Systematic Review on the Use of the Children’s Depression Inventory-2 Among Hispanics

Abstract

We systematically reviewed studies reporting the use of the Children’s Depression Inventory-2 (CDI-2) in samples with at least 30 children and significant Hispanics enrollment (≥14.5% of the sample and at least 20 Hispanics completing the scale). We grouped studies by form (short or full-length) and language used, developmental stage, report of psychometric data (particularly for Hispanics), and other characteristics. From 252 full-texts revised, 22 met selection criteria. Six reported psychometric data for Hispanics, either for the English full-length (α = .86–.92) and short form (α = .76–.81) or for the Spanish short form (α = .69–.80). Criterion-related validity was supported via correlation/regression or comparing group means, but not using another depression self-report scale. Current knowledge on the CDI-2 psychometrics among Hispanics is mostly based on studies with the English-language version. No study has reported the psychometrics of the full-length Spanish-language CDI-2 with Hispanics.

Keywords

Children’s Depression Inventory-2 psychometrics Hispanics systematic review

Since 1977, the Children’s Depression Inventory (CDI) has been used to screen youth for depression and to assess the efficacy of depression treatment. A second edition of its self-report form, known as the CDI-2 SR, was published, and evidence of its psychometric properties with U.S. children was provided in a technical manual (Kovacs & Multi-Health Systems Staff, 2011). After conducting some changes in the scale, data were collected for a normative sample of 1,100 U.S. children and adolescents and a clinical sample of 319 youths, which formed a total sample of 1419 children.

The 28-item CDI-2 Total score has an alpha (α) coefficient of .91. This value was .90 for children aged 7 to 12 years and .92 for those aged 13 to 17 years. The measure has two higher order scales, named Emotional (α = .85) and Functional Problems (α = .83). The first includes the Negative Mood/Physical Symptoms (NM/PS; α = .75) and the Negative Self-Esteem (NSE; α = .77) subscales. Subscores on Ineffectiveness (INE; α = .76) and Interpersonal Problems (IP; α = .73) comprise the second scale. A 12-item short form (SF) is available (α = .82). It correlated .95 with the full-length form (Bae, 2012; Kovacs & Multi-Health Systems Staff, 2011). CDI-2 scores correlated (N = 266) with scores on the Beck Depression Inventory-Youth (.37) and the Conners Comprehensive Behavior Rating Scales-Major Depressive Episode scale (.58; Kovacs & Multi-Health Systems Staff, 2011).

Despite the normative sample for the English version of the CDI-2 aligned with the U.S. Census 2000 figures, with 14.5% of Hispanic youth (N = 160), psychometric analyses for racial or ethno-cultural groups are not available. Although White and African Americans had similar mean scores, Hispanics obtained significantly lower T-scores than Whites in the CDI-2 Total, Emotional Problems scale and NM/PS subscale (Kovacs & Multi-Health Systems Staff, 2011). Differences were significant after adjusting for age and sex, but effect sizes (partial eta squared) were very small, (.012 for the Total score). Similar differences were reported for the 12-item SF, with an effect size of .011 (Kovacs & Multi-Health Systems Staff, 2011). Yet, differences among Hispanics and Whites were not found in two studies that used the SF (Seegan, 2018) or the full-length form (Babakhanyan, 2013) with U.S. youth. In a third study, Hispanic youth were compared with a collapsed group of non-Hispanic Whites, Asian American, Native American, African American, and youth from mixed background. No difference was found in outcomes, including CDI-2 SF scores (Rogers et al., 2017). In a fourth study, being non-Caucasian was related with higher scores on the standard CDI-2 than among Caucasians (Becker et al., 2013).

Hispanics are an 18.3% (59.87 millions) of the U.S. population, being the nation’s largest ethnic/racial minority (U.S. Census Bureau, 2019a). As of July 1, 2018, about 25.2% of children 7 to 17 years in the United States were Hispanic (U.S. Census Bureau, 2019b). Among foreign-born Spanish speakers aged 5 years and over who live in the United States and speak Spanish at home, 68.2% speak English less than very well. The rate of children aged 5 to 17 years who speak Spanish at home but speak English less than very well was 19.7% (U.S. Census Bureau, 2018). Thus, knowing the state of evidence on the use and psychometrics of the Spanish CDI-2 (and not only the English version) may help to improve the quality of depression assessment not only for Spanish-speaking youth in Hispanic American countries but also for a significant amount of children in the United States.

In 2011, an official Spanish-language version of the CDI-2 was published (Kovacs & Multi-Health Systems Staff, 2011). Yet, no literature review has been conducted on studies using the CDI-2 with samples of Hispanic youth. We conducted a systematic review on studies reporting on the use of the CDI-2 full-length or SF in samples with significant enrollment of Hispanic youth and provided a qualitative synthesis of research that have reported the use of the English- or Spanish-language CDI-2 with prepubertal children or adolescents of Hispanic origin. Emphasis was given in the report of data on the psychometric properties of the scale among this population.

Method

Eligibility Criteria and Search Strategy

Journal articles, dissertations, and other electronically available sources were part of this review. To be eligible, studies must (a) have been published between January 1, 2010, and July 31, 2019; (b) have a sample size of at least 30 youths; (c) have used the English- or Spanish-language CDI-2; and (d) have a meaningful enrollment of Hispanics (≥ 14.5% of the overall sample, and at least 20 Hispanics completing the scale). Thirty was the minimum sample size due to Nunnally’s (1978) argument against alpha estimation with sample sizes lower than the number of items in the scale. Criteria for the inclusion of Hispanics were based in the percentage of their enrollment in the CDI-2 normative sample (Kovacs & Multi-Health Systems Staff, 2011). Requiring at least 20 Hispanics completing the CDI-2 in any given sample provided certainty that studies with small samples will have a predominantly (i.e., at least two thirds if N = 30) Hispanic sample.

We identified studies using redundant search procedures, including an electronic search, followed by a hand search of the reference lists of selected articles. PsycINFO, Psychology and Behavioral Science Collection, Psychology Database, Web of Science, ScienceDirect, SpringerLink, Google Scholar, Education Source, Academic Search Ultimate, ERIC, Sociology Source Ultimate, Open Dissertations, CINAHL, Caribbean Search, Fuente Académica Plus, eBook Collection, E-Journals, and MEDLINE articles from January 2010 to July 2019 were searched. The following keywords were used in any text field: Children’s Depression Inventory-2, Children’s Depression Inventory-2nd Edition, CDI-2, CDI2, or CDI-II.

Data Analysis

After retrieving records, duplicates were eliminated. Titles, abstracts, and keywords of nonduplicate records were revised to identify articles, book chapters, manuals, dissertations, or thesis that specifically reported about the use of the CDI-2. We excluded records reporting on another scale with the same acronym or on the CDI first edition, and any other unrelated record. Once target records were identified, full texts were collected and revised to distinguish between studies using the English- or Spanish-language version vs. other versions. We also excluded records that described the English version or previous psychometric reports but did not provide results from an independent sample. Next, we examined which of the selected full texts reported on samples with sizes ≥ 30 with a meaningful inclusion of Hispanic youth (as defined earlier).

