2017), the need for the inclusion of diversity in research (i.e., Sinclair et al. 2018), and more specifically, understanding linguistic diversity (Lim et  al. 2018). It is likely that the need for social validity should extended to utilizing socially sig- nificant assessments when working with individuals who are dual language learners (Lim et al. 2018). In doing so, behavior analysts will ensure adherence to our ethics code (BACB Ethical Compliance Code for Behavior Analysts 2015) by including

1 3

Journal of Behavioral Education

culturally responsive practices, which may also help to ensure reliable assessment findings. Becoming more aware of the implications of language diversity may help practitioners address more socially valid behaviors within the treatment plans they develop. In addition, future research might extend this research by evaluating other aspects of culture (e.g., cultural values) and the impact on assessment results.

This extension provides a methodological framework for linguistic analysis dur- ing an FA, the evaluation of the impacts on problem behavior, and highlights the importance of the inclusion of linguistic diversity in the assessment of problem behavior. We recommend that the inclusion of linguistic diversity in the analysis of problem behavior becomes standard practice for the assessments and treatments of problem behavior.

Funding This article was not funded.

Compliance with Ethical Standards

Conflict of interest The authors declare that they have no conflict of interest.

Ethical Approval This article complies with all ethical requirements and was approved by the University of Texas at San Antonio Institutional Review Board #16-247.

References

Aguilar, J. M., Chan, J. M., White, P. J., & Fragale, C. (2017). Assessment of the language preferences of five children with autism from Spanish-speaking homes. Journal of Behavioral Education, 26, 334–347. https ://doi.org/10.1007/s1086 4-017-9280-9

Baer, D., Wolf, M., & Risley, T. (1968). Some current dimensions of applied behavior analysis. Journal of Applied Behavior Analysis, 1, 91–97. https ://doi.org/10.1901/jaba.1968.1-91

Brodhead, M. T., Durán, L., & Bloom, S. E. (2014). Cultural and linguistic diversity in recent verbal behavior research on individuals with disabilities: A review and implications for research and prac- tice. Analysis of Verbal Behavior, 30, 75–86. https ://doi.org/10.1007/s4061 6-014-0009-8

Conners, J., Iwata, B. A., Kahng, S. W., Hanley, G. P., Worsdell, A. S., & Thompson, R. H. (2000). Differential responding in the presence and absence of discriminative stimuli during multielement functional analyses. Journal of Applied Behavior Analysis, 33, 299–308. https ://doi.org/10.1901/ jaba.2000.33-299

Fong, E. H., Ficklin, S., & Lee, H. Y. (2017). Increasing cultural understanding and diversity in applied behavior analysis. Behavior Analysis: Research and Practice, 17, 103–113. https ://doi.org/10.1037/ bar00 00076

Hanley, G. P., Iwata, B. A., & McCord, B. E. (2003). Functional analysis of problem behavior: A review. Journal of Applied Behavior Analysis, 36, 147–185. https ://doi.org/10.1901/jaba.2003.36-147

Iwata, B., & Dozier, C. L. (2008). Clinical application of functional analysis methodology. Behavior Analysis in Practice, 1, 3–9. https ://doi.org/10.1007/BF033 91714

Lang, R., Rispoli, M., Sigafoos, J., et al. (2011). Effects of language of instruction on response accuracy and challenging behavior in a child with autism. Journal of Behavioral Education, 20, 252–259. https ://doi.org/10.1007/s1086 4-011-9130-0

Lim, N., O’Reilly, M. F., Sigafoos, J., & Lancioni, G. E. (2018). Understanding the linguistic needs of diverse individuals with autism spectrum disorder: Some comments on the research literature and suggestions for clinicians. Journal of Autism and Developmental Disorders, 48, 2890–2895. https :// doi.org/10.1007/s1080 3-018-3532-y

Journal of Behavioral Education

1 3

Neely, L., Graber, J., Kunnavatana, S., & Cantrell, K. (2019). Impact of language on behavior assess- ment and intervention outcomes. Journal of Applied Behavior Analysis, Online First. https ://doi. org/10.1002/jaba.626

Ninci, J., Neely, L., Hong, E., Boles, M., Gilliland, W., Ganz, J., & Vannest, K. J. (2015). Meta-analy- sis of interventions to improve functional living skills for people with autism spectrum disorder. Review Journal of Autism and Developmental Disorders, 2, 184–198. https ://doi.org/10.1007/s4048 9-014-0046-1

Parker, R. I., Vannest, K. J., & Davis, J. L. (2011). Effect size in single-case research: A review of nine nonoverlap techniques. Behavior Modification, 35, 303–322. https ://doi.org/10.1177/01454 45511 39914 7

Rispoli, M., O’Reilly, M., Lang, R., Sigafoos, J., Mulloy, A., Auilar, J., & Singer, G. (2011). Effects of language of implementation on functional analysis outcomes. Journal of Behavioral Education, 20, 224–232. https ://doi.org/10.1007/s1086 4-011-9128-7

Schlichenmeyer, K. J., Roscoe, E. M., Rooker, G. W., Wheeler, E. E., & Dube, W. V. (2013). Idiosyn- cratic variables that affect functional analysis outcomes: A review (2001–2010). Journal of Applied Behavior Analysis, 46, 339–348. https ://doi.org/10.1002/jaba.12

Sinclair, J., Hansen, S., Machalicek, W., Knowles, C., Hirano, K. A., & Murray, C. (2018). A 16-year review of participant diversity in intervention research across a selection of 12 special education journals. Exceptional Children, 84, 312–329. https ://doi.org/10.1177/00144 02918 75698 9

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.