THE DEVELOPMENT OF THE SELF-EFFICACY DURING EXPOSURE â•fi CHILD (SEE-C) MEASURE: A PILOT STUDY

Cognitive-behavioral therapy (CBT) with exposure is an effective treatment for anxiety disorders but involves acute discomfort, rendering treatment engagement a challenge. Willingness to engage in exposure is of interest, because a child has to willingly approach a feared stimulus. This preliminary study investigated child engagement during exposures in CBT for anxiety disorders, with the development, validation, and field testing of a measure of self-efficacy over three time points. The measure, Self-Efficacy During Exposure-Child (SEE-C), is a 9-item, self-report measure of self-efficacy during exposure for youth 8-17 years old. A sample of eight reviewers at least 2-years post licensure (M = 14.06; SD = 4.71) with expertise in CBT for childhood anxiety provided feedback on the SEE-C’s face and content validity. Field testing included a sample of 24 child-parent dyads recruited from an anxiety clinic in the Northeast U.S. Child-parent dyads were asked to complete measures of child selfefficacy; child anxiety symptoms; and child school, social, and family functioning. Children were also asked to complete a measure of motivation. Analyses revealed the SEE-C to demonstrate significant increases in child self-efficacy, reductions in child anxiety symptoms, and increases in social and family functioning over treatment. Internal consistency of the SEE-C was acceptable to excellent, and exploratory principle component analysis suggested a three-factor solution, with loadings ranging from 0.5 to 0.9. The SEE-C adds to the literature as the first measure of child self-efficacy designed for use during exposure in CBT for anxiety. Findings provide insight into those factors that contribute to a child’s engagement during exposure. Reported effect sizes are promising and warrant greater investigation of the SEE-C’s utility.


LIST OF TABLES
. Cognitive-behavioral therapy (CBT) for anxiety disorders among children and adolescents, particularly those incorporating exposure practice, have been empirically supported and denoted efficacious by a considerable amount of clinical outcome research (e.g., Compton, Peris, Almirall, Birmaher, Sherrill, Kendall, et al., 2014;Higa-McMillan, Francis, Rith-Najarian, & Chorpita, 2015;Kendall, Flannery-Schroeder, Panichelli-Mindel, Southam-Gerow, Henin, & Warman, 1997). CBT with exposure is a variant of CBT that is purported to work via exposure, where exposure is defined as "a controlled therapeutic task in which a person confronts an anxietyprovoking stimulus or situation" (Marks, 1973). As such, anxiety and fear reduction occurs through contact with the feared stimulus.
While CBT with exposure is identified as an effective treatment for childhood anxiety disorders, it involves much acute discomfort, thus rendering treatment engagement during exposure with children and adolescents a challenge. A focus on exposure practice during treatment is necessary, however, as it is considered a primary mechanism and an active "ingredient" in anxiety reduction (Peris et al., 2015). As a result, questions about the exposure process have arisen. For example, how does a child in CBT for anxiety begin to engage in and utilize exposure to "fight" their anxiety? Also, when does the child begin to experience a reduction in anxiety during exposure? Further, does a reduction of a child's anxiety during exposure within session lead to future reductions of anxiety during exposure across session within the CBT with exposure treatment trajectory?
Social psychology, sports psychology, and behavioral economics each study the performance-enhancing concept of positive psychological momentum (PPM).
Rooted in attribution theory (the study of the processes by which individuals explain the causes of behavior and events), PPM is defined as the "perception, attitude, belief, or state-of-mind an individual experiences, in which their initial success leads to more successes" (Iso-Ahola & Blanchard, 1986). Positive psychological momentum theory posits that perceived positive momentum leads to increased confidence; thus, in turn, leading to more active and better performance (Rosenqvist & Nordström Skans, 2015).
In line with this phenomenon, Compte and Postlewaite (2004) suggest that a causal link may exist from past successes to future performance through "confidence," where confidence is defined as one's belief in one's ability to succeed in specific situations or accomplish a task (Iso-Ahola & Dotson, 2014). However, given the definition of confidence, it seems that before an individual can perceive confidence, they have to first perceive self-efficacy. Self-efficacy is the belief one has about their ability to perform a certain task (Bandura, 1997(Bandura, & 1988; thus, one will have confidence when self-efficacy is utilized to competently complete a task multiple times. The concept of PPM may also be used to explain enhanced performance within the context of enhancing a child/adolescent's motivation (the process that initiates, guides, and maintains goal-oriented behaviors). The theory of PPM, if shown to have a causal relationship with self-efficacy and motivation by which "success leads to success," could be used to enhance performance not only within sports or entrepreneurial endeavors, but also within the context of enhancing youth motivation and self-efficacy to complete exposure tasks in CBT treatments.