Such records were classified into those providing some reliability or validity data for the CDI-2 in the general sample, studies that did not, and those with psychometric data specific for Hispanics. We distinguished studies considering the use of the full-length vs. the SF of the scale. Third, a distinction was made among studies using the English- vs. the Spanish-language version of the measure. In addition, studies were organized considering the developmental stage of participants. For the purpose of this review, a sample was considered to be of adolescents only if most participants were aged at least 12 years but no older than 18 years. Samples in which most participants had ages 11 years or below (such as samples of students from sixth grade and under) were considered to be of prepubertal children. In the fifth place, studies were analyzed with regard of their sample size. If using the full-length form, a sample was considered small if less than 140 youths were enrolled (rate of participants per scale item is lower than 5:1). A sample size was labeled as medium if enrollment was between 140 and 279 youths (a rate of at least 5:1 but less than 10:1). Samples with 280 participants or more were considered large. If using the CDI-2 SF, sample size was labeled as follows: small (less than 100 participants), medium (from 100 to 199), and large (200 and above). Sixth, studies with a majority of Hispanic participants were identified. Seventh, we identified studies with samples comprised by youth from families of middle-low or low socioeconomic status (SES), who were living in economic hardship, or who were facing very stressful life situations. Eighth, we examined if youth from any biological sex predominated in each sample and estimated its sex ratio. Ninth, studies were grouped as peer-reviewed articles or thesis/dissertations. Last, we counted studies with samples recruited in the United States, in Hispanic American countries, or part in the United States and part in a Hispanic American country.

Results

Selection, Number, and Categories of Eligible Studies

We collected 975 records. Duplicates (419) were deleted, and a pool of 556 records was screened for eligibility. Potential useful records were reduced to 252, by eliminating studies that reported on the use of other scale or unrelated to this review. We obtained full texts for these records and excluded 230 of them, as described in the “Data Analysis” section (Figure 1).

Figure 1.

Flowchart for the systematic literature review on the use of the CDI-2 among samples with meaningful inclusion of Hispanic youth.

Table 1 includes a profile of the 22 studies that met the eligibility criteria. Six reported some psychometric data for Hispanics and 16 did not. Eleven of those 16 records (five of which did not report any psychometric data at all) described studies in which the full-length form was used. In seven of those 11 studies, the authors administered only the English version, while most participants in another study completed the Spanish version. In the remaining three, the authors did not clarify the proportion of youth who completed either version, although it was implicit that the English version predominated. In the other five studies in which the authors did not report psychometric data for Hispanics, they used the English-language SF. Among the six studies in which the authors did report some psychometric data for Hispanics, three used the full-length form (completed mostly in English) and three used the SF: one using the English-language version and two using the Spanish-language version. Psychometric data for Hispanics were reported in four peer-reviewed articles (Gulbas et al., 2016; Park et al., 2017; Suárez-López et al., 2019; Zayas et al., 2015) and two dissertations (Marchante-Hoffman, 2018; Scanlon, 2016). In these six studies, as well as in Tropez-Arceneaux et al. (2017), the sample was exclusively Hispanic.

Table 1.

Descriptive Profile of Eligible Studies According to 10 Categories for Classification.

Studies	Type	EV or SV	CDI-2 form	Sample is mostly Hispanic	Psychometric data for Hispanics	Psychometric data but not for Hispanics	No psychometric data	Sample size	Sample origin	Economic hardship or extreme stress	Developmental stage	Main biological sex (sex ratio)
DeLay et al. (2017)	PRA	EV	SF			X		Large	United States	ML/L SES	PPC	Girls (5.3:4.7)
Rogers et al. (2017)	PRA	EV	SF			X		Large	United States	ML/L SES	PPC	Girls (5.4:4.6)
Glynn et al. (2019)	PRA	EV	SF			X		Medium	United States		ADOL	Girls (5.2:4.8)
Swift (2016)	DD	EV	SF			X		Large	United States		PPC	Equal ratio
Seegan (2018)	DD	EV	SF			X		Medium	United States		COMB	Boys (5.7:4.3)
Park et al. (2017)	PRA	Both	SF	Yes	X			Large	United States	ML/L SES	ADOL	Girls (5.7:4.3)
Marchante-Hoffman (2018)	DD	SV	SF	Yes	X			Medium	United States	ML/L SES	COMB	Girls (5.3:4.7)
Suárez-López et al. (2019)	PRA	SV	SF	Yes	X			Large	ECU	ML/L SES	ADOL	Girls (5.1:4.9)
Hockenberry et al. (2017)	PRA	Both	FLSR				X	Medium	United States	ALL	COMB	Boys (5.5:4.5)
Hooke et al. (2018)	PRA	Both	FLSR				X	Large	United States	ALL	COMB	Boys (5.2:4.8)
Rodgers et al. (2019)	PRA	Both	FLSR				X	Large	United States	ALL	COMB	Boys (5.2:4.8)
Esposito-Smythers et al. (2019)	PRA	EV	FLSR				X	Medium	United States	Suicidal crisis	ADOL	Girls (7.6:2.4)
Tropez-Arceneaux et al. (2017)	PRA	SV	FLSR	Yes			X	Small	NIC	Burned patients	ADOL (61%) and young adults	Girls (5.8:4.2)
Babakhanyan (2013)	DD	EV	FLSR			X		Medium	United States		COMB	Boys (5.3:4.7)
Gelley (2014)	DD	EV	FLSR			X		Medium	United States	ML/L SES	ADOL	Girls (5.2:4.8)
Morton (2018)	MT	EV	FLSR			X		Small	United States		PPC	Girls (5.9:4.1)
Mathai (2018)	MT	EV	FLSR			X		Small	United States		PPC	Girls (5.9:4.1)
Nelson et al. (2018)	PRA	EV	FLSR			X		Large	United States	ML/L SES	PPC	Boys (5.2:4.8)
Becker et al. (2013)	PRA	EV	FLSR			X		Medium	United States		PPC	Boys (5.9:4.1)
Scanlon (2016)	DD	Both	FLSR	Yes	X			Medium	United States		COMB	Girls (5.7:4.3)
Zayas et al. (2015)	PRA	Both	FLSR	Yes	X			Small	United States/MX	GEH	COMB	Girls (6.0:4.0)^a
Gulbas et al. (2016)	PRA	Both	FLSR	Yes	X			Small	United States/MX	GEH	COMB	Girls (5.8:4.2)

Note. To increase clarity, the category about reporting of psychometric data was divided into three mutually exclusive columns. PRA = peer-reviewed article; DD = doctoral dissertation; MT = master thesis; EV = English version; SV = Spanish version; Both = English and Spanish version were available but most youth completed the English version; SF = short form; FLSR = full-length self-report form; NIC = Nicaragua; ECU = Ecuador; MX = Mexico; ALL = acute lymphoblastic leukemia patients; GEH = general economic hardship was mentioned in the sample description, although no data were provided on SES; ML/L SES = most participants were from middle-low or low socioeconomic status according to specific data provided (e.g., mean income, percent eligible for free or reduced meals at school, percent of parents with only a high school education or less, etc.); PPC = prepubertal children; ADOL = adolescents; COMB = prepubertal children and adolescents are combined in the sample; CDI-2 = Children’s Depression Inventory-2.