Child Engagement During Exposure in CBT for Childhood Anxiety Disorders
A child's willingness to engage in and adhere to exposure practice during CBT is of great interest, because exposures cannot be "done" to a child. Rather, a child has to willingly participate in a task that focuses on a feared stimulus in order for the fear to reduce. Engagement (the act of occupying the attention or efforts of a person, and in this case, a child) in exposure is made up of multiple factors. However, to date, there is a significant gap in the literature, as no investigations have been found related to child factors that contribute to engagement during the utilization of exposure. Thus, the study of these factors is needed.
While no studies have examined child engagement during exposure sessions, some research has investigated child engagement more generally during CBT treatment. For example, Morgan and colleagues (2013) explored the relationship between poor treatment adherence and attenuated treatment response in pediatric OCD and found that a child's willingness to engage in exposure mediated overall treatment outcomes. Morgan and colleagues' (2013) study supports the need for additional research regarding treatment engagement in exposure, yet, does not examine the childspecific factors that may be identified as necessary for engagement during exposure.
Relatedly, King, Currie, and Petersen (2014) examined factors of child engagement in mental health treatment and assert that it involves a motivational commitment to the intervention process. They suggested that this process includes behavioral involvement such as child in-session participation, therapist and child collaboration, and the child's self-efficacy to continue in the identified intervention.
Thus, a child's in-session motivation to experience and feel distress in the face of a feared stimulus along with their perceived self-efficacy may be two important factors in engagement during exposure.
While there are many factors that make up child and adolescent treatment engagement during exposure in CBT for childhood anxiety, including caregiver (hereinafter referred to as "parent") factors and clinician factors, motivation and perceived self-efficacy are two child factors purported to contribute to engagement within the theory of positive psychological momentum. Yet, there is a real need for the development of assessment tools to measure these constructs. The present study will focus on the latter (i.e., self-efficacy), as a review of the literature demonstrates the absence of such a measure.

Self-efficacy: A Factor to be Explored
Perceived self-efficacy has been studied within the context of fear reduction and phobias (Bandura, 1977(Bandura, , 1978(Bandura, , 1982(Bandura, , 1998 and is posited to improve one's expectation of achievement. Though perceived self-efficacy does not cause the reduction of fear (Tryon, 2005), it may function as an anchor for a child to engage in exposure practice and a platform for a child to engage in additional and more difficult exposure practice between sessions. This is notable, as Bandura (1997) maintained that perceived self-efficacy influences one's motivation to act or to persevere in the face of difficulties.
There are multiple empirical investigations on child and adolescent perceived self-efficacy reported in the literature. For instance, Tonge, King, Klimkeit, Melvin, Heyne, and Gordon (2005) developed and tested a measure of perceived self-efficacy about coping with depressive symptoms in adolescents and found the measure's psychometric properties to be acceptable. Study results indicated higher pre-treatment self-efficacy scores predicted better outcomes at three and six months post-treatment.
Similarly, Bandura and colleagues (1999) determined that perceived self-efficacy influenced childhood depression and contributed to concurrent and subsequent depression. Additionally, Warren and Salazar (2015) observed self-efficacy to be associated with improvements in youth-reported symptoms of emotional distress and behavioral dysfunction in routine mental health services. Self-efficacy has also been found to be predictive of performance on cognitive tasks, such as academic achievement and social adaptation (Magno & Lajom, 2008;Pastorelli, Caprara, Barbaranelli, Rola, Rozsa, & Bandura, 2001). Multiple child and adolescent health studies have also revealed high perceived self-efficacy to be predictive of proper management of chronic conditions, such as weight loss behaviors (Walpole, Dettmer, Morrongiello, McCrindle, & Hamilton, 2011), smoking behaviors (Ford, Oladopo, Sterling, Diamond, Kelder, & McAlister, 2013), chronic pain (Bursch, Tsao, Meldrum, & Zeltzer, 2006), chronic illness (Emerson et al., 2018) and exercise-related behaviors (Pakarinen, Parisod, Smed, & Salantera, 2017). Considering the extant literature on child perceived self-efficacy and child engagement during treatment, selfefficacy is an important factor to be explored in a study of engagement during exposure.

The Current Study
The current study reports the development, psychometric evaluation, and field testing of a measure of child/adolescent perceived self-efficacy to be used during exposure in CBT for children and adolescents with anxiety disorders. The purpose of the measure is to facilitate the understanding of child/adolescent factors that contribute to treatment engagement during exposure.
The study was conducted in two parts: 1. The development and psychometric evaluation of the measure, including expert review and child response feedback, and 2.
Field testing of the measure to examine the relationship of child perceived selfefficacy to other variables during CBT for anxiety among children and adolescents.
See Study Flow Chart in Figure 1.    Table 1 for Iterative Process).