Over 60% of the sample recruited comprised girls, but the specific percent was not provided.

Samples of prepubertal children (n = 7) or that combined them with adolescents (n = 9) were more common than those including only or mostly adolescents (n = 6). Nine of the 22 studies had medium-size samples. Among those studies, four reported on the SF and five on the standard form. Eight of the 22 studies had large samples: four used the SF, and four, the standard form. Seven of those eight studies were peer-reviewed articles. In contrast, the five studies with small-size samples reported on the use of the standard CDI-2 and enrolled more girls than boys. Three of the latter studies were exclusively with Hispanics (Gulbas et al., 2016; Tropez-Arceneaux et al., 2017; Zayas et al., 2015). Among other studies with samples in which all participants were Hispanics, two had a medium (Marchante-Hoffman, 2018; Scanlon, 2016) and two had a large sample (Park et al., 2017; Suárez-López et al., 2019). Four of the eight studies with large samples enrolled prepubertal children. Seven of the nine studies with samples combining children and adolescents had samples of at least medium size. The exceptions were studies that exclusively enrolled Hispanics. Overall, 14 studies had samples mostly comprising girls; seven had samples mostly comprising boys, and one had a sample with equal distribution by sex. The study with the largest difference in the biological sex ratio had a girls-to-boys’ ratio of 7.6:2.4 (Esposito-Smythers et al., 2019). The average ratio across studies was 5.3:4.7 in favor of girls.

Most studies (18) had samples entirely recruited in the United States. The samples from two studies were recruited in Hispanic American countries (Nicaragua or Ecuador), and the remaining two samples were recruited partially in the United States and Mexico (Gulbas et al., 2016; Zayas et al., 2015). Fourteen studies were conducted with samples mostly comprising youths living in families with middle-low/low SES or with economic hardship (9), or by youths living in extremely stressful circumstances, such as having a severe physical illness (3), being through a suicidal crisis (1), or having experienced a physical trauma (1). Seven of the eight studies with a large sample were conducted with participants who met criteria for one of these subcategories. Fourteen eligible studies were peer-reviewed articles and eight were either dissertations (6) or theses (2). Nine of the 14 studies reporting about the full-length CDI-2, and five of the eight studies that reported on the CDI-2 SF, were peer-reviewed articles. Next, we present a qualitative synthesis of the findings of studies using the SF and the full-length form.

Qualitative Synthesis of Studies Using the CDI-2 SF

In six studies, the CDI-2 SF was administered using mostly or only its English-language version (Table 2). In the only of those studies that reported data for Hispanics, 97.8% of the U.S. adolescents of Mexican origin completed the scale in English. Alpha values from .76 to .81 were found along three assessments (Park et al., 2017). Positive associations for the SF with anxiety, discrimination, and outward anger expression scores, and a negative one with anger control were observed. In Arizona, a modified 12-item English version was used with 299 sixth graders (37% Hispanic). Alpha values from .70 to .74 were reported (DeLay et al., 2017). Significant associations were found among Time 1 experiences of homophobic name-calling and Time 2 CDI-2 SF scores. In related analyses with a subsample of 280 children (44% Hispanic), a negative relation between the SF and academic engagement and a positive link with conformity to traditional masculinity were reported (Rogers et al., 2017). Glynn et al. (2019), on their behalf, found significant correlations between the English SF and measures of anxiety, childhood unpredictability, and anhedonia among 175 adolescents (30% Hispanic) from California. After assessing 1409 children aged 8 to 12 years (15% Hispanics) using the English SF in Delaware, Swift (2016) reported alphas from .81 to .82. She also found positive correlations with self-reports on bullying victimization and perpetration, with anxiety symptoms, and with teachers’ reports of youth’s depressive symptoms. Yet, in the last four studies, analyses were conducted without considering race/ethnicity. This limitation also applied to a study by Seegan (2018), who reported an alpha of .74 among 197 youth from Texas, aged 9 to 18 years (18% Hispanic), positive correlations with anxiety and peer victimization, and negative ones with hope and social support.

Table 2.

Studies That Reported About the Use of the Children’s Depression Inventory-2 Self-Report Short Form Among Hispanics.

Study	Sample	Measures	Internal reliability	Validity	Other relevant information
Studies that did not report psychometric data specifically for Hispanics
DeLay et al. (2017)	299 sixth graders (M age = 11.13 years); 53% girls; 37% Hispanic; 71% qualified for free meals at a large public school (Arizona)	CDI-2 SF (EV) with a modified 1 to 4 response scale, HCAT	Alpha coefficients of .70 in T1 (fall) and .74 (T2) (spring)	Experiences of homophobic name-calling at fall predicted higher levels of depressive symptoms at spring (estimate = 1.33, SE = 0.46, odds ratio = 1.56).	Correlation of CDI-2 scores between T1 and T2 was of .61 (5-month interval, p ≤ .001).
Rogers et al. (2017)	280 sixth graders (M age = 11.13 years); 44% Hispanic 54.3% girls; 75.9% qualified for free or reduced meals at a public school (Arizona)	CDI-2 SF (EV) with a 1 to 4 response scale, AMIRS, ESF	Same as above	CDI-2 scores were positively related with conformity to traditional masculinity (.20–.31) and negatively related (r = −.41 to −.53) with academic engagement.	Correlation of T1 and T2 CDI-2 scores was of .60 (5-month interval, p ≤ .001).
Glynn et al. (2019)	175 adolescents (M age = 13.7 years); 30% Hispanic; 52% girls; M annual income = US$145,314; study conducted in California	CDI-2 SF (EV), QUIC, STAIC, MASQ-AD, CLES	Not reported	Significant correlations with trait anxiety (.63), anhedonia (.61), and childhood unpredictability (.42) but no association with childhood life events (.08).	The study was focused on the validation of the QUIC.
Swift (2016)	1409 children from fourth and fifth grade; aged 8 to 12 years (M age = 10.1 years); 15% Hispanics; 50% girls; from nine schools in a diverse district in Delaware; 43.5% low income	CDI-2 SF (EV), PVS, CSTPV, BBS, BASC-2 DEP, MASC-10	Alpha values from .81 (T1, N = 1293) to .82 (T2, N = 1218)	Significant correlations with self-reports of bullying victimization (from .33 to .48) and perpetration (.24) and anxiety symptoms (.42 to .43), and with teachers reports of youth’s depressive symptoms (.21–.24).	Correlation between T1 and T2 CDI-2 scores was .64 (7-month interval, p ≤ .001, N = 1188).
Seegan (2018)	197 youth aged 9 to 18 years (M age = 13.6 years); 18% Hispanic; 57% boys; from an urban summer program and a rural public school in Texas; mean monthly income = US$6,177.38	CDI-2 SF (EV), SEQ, RASEQ, SCAS, CASSS, CASH	Alpha = .74	Positive correlations with anxiety (.52) and peer victimization (.38), but negative with social support from friends (−.20) and hope (−.41) scores.	Girls had higher CDI-2 scores. Income was negatively related to CDI-2 scores. No differences were found by ethnicity/race.
Studies that reported psychometric data specifically for Hispanics
Park et al. (2017)	269 adolescents of Mexican origin; ages 12 to 17 years (M age = 14.1 years); 71.0% U.S. born; 56.9% girls; most parents had a high school education or less; recruited in public schools, churches, and community-based organizations in a mid-sized Midwestern region in the United States	CDI-2 SF (EV & SV), PRaCY, MASC-10, STAXI-2 C/A	Alpha of .81 (T1; N = 269), .76 (T2, N = 246), and .81 (T3, N = 246), respectively	Depression scores were positively related with anxiety (.39–.51), discrimination (.33–.39) and outward anger expression (.47–.50) considering three time points. They were negatively related with anger control (−.16 to −.24 at T2 and T3). Correlation at T1 was nonsignificant (−.10).	Only six youths completed the survey in Spanish via face-to-face interviews with a bilingual interviewer or a Spanish-language written questionnaire.
Marchante-Hoffman (2018)	152 first-generation foreign-born Latino youth aged 8 to 17 years (M age = 13.02 years); 53% girls; 65.8% from Venezuela; low-income patients of a mobile clinic in South Florida	CDI-2 SF (SV), CSI-24, FAD-GF, ASS, SPC, UCLA PTSD RI-DSM-5	Alpha = .80	Positively related with somatic symptoms (.60), sleep problems (.48), number of PTE (.44), severity of PTSD symptoms (.65), and acculturative stress (.46), and negatively related to family functioning (−.31).	About 90.8% of them preferred to complete the scale in Spanish. Sample mean score was similar to the U.S. normative sample.
Suárez-López et al. (2019)	529 adolescents aged 11 to 17 years (median age = 14.38 years); 51% girls; from a rural, low-income county in the Ecuadorian Andes	CDI-2 SF (SV), AChE activity (finger-stick sample); MASC-2	Alpha = .69, N = 516 (Data were provided by the first author)	A 1.09 T-score increase per SD decrease of AChE, and a 2.11 T-score decrease per SD increase of AChE was found in CS and longitudinal data, respectively.CDI-2 as MASC-2 scores correlated .22 (p < .001).	The associations between AChE and CDI-2 scores were stronger among girls and among younger children (<14.38 years old).