Aggregation and Consensus
Quantitatively: • Items will be summarized descriptively using frequencies, means and standard deviations Qualitatively: • Item statements will be aggregated as to theme and consensus (i.e., frequency) will be determined • Frequency of consensus items will be analyzed as to outline considerations

Outlined considerations
Quantitatively: • 100% participant agreement on a single item with a Mean = 7 or disagreement with a Mean = 1 will be an item for editing • Participant agreement on a single item with a Mean = 5 or 6 or disagreement with a Mean = 2 or 3 will be an item considered for editing but additional reference to the literature and advisor consultation will determine and justify scale adjustments • Participant agreement on a single item with a Mean = 4 will not be considered for editing Qualitatively: • Consensus statements on a single item endorsed by each reviewer (N = 8) will be considered an item for editing • Consensus statements on a single item endorsed by one to seven reviewers (N = 1 to 7) will be considered for editing but additional reference to the literature and advisor consultation will determine and justify scale adjustments Field testing. A sample of 24 child-parent dyads was recruited from the Child Anxiety Program (CAP) at the University of Rhode Island (URI). Children ranged in age from 8 to 15 years (M = 9.9, SD = 1.93), with almost half the sample identifying as female (46%) and identified as White, non-Hispanic. Children were in the 2 nd to the 10 th grade, with more than half of the sample in the 3 rd through the 6 th grades. Additionally, 29% reported having a 504 plan in school. Parent participants identified as biological parents, a mean age of 42.1 (SD = 5.67), predominantly female (83%), and identified as White, non-Hispanic. The majority of children had biological parents living together (92%) and an average annual income ranging from $100,000-$120,000. More than half of parents reported a college degree or higher (92%). See Table 2 for an outline of demographic information.
Retention rates. Twenty-four parent-child dyads consented to the field testing portion of the study. Out of the 24 dyads, one child decided they no longer wanted to complete study measures during the 1st session; however, the parent continued in the study. At mid-and post-treatment, study retention rates remained the same, with 23 child and 24 parent participants. All child participants met criteria for at least one anxiety disorder. Eighty-six percent of the children met criteria for more than one disorder (anxiety or another type), ranging from two to five (M = 2.58, SD = 1.06) comorbid disorders. Primary anxiety diagnoses at pre-treatment included Generalized Anxiety Disorder (34%), Specific Phobia (23%), Separation Anxiety Disorder (14%), and Social Anxiety Disorder (14%).
See Table 3 for a complete listing of pre-treatment diagnoses and comorbid conditions counts. The first two participants of the 24 child-parent dyads enrolled were queried after their completion of the SEE-C to glean information for response process validity. Two child participants (M = 12 years; Male = 2) were queried about their experience and understanding completing the SEE-C including the scale's directions, questions, and response items. Both child participants identified as White, non-Hispanic. Response process feedback was aggregated and findings integrated.
Inclusion Criteria for the study were the following: 1) primary diagnosis of an anxiety disorder using the Anxiety Disorder Interview Schedule -IV (ADIS-IV; Silverman & Albano, 1996), adapted for DSM 5,2) Child is between the ages of 8 and 17 years old and has a parent or legal guardian available to participate in treatment, and 3) Child participant and parent are English speaking. Exclusion Criteria included a documented child diagnosis of Psychosis, Autism, or Intellectual Disability and child use of anti-depressant and/or anti-anxiety medications that has not been stable for more than six weeks. Exclusion criteria were designed to be minimal and exclude only those patients for whom CBT is not likely to be beneficial or may be risky.

Measures
Expert reviewer questionnaires were completed and edits to the developed measure were made in prep for field testing. Field testing questionnaires were completed at pre-treatment, mid-treatment (after completing 8 weeks of CBT and prior to initiation of exposure practice), and post-treatment (session 16). All field testing questionnaires were completed during scheduled intake or therapy sessions. See Appendix A for table of administration time points.

Self-Efficacy During Exposure -Child (SEE-C).
The SEE-C is a 9-item, child (ages 8 to 17 years old) self-report measure of perceived state self-efficacy during exposure using a 5-point Likert scale from 1 ("Not sure at all") to 5 ("Completely sure").
Total scores range from 0 to 45, and higher numbers on total score indicate greater perceived self-efficacy. Three subscales were defined in the construction of this measure: 'success of handling distress during an exposure' (items 1, 2, and 3), 'success of individual exposures themselves' (items 4, 5, and 6), and 'success of exposure treatment' (items 7, 8, and 9). Items were selected following Bandura's (2006)  An identical parent version of the measure was developed, where parents reported their perception of their child's state self-efficacy during exposure. The parent version was created to corroborate the child version, as a method to assist with the validity of the child measure.

Expert Reviewer Questionnaires
Clinician Demographic Questionnaire (CDQ). The CDQ was completed by expert clinicians (e.g., gender, education, institution). This information was used in preliminary descriptive analyses to describe the expert clinician sample. See Appendix D.

Clinician Feedback Survey (CFS). The CFS was adapted from the Pediatric
Motivation Scale Service Provider Survey (Tatla, 2014) and used to obtain feedback from expert clinicians upon review of the SEE-C. Questions on this survey include Likert scale and open-ended questions related to the measure's face validity, clarity (i.e., conciseness, grammar, readability, layout, reading level, and redundancy of questions) and clinical utility (i.e., ease of administration, time to administer, and challenges in use). To corroborate findings, information was reviewed qualitatively and quantitatively and then aggregated to revise and refine the SEE-C. See Appendix D.
Semi-structured Administration Questionnaire (SSAQ). The SSAQ was adapted from the Administration Questionnaire (Tatla, 2014). This questionnaire was used to query child/adolescent participants about their experience completing the SEE-C.
Child/adolescent participants were asked about their understanding of the scale's directions, questions, and response items. They were asked to provide any comments they have in regards to each. To corroborate findings, queried information was summarized and aggregated to revise and refine the SEE-C. See Appendix D.

Field Testing Questionnaires
Caregiver Demographics. This demographics questionnaire was completed by parents at baseline and includes questions regarding family composition, parent information (e.g., age, gender, race, ethnicity, education, income, occupation), and child information (e.g., age, gender, race, ethnicity, education). See Appendix D. In tandem with the CAP clinic procedures, a short phone screen by the PCC coordinator was used to identify potentially eligible participants. The research study was introduced to families at this time and interested parents were provided a short description of the study and an opportunity to ask questions. Consistent with the CAP clinic procedures, eligible families were invited to the clinic for a two-hour visit during which an intake assessment occurred.

Anxiety Disorder Interview Schedule for Children for DSM-IV (ADIS-IV),
Informed parent consent and child assent for the present study was conducted during the intake following the usual treatment consenting procedures for CAP. Families were reminded that study participation was voluntary and could be discontinued at any point during their treatment and the termination of their participation would not affect their treatment status.
Following informed consent/assent, pre-treatment assessments were administered with the parent and child. Families not interested in participating in the study were not penalized and were free to continue with treatment in CAP, provided they met with entry requirements specific to the CAP program. Families who agreed to participate in the present study were provided a complementary treatment manual (value ~$25) as compensation for their participation. The consent, assessment, and treatment procedures were video recorded (as consistent with CAP procedures) to ensure accuracy of procedures, and videotapes were destroyed in alignment with APA, Health Insurance Portability and Accountability Act (HIPAA), and PCC requirements.
Child participant reviewers. The first two participants of the 24 child-parent dyads recruited were queried about their experience completing the SEE-C using the SSAQ. Participants' response processes was observed and recorded while they complete the SEE-C pre-session. Research staff queried on items from the SEE-C that seemed difficult to answer or appeared confusing to participants. Participants were also asked to explain the rationale for their response selections to further evaluate participants' understanding of the SEE-C items. In order to minimize the effects of social desirability, a study research assistant administered the scale to the child, and the treating therapist was not present. Information gleaned on the SSAQ was reviewed quantitatively and qualitatively and then used to revise and refine the SEE-C.

Treatment. The treatment provided was CBT with exposure per the 16-session
Coping Cat treatment protocol (Kendall & Hedtke, 2006) and involved coping skill instruction and practice to reduce anxiety. Length of treatment sessions was 50 minutes.
At each session, child and parent participants completed paper and pencil measures about child perceived self-efficacy and motivation. At mid-and post-treatment, participants were again asked to complete the study measures.

Data Analyses
Preliminary statistical analyses. All statistical analyses were conducted using IBM SPSS Statistics version 24 (IBM Corp., 2016). Data were cleaned (frequencies, means, standard deviations and ranges were examined) and scored, and tests of assumption and normality for skewness and kurtosis were completed. Overall, study measures were deemed normally distributed (See Table 4). Descriptive statistics were used to summarize the sample characteristics. Continuous variables were summarized using means and standard deviations. Categorical variables were described with frequencies and percentages. Pre-treatment report of child prescribed anxiety medication (n = 2; Zoloft and Lorazepam) versus no medication were examined and showed no significant differences on the SEE-C child version or anxiety symptom measures; as a result, medication usage was not controlled for in the overall study analyses. Consistent with an intent-to-treat approach, all participants were included in the data analyses.
Missing data. Utilizing maximum likelihood, expectation-maximization algorithm (Allison, 2012), missing outcome data on randomly assigned participants were

replaced. A nonsignificant Little's Missing Completely At Random (MCAR) test
suggests that the data were missing completely at random (Little, 1988). Maximum likelihood imputation, using the expectation-maximization algorithm, was used to impute the missing data (less than 5%) to improve statistical power with unbiased parameter estimates (Enders, 2001;Scheffer, 2002). Expert reviewer and child participant feedback was integrated per outlined considerations (including both quantitative and qualitative methods, which are outlined in Table 1) and literature review. Item responses were summarized descriptively using frequencies, means, and standard deviations and evaluated qualitatively through an examination of themes and their frequency.