Note. Only results with at least one significant correlation showing an absolute value ≥.20 were considered meaningful evidence of validity. Only measures that were analyzed in relation to CDI-2 SF scores are included. CDI-2 SF = Children’s Depression Inventory-2 Short Form; EV = English version; SV = Spanish version; T = time; HCAT = Homophobic Content Agent Target Scale; AMIRS = Adolescent Masculinity Ideology in Relationships Scale; ESF = Experience Sampling Form (6-item academic engagement measure); QUIC = Questionnaire of Unpredictability in Childhood; STAIC = State-Trait Anxiety Inventory for Children; MASQ-AD = Mood and Anxiety Symptom Questionnaire-Anhedonic Depression scale; CLES = Coddington Life Events Scale for Children; PVS = Peer Victimization Scale; CSTPV = Comprehensive Scales of Traditional Peer Victimization; BBS = Bullying Behavior Scale; BASC-2 DEP = Depression subscale from the Behavioral Assessment Scale for Children-Teacher Form; MASC-10 = Multidimensional Anxiety Scale for Children-10 item version; SEQ = Social Experiences Questionnaire; RASEQ = Revised Adolescent Social Experiences Questionnaire; SCAS = Spence Children’s Anxiety Scale; CASSS = Child and Adolescent Social Support Scale; CASH = Child and Adolescent Scales of Hope; PRaCY = Perceptions of Racism in Children and Youth; STAXI-2 C/A = State-Trait Anger Expression Inventory-2 Child/Adolescent version; CSI-24 = Children’s Somatization Inventory; FAD-GF = Family Assessment Device-General Functioning Scale; ASS = Acculturative Stress Scale; SPC = Sleep Problems Composite (2-items); UCLA PTSD RI-DSM-5 = UCLA Post-Traumatic Stress Disorder Reaction Index for DSM-5; PTE = potentially traumatic events; AChE = acetylcholinesterase; MASC-2 = Multidimensional Anxiety Scale for Children-2nd Edition; CS = cross-sectional.

In a couple of studies, youth completed the CDI-2 SF using mostly or only its Spanish version. In one of those, the authors reported an alpha value of .80 among 152 first-generation foreign-born Latino youth aged 8 to 17 years (Marchante-Hoffman, 2018). About 90.8% of them preferred to complete the scale in Spanish. CDI-2 SF scores were positively related with somatic symptoms, sleep problems, number of potentially traumatic events, severity of post-traumatic symptoms, and acculturative stress, and negatively related to family functioning. In a recent publication, Suárez-López et al. (2019) used the Spanish-language SF with over 500 adolescents from the Ecuadorian Andes. In cross-sectional analysis, greater exposure to cholinesterase inhibitors (lower acetylcholinesterase activity) was associated to higher T-scores in depressive but not in anxiety symptoms. In longitudinal analysis (n = 223), reductions in T-scores on the SF were related to increases in acetylcholinesterase activity from 2008 to 2016. Although these data were not included in the published report, the alpha coefficient obtained within the sample was .69 and the CDI-2 SF showed a significant positive correlation (.22) with self-reports of anxiety symptoms (Suárez-López, personal communication, July 2, 2019).

Qualitative Synthesis of Studies Using the CDI-2 Full-Length Self-Report Form

In the first 11 studies on Table 3, the English or the Spanish full-length CDI-2 was used, but psychometric data were not reported or not analyzed separately for Hispanics. For example, Hockenberry et al. (2017) administered the CDI-2 to 236 children with acute lymphoblastic leukemia (ALL; 45.8% Hispanic). Only those aged ≥ 7 years filled the self-report form. Even though English- and Spanish-speaking children participated, the authors did not report how many youths completed either version. Two related reports (Hooke et al., 2018; Rodgers et al., 2019) were based on children with ALL who completed the standard CDI-2 but using a larger sample (N = 327; 47.4% Hispanic). Esposito-Smythers et al. (2019) used the English full-length form during the assessments of a Randomized Control Trial (RCT) for youth (N = 147; 16.55% Hispanic) who had experienced a suicidal crisis. In a study with 33 Hispanics in Nicaragua (aged 12–25 years) who have suffered burns, Tropez-Arceneaux et al. (2017) administered the Spanish version to those aged younger than 18 years (n = 20). However, none of these five studies reported any data on reliability or validity.

Table 3.

Studies That Reported About the Use of the Children’s Depression Inventory-2 Full-Length Self-Report Form Among Hispanics.