Expert Review
Tailoring the SEE-C per expert review. A sample of eight expert reviewers were recruited to tailor the SEE-C to the age and population of the proposed sample with emphasis on content and face validity. The CFS was used to obtain feedback on the measure's face validity, clarity, and potential clinical utility.
Responses on the CFS were reviewed, summarized, and aggregated for incremental integration of the best representation of the construct of self-efficacy, ease and clarity of reading the instructions for the scale, ease and clarity of reading the scale, layout attractiveness, appropriateness of reading level appropriate for those as young as 8 years old, and ease of completion by children/adolescents. See Table 5 for summary of CFS responses.
The CFS revealed that >50% of the expert reviewers thought the results of the SEE-C would inform their intervention planning and treatment engagement during exposure, the SEE-C items represented self-efficacy, and that youth with anxiety could understand responses. More specifically, 50% "agreed" or "strongly agreed" that items represented self-efficacy; 75% "agree" to "strongly agree" instructions were easy to follow; 62% "agree" to "strongly agree" youth with anxiety could understand responses; 87.5% thought the layout of questions was attractive and appealing; 62.5% "somewhat agree" and 25% "agree" reading level of SEE-C is appropriate for an 8-year old child; 100% "agree" to "strongly agree" a child with anxiety would not object to answering any items on the SEE-C; 50% endorsed "<5 minutes" and the other 50% "5-10 minutes" regarding how long they thought it would take to complete SEE-C; 100% thought "5-10 questions" was a reasonable number of questions to include on the SEE-C. Reviewers indicated the measure results would inform intervention planning and treatment engagement during exposure with 62.5% responding "yes," 25% "possibly," and 12.5% left the question blank. Sixty-two percent believed this scale would be helpful to them as a therapist. Thirty-seven percent believed that it might be useful. See Figure 2 for a summary of expert reviewer feedback.

Table 5. Expert Reviewer Feedback
Question prompt: Q.) "In your opinion, would this scale be helpful to you as a therapist?" Responses: "Yes" = 5 (62.5%); "Maybe" = 3 (37.5%); "No" = 0 (0%) If you answered "Yes" to Q., explain how this scale could be helpful to you: Predict Better Treatment Outcomes: "Good to prompt and guide clinical discussion about areas where there is less selfefficacy and may be able to predict outcome" "Towards end of treatment, it would give therapist another source of ERP efficacy & possibly predict treatment outcomes and maintenance of treatment gains" Fill an Important Niche: "I agree with you, the Child Self-Efficacy is not adequately assessed in CBT and this would fill an important niche. Especially, if it is given as a repeated measure across treatment to assess treatment-related changes in children's confidence. Nice job."

Self-efficacy may be a Mechanism of Anxiety Reduction:
"If change in self-efficacy cognitions are a mechanism of anxiety reduction, this measure may help test this hypothesis" Assist with Engagement During Exposure: "It might be helpful by providing a structured way to assess for the nuances associated with the child's thoughts/beliefs about exposure tasks" "It would help at beginning of treatment to assess a patient's expectations about ERP and allow the therapist to provide psychoeducation to enhance acceptability of treatment" "Specifying child's belief in his/her efficacy can then become a target for intervention (e.g., self-talk)" "Yes, because you can assess expectations (in items 7-9) & attributions (in items 4-5)" If you answered "Maybe," explain how this scale may or may not be helpful to you: "I am somewhat concerned that many of the items have to do more with outcome expectancy and other constructs that are related to self-efficacy but are not really selfefficacy" "I just wonder if children can rate their response to exposures in general. They may say they can do some lower-level anxiety exposures but not believe they can do higher-level exposure"

Figure 2. Expert Reviewer Feedback
Thought the SEE-C Child would be helpful to them as a therapist: 62.5% = "yes" and 37.5% = "maybe" If "Yes," explain how this scale could be helpful to you: • Predict better treatment outcomes • Fill an important niche • Self-efficacy may be a mechanism of anxiety reduction • Assist with engagement during exposure

Thought the SEE-C Child would inform intervention planning and treatment engagement during exposure:
Thought the SEE-C Child items represented self-efficacy: 12.5% = "strongly agree," 37.5% ="agree," and 25% = "somewhat agree" SEE-C response processing per child feedback. The SSAQ was used with two child participants to elicit qualitative and quantitative feedback on the comprehension of scale items and the understanding of response format. This information was used to revise and refine the SEE-C. Both child participants described the developing measure as needing more information regarding the definition and examples of exposure. Then, once the SEE-C was revised and the additional information included, child participants reported that the questionnaire was easy to understand and both required minimal assistance to complete the questionnaire. They also indicated that they would willingly complete the SEE-C, the questions were easy to answer, and they liked the formatting of the responses. The SEE-C was edited via a formal iterative process with regard to agelevel readability, conciseness, and the need to add a definition and example of exposures.

Field Testing
Sample characteristics. A clinical sample of 24 treatment-seeking parent-child dyads was enrolled into the study. One-way ANOVA's were conducted on the number of pre-treatment comorbid diagnoses (ranging from 1 to 5 diagnoses) by pre-, mid-, and post-treatment on the child's self-report of self-efficacy on the SEE-C. Due to the low frequency of four (n = 3) and five (n = 1) diagnoses, these two values were collapsed to one value (n = 4) for these analyses. Results indicated no difference in self-efficacy at pre-and mid-treatment on the SEE-C child version; however, a significant difference on child self-efficacy was seen at post-treatment by the number of pre-treatment comorbid diagnoses (F(3,19) = 3.260, p=.044), where post hoc test revealed a difference approaching significance (p = .056) between a comorbidity of two pre-treatment diagnoses (n = 8) and four/five diagnoses (n = 4).See Table 6 for post-treatment diagnoses frequencies and treatment counts.

Examination of the SEE-C at Pre-treatment
Bivariate Pearson Correlation Coefficients were run between the pre-treatment measures including the SEE-C, SCAS, CSDS composite scores and demographic variables (child age and gender). Results showed that the SEE-C was significantly related to parent report of child anxiety symptoms as measured by the SCAS (r = -.417, p = .048). The parent and child report of anxiety symptoms were also significant and positively correlated to each other (r = .715, p = .000). In addition, significant relationships were found between the child report of anxiety symptoms and the child report of functional impairment within the family (r = .502, p = .017), the parent report of child functional impairment socially (r = .625, p = .001), and the parent report of child functional impairment within the family (r = .688, p = .000). The child report of anxiety symptoms was also significantly related to the clinician report of symptom severity on the CGI (r =. 477, p = .018). See Table 6 for complete list of pre-treatment correlations.
Given the significant pre-treatment relationship between the SEE-C child version and the SCAS parent version, as well as other pre-treatment measures, regression analyses were conducted to examine whether the child symptoms of anxiety, functional impairment, and motivation predicted to child self-efficacy. The first regression model examined whether parent report and child self-report of pre-treatment anxiety symptoms predicted child self-report of self-efficacy. Results approached significance, R 2 = .232, F(2, 22) = 3.019, p = .071, where only the parent report of child anxiety symptoms was significantly predictive (β = -.646, t = -2.389, p = .027). Three additional regression analyses examining (1) child self-report of state motivation, (2) child self-report and parent report of child functioning, and (3) clinician report of child symptom severity and improvement as well as child age and gender as predictors of child self-report of selfefficacy were not significant.
Internal A three-factor solution, with loadings ranging from 0.5 to 0.9, emerged from the analysis.
Item one did not clearly load on a single factor but overlapped on two (Factor 1, 'belief in success of handling distress during an exposure,' and Factor 3, 'belief in success of exposure treatment'). Items two, three, and four loaded on Factor 1 ('belief in success of handling distress during an exposure'); items five, six, and nine loaded on Factor 2 ('belief in success of individual exposures themselves'); and items seven and eight loaded on Factor 3 ('belief in success of exposure treatment'). Cronbach alphas for each factor showed promising results with acceptable to good internal consistency, as Factor 1 had an α =. 748, Factor 2 had anα = 708, and Factor 3 had an α = .831 (item one was eliminated from these analyses due to the overlap between two factors). While three meaningful constructs are suggested, no clear determination of factor loadings can be made, at this point, due to the small sample size.