Study	Sample	Measures	Internal reliability	Validity	Other relevant information
Studies that did not report psychometric data at all
Hockenberry et al. (2017)	236 children with ALL; aged 3 to 18 years; 123 aged ≥ 7 years; 45.8% Hispanic; 54.7% boys; from four oncology treatment centers in the United States	CDI-2 SR (EV and SV), Measures completed with iPADs	NDR	NDR	The authors did not report how many youth completed either the English or the Spanish version.
Hooke et al. (2018)	327 children with ALL; aged 3 to 18 years; 47.4% Hispanic; 179 aged ≥ 7 years; 52% boys; recruited as in Hockenberry et al.	Same as above	NDR	NDR	Same as above. CDI-2 scores decreased over time (from T1 to T4; about 12-month interval).
Rodgers et al. (2019)	Same as in Hooke et al.	Same as above	NDR	NDR	Same as above.
Esposito-Smythers et al. (2019)	147 youth aged 12 to 18 years (M age = 14.97 years); 16.55% Hispanic; 76.2% girls; from an U.S. inpatient/partial hospitalization program	CDI-2 SR (EV)	NDR	NDR	Significant changes in CDI-2 scores were observed at post-treatment and follow-up.
Tropez-Arceneaux et al. (2017)	33 Hispanic youth (aged 12 to 25 years) who have suffered burns and participated in a burn camp in Nicaragua; 58% girls; 79% from urban zones; 61% (n = 20) were aged 12 to 17 years	CDI-2 SR (SV), CDI-2 Parent version (SV); youth aged ≥18 years completed the BDI	NDR	NDR	Lower parent-rated CDI-2 scores at 6-month post camp. Analysis for youth who completed the CDI-2 SR was not provided.
Studies that did not report psychometric data specifically for Hispanics
Babakhanyan (2013)	168 youth aged 7 to 17 years (M age = 12.73 years); 22.6% Hispanic; 53% boys; 55.3% with family income of US$60,000 or more recruited in the Los Angeles county	CDI-2 SR (EV), DPS-C, History of any depressive disorder	NDR	CDI-2 scores significantly predicted cases with a depressive disorder; significant correlation (.36) with DPS-C (a process-focused measure of depression).	The DPS-C was the focus of this study. History of any depressive disorder was obtained from parental report (single question).
Gelley (2014)	233 seventh and eight graders; 52.36% girls; 27.04% Hispanic; 85% qualified for free or reduced meals in a middle school within a large, urban district in the Southeastern United States	CDI-2 SR (EV)	Alpha = .82	NDR	N/A
Morton (2018)	133 children aged 8 to 12 years (M age = 9.75); 59.4% girls; 27.8% Hispanic; most families from middle-upper SES levels; recruited at community events in a mid-sized city in Texas	CDI-2 SR and Parent form (EV), PCDS, BAI-Y, ARQ/Control, SASC-R, PCS-YSR	Alpha = .89	Positively related with general (.72) and social anxiety (.32–.59), hostility/rejection/neglect (.55), psychological (.45-.51) and behavioral control (.22-.32); negatively related with maternal warmth (−.38).	Correlation between CDI-2 self-reports and maternal CDI-2 reports was .50.
Mathai (2018)	Same as above	CDI-2 SR (EV), MGI, BAI, MASC-2; mental health Tx (nonpharmacological)	Same as above	Related to MASC-2 (.67) and MGI (.61) child self-reported scores, and to BAI (.22) parental self-reports. Children with an Hx of treatment showed higher scores.	SES was negatively related to CDI-2 SR scores.
Nelson et al. (2018)	280 children aged 8 to 10 years (M age = 8.95); 52.14% boys; 14.64% Hispanic; 56.4% receiving public medical assistance; recruited in a small Midwestern city in the United States	CDI-2 SR (EV), 9-tasks battery designed to measure major aspects of EC, RCMAS-2	Alpha = .80	Executive control scores at age 5 years and 3 months significantly predicted CDI-2 scores at follow-up (fourth grade), controlling for other variables. Depression and anxiety symptoms were significantly related.	Average time lapse between preschool EC assessment and elementary school depression/anxiety assessment = 4.7 years.
Becker et al. (2013)	188 children with ADHD-I; aged 7 to 11 years (M age = 8.67 years); 58.51% boys; 15.43% Hispanic; recruited from the San Francisco Bay Area; 81% of parents had completed college; mean annual income = US$85,000	CDI-2 SR and Parent form (EV), BASC	NDR	Positive correlations with self-reported anxiety (.62) and teachers’ ratings of youth externalizing problems (.22).	Correlation between CDI-2 self-reports and maternal CDI-2 reports was .25. Higher CDI-2 scores were related (r = −.24, p = 001) with being non-Caucasian.
Studies that reported psychometric data specifically for Hispanics
Scanlon (2016)	141 Hispanic children aged 9 to 13 years (M age = 10.92 years); 56.7% girls; SES reflected a predominantly middle-class sample; recruited from a community in West Texas	CDI-2 SR (EV and SV), SCS, PARQ/Control, RCMAS-2, SASC-R,	Alpha = .87	Negative correlations with youth reports of independent (−.35) and interdependent (−.28) self-construal. Positive correlations with self-reports of manifest (.63) and social (.47) anxiety, maternal hostility/rejection/neglect (.55) and acceptance problems (.43), and with mothers’ reports of child anxiety symptoms (.27). CDI-2 scores were higher for children with history of psychological treatment.	Only one child completed the SV. Correlation between CDI-2 self-reports and maternal CDI-2 reports was .40. M score was 8.85 (T = 54.76), which was higher than the one reported for the normative sample and for age- and sex-defined subgroups.
Zayas et al. (2015)	83 US-born citizen-children of Mexican origin, aged 8 to 15 years (M age = 11.4); more than 60% were girls; youth either lived in Mexico with deported parents (n = 31), lived in the United States but have a parent affected by detention or deportation (n = 18), or lived in the United States with undocumented parents who were not under removal proceedings (n = 34)	CDI-2 SR (EV and SV); other measures were used but relationships with CDI-2 scores were not examined	Alpha = .86	Youth affected by a parent’s deportation or detention (n = 49) reported significantly higher scores than their counterparts (n = 34) in CDI-2 Total and Emotional Problems scores, and in the NM/PS and NSE subscales.	No information was published about how many youths completed the scale in Spanish. Nobody endorsed Item 8 (suicide ideation or intent). Alpha, thus, was estimated with 27 items.
Gulbas et al. (2016)	48 US-born citizen-children of Mexican origin, aged 8 to 15 years (M age = 11.3 years; 58.3% girls; youths with extreme high or low T-scores (selected from Zayas et al.’s sample) were classified as probably depressed (n = 16) or not depressed (n = 32)	CDI-2 SR (EV and SV; qualitative data were collected via individual interviews	Alpha = .92	Half of the youth with probable depression cited the stressed relation with parents as a factor that was salient to their experience of suffering, compared with 9% of those without depressive symptoms.	Same as above. In this study, half of the participants (24) had undocumented parents affected by detention or deportation and half (24) had undocumented parents who did not.

Note. Only results with at least one significant correlation showing an absolute value ≥.20 were considered meaningful evidence of validity. Only measures that were analyzed in relation to CDI-2 SR scores are included. NDR = no data reported; ALL = acute lymphoblastic leukemia; CDI-2 SR = Children’s Depression Inventory-2 Full-Length Self-Report Form; EV = English version; SV = Spanish version; T = time; BDI = Beck Depression Inventory; DPS-C = Depression Process Scale-Children; SES = socioeconomic status; BAI-Y = Beck Anxiety Inventory for Youth; BAI = Beck Anxiety Inventory; SASC-R = Social Anxiety Scale for Children-Revised; PARQ/Control = Parental Acceptance and Rejection/Control Questionnaire; PCS-YSR = Psychological Control Scale-Youth Self Report; PCDS = Psychological Control-Disrespect Scale; PCM = Psychological Control Measure (Parental report); Hx = history; MASC-2 = Multidimensional Anxiety Scale for Children (Child and Parent versions); MGI = Maladaptive Guilt-Induction Measure; EC = executive control; RCMAS = Revised Children’s Manifest Anxiety Scale; RCMAS-2 = Revised Child Manifest Anxiety Scale, second edition; BASC = Behavior Assessment System for Children; ADHD-I = attention-deficit hyperactivity disorder-predominantly inattentive type; NM/PS = CDI-2 Negative Mood/Physical Symptoms subscale; NSE = CDI-2 Negative Self-Esteem subscale; SCS = Self-Construal Scale.

On the other hand, Babakhanyan (2013) studied 168 youth aged 7 to 17 years (22.6% Hispanic) from Los Angeles and found CDI-2 scores to be significant predictors of depression diagnoses and to correlate with scores in the Depression Process Scale (DPS). Gelley (2014) used the full-length English version with 233 youth (27.04% Hispanic) in Grades 7 to 8, reporting an alpha of .82. Similarly, Morton (2018) reported an internal consistency of .89 among 133 youth aged 8 to 12 years (27.8% Hispanic), as well as positive correlations with measures of anxiety and maternal psychological and behavioral control, but a negative association with maternal warmth. Inter-rater agreement between youth and maternal ratings on the CDI-2 was .50. Within the same sample, Mathai (2018) reported correlations for the CDI-2 with scores of parental and youth anxiety, and mother’s maladaptive guilt-induction on children. Children with history of nonpharmacological mental health treatment showed higher CDI-2 scores. Nelson et al. (2018) reported an alpha of .80 among 280 youth aged 8 to 10 years (14.64% Hispanic). CDI-2 scores and anxiety symptoms correlated. Moreover, among 188 attention-deficit hyperactivity disorder (ADHD) children aged 7 to 11 years (15.43% Hispanic), Becker et al. (2013) found correlations with self-reported anxiety, ratings of youth symptoms in the Parent version of the CDI-2, and teachers’ ratings of child’s externalizing problems. However, in these latter six studies, no separate analysis for Hispanics was reported.

Only three studies have reported psychometric data for the full-length CDI-2 among Hispanics. Scanlon (2016) used the standard form with 141 Hispanic children aged 9 to 13 years. Internal consistency was .87. Symptoms were higher for children with a history of psychological treatment. CDI-2 scores were negatively related with reports of independent and interdependent self-construal, and maternal acceptance, and positively related to self-reports of manifest and social anxiety and maternal hostility/rejection/neglect, and with mothers’ reports of their child depressive (r = .40) and anxiety symptoms. Although available in English and Spanish, only one child completed the Spanish version. Zayas et al. (2015) used the full-length form with 83 US-born citizen-children of Mexican origin, aged 8 to 15 years. Some lived in Mexico with deported parents. Some lived in the United States but had at least a parent affected by detention or deportation (D/D), and others lived in the United States with undocumented parents not affected by D/D. Youth affected by a parent’s D/D reported higher scores than others in Total and Emotional Problems scores, and in the NM/PS and NSE subscales. An alpha of .86 was found. Although data were collected in either Spanish or English, separated psychometric data for each version were absent and no information was given on how many youths completed the scale in Spanish (Zayas et al., 2015). This suggests that most children completed the English version. Further analyses with a subsample of 48 youth (α = .92) showed that 50% of those (n = 16) with probable depression (independently of their parents’ D/D status) cited the stressed relation with parents as a factor that was salient to their experience of suffering, compared with only 9% of those (n = 32) without depressive symptoms (Gulbas et al., 2016).

Discussion

Some important points emerged from our review. Although 17 studies reported some psychometric data, only six included psychometric data for Hispanics. Four reported about either the short (Park et al., 2017) or full-length (Gulbas et al., 2016; Scanlon, 2016; Zayas et al., 2015) English version and two reported about the 12-item Spanish version (Marchante-Hoffman, 2018; Suárez-López et al., 2019). These six studies included data about internal consistency (α) and criterion-related validity. Four of them documented validity using correlations (Marchante-Hoffman, 2018; Park et al., 2017; Scanlon, 2016; Suárez-López et al., 2019) or a linear regression (Suárez-López et al., 2019) and three used analyses to compare group means (Gulbas et al., 2016; Scanlon, 2016; Zayas et al., 2015). Studies documenting validity with correlations or regression used both some criteria expected to be negatively related and criteria expected to be positively related to CDI-2 scores. No study that provided validity data on the CDI-2 when used among Hispanics included another depression self-report scale as criterion. The only study that reported a correlation with another depression scale (not just with another form of the same measure) only enrolled 22.6% of Hispanics, did not report separate analyses for this group, and was aimed to support the validity of another scale (Depression Process Scale-Children [DPS-C]) not the CDI-2 (Babakhanyan, 2013).

Among studies that included evidence on the validity of the CDI-2 SR among Hispanics, one provided inter-rater correlations (IRCs) with mother’s ratings on the parent-rated English version (Scanlon, 2016). Other two studies reported IRCs using these two forms of the English-language CDI-2, in a range from .25 (Becker et al., 2013) to .50 (Morton, 2018), but these studies only enrolled from 15.43% to 27.8% of Hispanics. By contrast, the IRC of .40 reported by Scanlon (2016) was obtained in a 100% Hispanic sample. This IRC value compares favorably with coefficients (.34–.36) reported in the CDI-2 manual (Kovacs & Multi-Health Systems Staff, 2011).

Many eligible studies (11) reported psychometric properties that were not specific to Hispanics. That was the case for five studies that reported on the use of the English-language CDI-2 SF, finding alpha coefficients from .70 to .82 and a wide range of convergent and divergent associations with validity criteria. Similarly, six studies reported on the use of the English full-length form without considering ethnicity/race, finding alpha values from .80 to .89 and a variety of correlations with validity criteria. In samples of the aforementioned studies that focused on the SF, inclusion of Hispanics ranged from 15% (Swift, 2016) to 44% (Rogers et al., 2017), while such inclusion ranged from 14.6% (Nelson et al., 2018) to 27.8% (Morton, 2018) in studies focused on the full-length form. Thus, in those 11 studies, the minority of children were Hispanics, and their inclusion was considerably lower in studies that used the full-length form.