Examination of the SEE-C at Mid-and Post-Treatment
An exploratory examination at both mid-and post-treatment of the relationship between parent and child report of child anxiety symptoms, functioning, and clinician report of symptom severity and improvement were conducted (See Tables 7 & 8).
Specifically, at mid-treatment, clinician report of child symptom severity was negatively correlated to child self-report of self-efficacy (r = -.603, p = .003). At post-treatment, child self-report of self-efficacy was inversely related to child self-report of anxiety symptoms (r = -.611, p = .004), parent report of child anxiety symptoms (r = -.545, p = .016), parent report of child functioning within the family (r = -.605, p = .006), and positively correlated to both parent report of child self-efficacy (r = .484, p = .023) and clinician report of improved symptoms (r = .527, p = .012).
Due to significant relationships at mid-treatment and at post-treatment between the SEE-C and other constructs, hierarchical multiple regressions were performed to assess whether certain variables predicted to lower child anxiety symptoms at midtreatment and at post-treatment. Significant results were seen when modeling posttreatment child self-report of self-efficacy on the SEE-C and PMOT (motivation) and predicting to post-treatment child self-report of anxiety symptoms via the SCAS after controlling for pre-treatment child self-report of anxiety symptoms, self-efficacy, and motivation. Results demonstrated that there was a significant effect (F(5, 13) = 4.348, p = .015, R 2 = .626), where, after controlling for pre-treatment variables, at posttreatment higher child-report of self-efficacy predicted to lower anxiety symptoms via child-report (β = -.574, p = .013). Individual predictors were examined and are reported in Table 9.

Evaluation of the SEE-C Overtime
Repeated measures MANOVAs were conducted on all measures and groups of measures with the same construct (i.e., child self-efficacy, child anxiety symptoms, child motivation, and child functioning) across pre-, mid-, and post-treatment.
Significant differences were seen overtime on most constructs. Particularly, findings showed that the SEE-C child and parent versions demonstrated significant, large effects from pre-, to mid-, to post-treatment (F(2, 18) = 7.976, p = .000, ηp 2 = .301).
Bonferroni post hoc tests showed that participants experienced a significant change on the child version, p = .044, from pre-treatment to mid-treatment and on the parent version with significant increases from pre-to post-treatment (p = .017), mid-to posttreatment (p = .010), and pre-to post-treatment (p = .000). See Similarly, with a medium effect size, parent and child report of child impairment (See between pre-to post-treatment (p = .000) and mid-to post-treatment (p = .001), and on child improvement (increased in improvement) between pre-to post-treatment (p = .003) and mid-to post-treatment (p = .000). See acceptable to excellent internal consistency/reliability from pre-treatment to posttreatment on the measure as a whole, and a promising three-factor structure that demonstrated acceptable to good internal consistency per factor.
Much of the literature on child anxiety treatment focuses on treatment outcomes of CBT with exposure but not on the effects of child engagement during exposure. Child factors that may influence this engagement, as postulated in the theory of positive psychological momentum, include perceived self-efficacy and motivation. As study findings suggest, self-efficacy and motivation are relevant and important to child engagement during exposure, as they predicted to reduced anxiety symptoms at posttreatment. Additionally, child self-efficacy was significantly related to child anxiety symptoms, child symptom severity, and level of child functioning within the family at pre-, mid-and post-treatment.

Sample Comorbidity
This clinical sample of treatment-seeking parents and youth, while consistent with previous research on many demographic characteristics, demonstrated at pre-treatment that greater than half of the youth met criteria for multiple diagnoses of anxiety disorders as well as other disorders. Specifically, 71% of the sample received at least two or three concurrent diagnoses determined via a semi-structure assessment (ADIS-IV revised for DMS-5 (APA, 2013)). Further, post-treatment outcomes indicated that 50% of children ended treatment with no diagnoses and another 42% met criteria for only one diagnosis (80% of which was a single anxiety disorder diagnosis). These statistics mirror diagnoses rates in other studies focused on CBT for pediatric anxiety disorders (i.e., Cartwright-Hatton et al., 2006).
This study's focus is novel in its examination of self-efficacy as a factor postulated to contribute to engagement during exposure and assist in explaining 'why' children engage in exposure. When considering the comorbid diagnoses, it seems that the compounded symptom presentations would dampen the treatment outcomes seen in this study; however, they did not. Interestingly, among this small clinical sample, child selfefficacy did not differ according to comorbidity (presence/absence) at pre-and midtreatment. However, significant differences in post-treatment child self-efficacy were found between those children who had two diagnoses versus those with four to five diagnoses in that those with greater comorbidity showed less perceived self-efficacy than those with fewer diagnoses at the end of treatment.