Although it is uncertain to what extent psychometric findings of those 11 studies apply to Hispanics in their samples, there are similarities between their results and those from samples in which all participants were Hispanics. For example, in studies using the CDI-2 SF with Hispanic youth (either in English or Spanish), alpha values ranged from .69 to .81 (Marchante-Hoffman, 2018; Park et al., 2017; Suárez-López et al., 2019). Meanwhile, in studies that reported on the full-length form, reliability coefficients were from .86 to .92 (Gulbas et al., 2016; Scanlon, 2016; Zayas et al., 2015). In both cases, the range of reliability indexes was similar to the one reported within samples for which separate psychometric data for Hispanic were unavailable. Although patterns of correlations supported the validity of both CDI-2 forms in samples for which separate data for Hispanic were unavailable and in those comprising only Hispanics, correlational evidence of validity for Hispanics was more frequent for the SF than for the full-length form.

Three of the eight studies that used the 12-item SF had a sample comprised exclusively by adolescents: two with the English (Glynn et al., 2019; Park et al., 2017) and one with the Spanish version (Suárez-López et al., 2019). The other studies had either samples comprising prepubertal children (DeLay et al., 2017; Rogers et al., 2017; Swift, 2016) or a combination of prepubertal children and adolescents (Marchante-Hoffman, 2018; Seegan, 2018). Alpha values from .69 to .81 were observed in adolescent samples, with two of the three studies being conducted in Hispanic samples. Alpha coefficients showed magnitudes from .70 to .82 in studies with prepubertal children, although none of these published separate analysis for Hispanics. Alpha values from .74 to .80 have been found in studies combining youths from both developmental stages, with the higher coefficient belonging to the one conducted only with Hispanics. This suggest that, based on available studies, the internal consistency of the SF of the CDI-2 may be similar for children and for adolescents of Hispanic origin, at least to the extent that results from studies conducted exclusively with prepubertal children apply to Hispanic children. However, it is noteworthy that, as far as studies combining prepubertal children with adolescents did not publish separate analysis for either developmental group, no study has published psychometric data on the 12-item CDI-2 exclusively derived from prepubertal children of Hispanic origin.

Four of the 14 studies that reported on the use of the full-length form enrolled only prepubertal children, with alpha values from .80 to .89. Two studies included only adolescents, and one had a sample comprising mostly (61%) adolescents, but only Gelley (2014) reported an internal consistency value (.82) estimated exclusively for adolescents. No study has published psychometric data on the full-length CDI-2 (in either English or Spanish) exclusively derived from prepubertal children or from adolescents of Hispanic origin. Among the seven studies that reported on the full-length form in samples that combined children and adolescents, alpha values varied from .86 to .92, and those reporting reliability data were all conducted among Hispanics. Yet, none of those studies reported the psychometrics of the full-length Spanish CDI-2 among Spanish-speaking youth of any age group. In fact, our current knowledge on the psychometrics of the full-length form among Hispanics is based on three studies that used only or mostly the English version within samples that combined youth from different developmental stages.

We identified additional limitations in the current state of literature on the use of the CDI-2 with Hispanic youth. For example, none of the studies that included psychometric data for Hispanics reported results from a formal test–retest reliability study. The three studies (conducted on two samples) that estimated a correlation coefficient between CDI-2 scores obtained at two different assessment points reported coefficients in a range from .60 to .64 (DeLay et al., 2017; Rogers et al., 2017; Swift, 2016), but the interval between the two assessments (from 5 to 7 months) was too long. All three studies were conducted with the CDI-2 SF and the authors did not report separate analyses for Hispanics. Among eligible studies, a factorial or structural validity analysis was also lacking. In addition, only one eligible study provided any evidence of the sensitivity to change of the measure when used to assess treatment gains (Esposito-Smythers et al., 2019), but separate analyses by race/ethnicity were not considered. Although in other two studies do the authors mentioned that CDI-2 full-length form scores decreased over time (Hooke et al., 2018; Rodgers et al., 2019), such reduction did not support the sensitivity to change of the full-length scale as no treatment for depression was provided. On the other hand, only one study was conducted in a clinical sample of depressed youth (Esposito-Smythers et al., 2019) and none compared a clinical sample with a nonclinical one. Finally, as mentioned, no study reported the psychometric properties for the full-length CDI-2 using only (or mostly) the Spanish version.

In summary, what we currently know about the psychometrics of CDI-2 scores among Hispanic youth is based mostly on the use of the SF (in Spanish or English), or in the use of the English full-length form in studies that combined prepubertal children with adolescents, but did not provide separate psychometric analysis for either age group. Further studies should explore the psychometric properties of the CDI-2 in samples comprising only Hispanic prepubertal children, using exclusively either the Spanish- or the English-language version. The need for such studies applies for both the full-length and the SF of the scale. In addition, research on the Spanish-language version of the full-length CDI-2 using samples comprising exclusively Hispanic adolescents is needed. Either within samples of Hispanic prepubertal children or adolescents, future research should also examine the concurrent validity of the short and full-length forms of the English- and Spanish-language CDI-2 using another depressive symptoms self-report scale as validity criterion, not just the parent or teacher version of the CDI-2. Finally, researchers should conduct test–retest reliability studies of the CDI-2 (particularly for the Spanish version) among Hispanic youth, as these are currently inexistent. The same is true for studies on its factor structure among Hispanic youths, and for examinations of its sensitivity to change.

The current systematic review may encourage the conduction of additional empirical studies and systematic reviews regarding the use and the psychometrics of the CDI-2 not only among Hispanics but also among other ethno-cultural groups in the United States and in other countries. Findings from this study suggest the need for researchers to include psychometric data about the CDI-2 in their publications, even if their main purpose is not psychometric in nature. If conducting state-of-the-art research and clinical assessments grounded on best practices and research evidence when using the CDI-2 is a desired goal, evidence of its validity and reliability with specific groups must be provided. As far as we know, this is the first systematic review on the use and psychometric properties of the CDI-2 with a particular ethno-cultural group. More research is needed to increase the cumulative of evidence on its validity, reliability, and applicability across different ethno-cultural groups.

Footnotes

Author’s Note

The content is solely the responsibility of the author and does not necessarily represent the official views of the National Institutes of Health.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was possible thanks to funds granted by the President of the University of Puerto Rico to the Institute for Psychological Research, and to funds granted to the author by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health, under award number R03DK092547.

Author Biography

Eduardo Cumba-Avilés received his doctorate in clinical psychology from University of Puerto Rico, Río Piedras (UPR-RP). He is an assistant research scientist at the Institute for Psychological Research at UPR-RP. He is interested in the manifestation of depression and comorbid physical illnesses and mental health problems, as well as in the development and validation of outcome measures to assess the efficacy of psychosocial interventions aimed to treat those conditions. His main research has been focused on the assessment and treatment of depression among Latino youth, with emphasis on evidence-based interventions such as cognitive-behavior therapy.

References

Babakhanyan

(2013). Development and validation of a process focused measure of childhood depression [Doctoral dissertation, Loma Linda University]. https://search.proquest.com/docview/1426563030

Bae

(2012). Test review: Children’s Depression Inventory 2 (CDI 2). Journal of Psychoeducational Assessment, 30(3), 304–308. https://doi.org/10.1177/0734282911426407

Becker

McBurnett

Hinshaw

Pfiffner

(2013). Negative social preference in relation to internalizing symptoms among children with ADHD predominantly inattentive type: Girls fare worse than boys. Journal of Clinical Child & Adolescent Psychology, 42(6), 784–795. https://doi.org/10.1080/15374416.2013.828298

DeLay

Hanish

L. D.

Zhang

Martin

C. L.

(2017). Assessing the impact of homophobic name calling on early adolescent mental health: A longitudinal social network analysis of competing peer influence effects. Journal of Youth and Adolescence, 46(5), 955–969. https://doi.org/10.1007/s10964-016-0598-8

Esposito-Smythers

Wolff

J. C.

Liu

R. T.

Hunt

J. I.

Adams

Kim

. . . Spirito

(2019). Family-focused cognitive behavioral treatment for depressed adolescents in suicidal crisis with co-occurring risk factors: a randomized trial. Journal of Child Psychology and Psychiatry, 60(10), 1133–1141. https://doi.org/10.1111/jcpp.13095

Gelley

C. D.

(2014). Accuracy of educators in identifying middle school students with elevated levels of anxiety or depression [Doctoral dissertation, University of South Florida]. https://scholarcommons.usf.edu/etd/5221

Glynn

L. M.

Stern

H. S.

Howland

M. A.

Risbrough

V. B.

Baker

D. G.

Nievergelt

C. M.

. . . Davis

E. P.

(2019). Measuring novel antecedents of mental illness: The Questionnaire of Unpredictability in Childhood. Neuropsychopharmacology, 44(5), 876–882. https://doi.org/10.1038/s41386-018-0280-9

Gulbas

L. E.

Zayas

L. H.

Yoon

Szlyk

Aguilar-Gaxiola

Natera

(2016). Deportation experiences and depression among U.S. citizen-children with undocumented Mexican parents. Child: Care Health and Development, 42(2), 220–230. https://doi.org/10.1111/cch.12307

Hockenberry

M. J.

Hooke

M. C.

Rodgers

Taylor

Koerner

K. M.

Mitby

. . . Pan

(2017). Symptom trajectories in children receiving treatment for leukemia: A latent class growth analysis with multitrajectory modeling. Journal of Pain and Symptom Management, 54(1), 1–8. https://doi.org/10.1016/j.jpainsymman.2017.03.002

10.

Hooke

M. C.

Rodgers

Taylor

Koerner

K. M.

Mitby

Moore

. . . Pan

(2018). Physical activity, the Childhood Cancer Symptom Cluster–Leukemia, and cognitive function: A longitudinal mediation analysis. Cancer Nursing, 41(6), 434–440. https://doi.org/10.1097/ncc.0000000000000634

11.

Kovacs

& Multi-Health Systems Staff. (2011). Children’s Depression Inventory 2nd Edition (CDI 2) technical manual. Multi-Health Systems.

12.

Marchante-Hoffman

A. N.

(2018). Giving voice to underserved foreign-born Latino youth: Trauma, stress, and health in the primary care setting (Open access dissertations, 2112) [Doctoral dissertation]. https://scholarlyrepository.miami.edu/oa_dissertations/2112

13.

Mathai

B. B.

(2018). The relationship between maternal depression, anxiety, and social anxiety symptoms and psychological control [Master thesis, Texas Tech University]. http://hdl.handle.net/2346/73937

14.

Morton

A. R.

(2018). The relation between parental psychological control and youth internalizing symptoms moderated by parenting constructs [Master thesis, Texas Tech University]. http://hdl.handle.net/2346/73817

15.

Nelson

T. D.

Kidwell

K. M.

Nelson

J. M.

Tomaso

C. C.

Hankey

Espy

K. A.

(2018). Preschool executive control and iternalizing symptoms in elementary school. Journal of Abnormal Child Psychology, 46(7), 1509–1520. https://doi.org/10.1007/s10802-017-0395-1

16.

Nunnally

J. C.

(1978). Psychometric theory (2nd ed.). McGraw-Hill.

17.

Park

I. J. K.

Wang

Williams

D. R.

Alegría

(2017). Does anger regulation mediate the discrimination-mental health link among Mexican-origin adolescents? A longitudinal mediation analysis using multilevel modeling. Developmental Psychology, 53(2), 340–352. https://doi.org/10.1037/dev0000235

18.

Rodgers

C. C.

Hooke

M. C.

Taylor

O. A.

Koerner

K. M.

Mitby

P. A.

Moore

I. M.

. . . Wei

(2019). Childhood cancer symptom cluster: Leukemia and health-related quality of life. Oncology Nursing Forum, 46(2), 228–237. https://doi.org/10.1188/19.onf.228-237

19.

Rogers

A. A.

DeLay

Martin

C. L.

(2017). Traditional masculinity during the middle school transition: Associations with depressive symptoms and academic engagement. Journal of Youth and Adolescence, 46(4), 709–724. https://doi.org/10.1007/s10964-016-0545-8

20.

Scanlon

N. M.

(2016). Within-group differences of Hispanic children’s symptoms of depression, anxiety, and social anxiety and their relations with various parenting behaviors [Doctoral dissertation, Texas Tech University]. http://hdl.handle.net/2346/72280

21.

Seegan

P. L.

(2018). Internalizing symptoms and health behaviors associated with relational peer victimization: An examination of perceived social support and hope in relation to youth adjustment [Doctoral dissertation, Texas Tech University]. http://hdl.handle.net/2346/74367

22.

Suárez-López

J. R.

Hood

Suárez-Torres

Gahagan

Gunnar

M. R.

López-Paredes

(2019). Associations of acetylcholinesterase activity with depression and anxiety symptoms among adolescents growing up near pesticide spray sites. International Journal of Hygiene and Environmental Health, 222, 981–990. https://doi.org/10.1016/j.ijheh.2019.06.001

23.

Swift

L. E.

(2016). Teacher factors contributing to implementation of the kiVa anti-bullying program [Doctoral dissertation, University of Delaware]. https://search.proquest.com/docview/1840889366?accountid=44825

24.

Tropez-Arceneaux

L. L.

Castillo Alaniz

A. T.

Icaza

I. L.

Murillo

E. A.

(2017). The psychological impact of first burn camp in Nicaragua. Journal of Burn Care & Research, 38(1), e1–e7. https://doi.org/10.1097/bcr.0000000000000465

25.

U.S. Census Bureau. (2018). 2013-2017 American Community Survey 5-year estimates: United States (B16005 & S1601). https://data.census.gov/cedsci/

26.

U.S. Census Bureau. (2019a). Annual estimates of the resident population by sex, age, race, and Hispanic origin for the United States and states: April 1, 2010 to July 1, 2018 (Online data). https://www.census.gov/content/census/en/data/tables/time-series/demo/popest/2010s-national-detail.html

27.

U.S. Census Bureau. (2019b). Annual estimates of the resident population by sex, single year of age, race, and Hispanic origin for the United States: April 1, 2010 to July 1, 2018 (Online data). https://www.census.gov/content/census/en/data/tables/time-series/demo/popest/2010s-national-detail.html

28.

Zayas

L. H.

Aguilar-Gaxiola

Yoon

Rey

G. N.

(2015). The distress of citizen-children with detained and deported parents. Journal of Child and Family Studies, 24(11), 3213–3223. https://doi.org/10.1007/s10826-015-0124-8