The SEE-C: Psychometric Evaluation
Expert reviewer and youth feedback indicated that the SEE-C child version showed acceptable and appropriate face and content validity. The measure was found to be adequate in design and content for the age and population under study. The measure also subjectively appeared to measure the construct it was supposed to measure. For example, expert reviewers indicated an interest in the creation of the SEE-C as a tool to address a child's in-session beliefs about self-efficacy during exposure practice. Expert reviews were also positive about the SEE-C's ability to measure one of the child factors, i.e., self-efficacy, that is likely to improve treatment engagement during exposure and resultant treatment outcomes.
Because of the acceptable to excellent internal consistency for the measure as a whole, exploratory PCA with Verimax rotation was utilized and suggested a three-factor structure. These three factors appear relevant to the measure, and the results are promising, as future confirmation of the factor structure (i.e., subscales) may better explain the results of a child's self-efficacy overtime, therefore providing greater information as to a child's engagement during exposure. For instance, in this pilot study, the child-report of self-efficacy demonstrated a significant increase from pre-to midtreatment; however, it then slightly reduced from mid-to post-treatment (although still significantly increased from pre-treatment). While these results are interesting, the reason for the increase and slight decrease can only be inferred. Where as, if we were able to utilize the measure's subscales, we could better explain a child's perceived self-efficacy overtime via the more concise definition of each subscale, as each of the factors would explain a unique component of the child's self-efficacy.
In addition, there was significant positive association and medium effect size between the SEE-C child version and the SEE-C parent version at post-treatment. While is does not provide evidence of construct validity (the degree to which a measure adequately evaluates the construct it claims to assess), since both measures are newly created, it does further assist in supporting the face and content validity. This is observed as both the parent and the child responder appeared to understand and be able to report on perceived child self-efficacy during exposure. While there is no simple metric to quantify this measure's construct validity (Westen & Rosenthal, 2003), particularly because this appears to be the first to measure self-efficacy in CBT for pediatric anxiety, correlations between the SEE-C child version and measures of child anxiety symptoms and functioning demonstrate relationships in directions expected in a measure of selfefficacy. For example, there was a negative association and large effect size of reported child anxiety symptoms with child self-reported self-efficacy as well as another negative correlation and large effect size between child functioning and child self-reported selfefficacy.

Engagement During Exposure
In order for CBT with exposure to be effective, exposures must be completed, as exposures are one of the main or "active" ingredients in CBT for pediatric anxiety disorders (Hudson & Kendall, 2002). As such, a focus on the engagement during exposure practice is necessary. Engagement is not always easy, as it involves multiple components including a motivational commitment and behavioral involvement and a child's participation, therapist and child collaboration, and the child's belief of selfefficacy to continue in the agreed-upon and identified intervention (King, et al., 2014).
Additionally, and as stated earlier, Positive Psychological Momentum (PPM) is the "perception, attitude, belief, or state-of-mind an individual experiences, in which their initial success leads to more successes" (Iso-Ahola & Blanchard, 1986). So, the hope is that once a child feels self-efficacious in completing a task competently multiple times (such as exposures), he/she will become confident in their abilities. As a result, a child's in-session self-efficacy and motivation to experience and feel distress in the face of a feared stimulus may be one of the important ingredients in the engagement in exposure practice.
Perceived self-efficacy and motivation. Study findings indicate that child perceived self-efficacy is negatively related to child anxiety symptoms over time (pre-, mid-, and post-treatment), whereby when a child's perceived self-efficacy is low, their anxiety symptoms are high. Furthermore, as treatment progresses, a child's self-efficacy increases and their anxiety symptoms decrease. This is also true in the relationship between child perceived self-efficacy and symptom severity as well as functional impairment (social, school, and family domains). These findings provide evidence into one, seemingly impactful, factor that contributes to a child's engagement during exposure.
This study's goal was to define child factors that contribute to engagement in exposure, including the perceived self-efficacy and motivation a child believes he/she has and uses to incrementally approach a feared situation/event. However, by identifying factors that influence engagement, we also need to consider that these factors under investigation may be stimulated by the anxiety (the "challenge") that maintains the momentum to generate self-efficacy and motivation. Once a "challenge" is conquered, another "cycle" of momentum needs to be generated to deal with the next challenge.
Depending on the child, it may take some time to build the stamina needed to wade through potentially multiple challenges presented by anxiety. A focus on additional child factors of engagement during exposure are outside the scope of this current study; however, the consideration of these additional factors is necessary and leads us to the future directions of this research, as well as highlights the importance of and need to consider parent and therapist factors that may exist.

Limitations
While this study has multiple strengths, it does not go without limitations. Expert reviewers reviewed the measure only once rather than multiple times through an iterative feedback process. While an iterative process would have provided more feedback regarding the edited measure's face validity, a consensus regarding the measure's clinical utility occurred with a singular review. This study was also limited by the child feedback portion of the study, as there was a small sample size of child reviewers (N=2) and no variation in gender and age. However, these were the first two recruited child participants into the study and defined by the research proposal to be the child feedback participants.
Nonetheless, the child reviewers' feedback was valuable and provided information on the developing measure in a structured interview after the measure was administered, and the children's answers were queried in real-time. This feedback was thorough and qualitatively as well as quantitatively assisted in shaping the measure.

Medication History:
Please provide information about all medications that your child is currently taking:

Current Dose
Please provide information about medications that your child has taken for psychiatric problems in the past: