1

THE WELL-BEING INDEX

The Well-Being Index (WBI) for Schools:

A Brief Measure of Adolescents’ Mental Health

Suniya S. Luthar 1,2

Ashley M. Ebbert 1.3

Nina L. Kumar 4

Prepublication version.

For published manuscript, please email NLKumar@AuthConn.com

1
Authentic Connections, Tempe, AZ
2
Columbia University’s Teachers College (Emerita)
3
Arizona State University, Department of Psychology, Tempe, AZ
4
Authentic Connections, Cambridge, MA

The authors gratefully acknowledge support provided by Authentic Connections.

Correspondence concerning this article or the Well-Being Index measure should be addressed to

Nina L. Kumar, 240 Sidney Street Unit 101, Cambridge, MA 02139.

Email: nlkumar@authconn.com
2
THE WELL-BEING INDEX
Abstract

Schools are increasingly concerned with the well-being of the whole child – likely, more

so since the COVID-19 pandemic – and goals here were to document the psychometric

properties of a brief new measure of adolescent mental health, the Well-Being Index (WBI). The

measure assesses 4 symptom areas, 2 each of internalizing and externalizing symptoms—

Depression, Anxiety, Rule-Breaking, and Substance Use—and an optional scale on Isolation at

School. A total of 2,444 students from 2 high schools completed the WBI, the Youth Self-Report

(YSR), and other related measures. Alpha coefficients showed acceptable internal consistency,

with values for the 5 WBI subscales at .83, .84, .78, .79, and .74, respectively. Both exploratory

and confirmatory factor analyses demonstrated consistent factorial validity. Correlations with

corresponding YSR subscales indicated good convergent and discriminant validity. The WBI

Substance Use and Isolation at School subscales, similarly, had high correlations with subscales

from preexisting measures. Criterion-related validity was indicated in significant correlations

between WBI subscales and conceptually related dimensions of close relationships. Also

examined was the percentage of youth falling above clinical cutoffs on both the WBI and YSR,

and findings demonstrated high concurrent validity. Collectively, results suggest the promise of

the WBI as a brief, psychometrically sound measure to assess the adjustment of adolescents,

along with perceptions of school climate that can be modified toward fostering their overall well-

being.

Keywords: adolescents, symptoms, COVID-19, schools, well-being

 Statement of Significance: This article establishes good psychometric properties of a

brief measure of common adolescent symptoms: depression, anxiety, rule-breaking, and

substance use. Its brevity allows for widespread administration in school-based

3
THE WELL-BEING INDEX
assessments of mental health; these are especially important given documented increases

in (a) overall distress levels among teenagers over the years and (b) expectations that

schools must play a major role in monitoring and promoting youth well-being.
4
THE WELL-BEING INDEX
The Well-Being Index (WBI) for Schools:

A Brief Measure of Adolescents’ Mental Health

Schools are increasingly concerned about the mental health of their students, given

reports of increases in anxiety, depression, rule-breaking behaviors, and substance use among

youth in recent years (see American Psychological Association (APA, 2018; Luthar, Kumar, &

Zillmer, 2019). However, measures used to assess adolescents’ mental health either take a

substantial amount of time to administer, or when brief, are narrow in scope. Additionally, there

is rarely consideration of students’ feelings of isolation at school: the degree to which they feel

alienated from others as opposed to connected. Our goal in this study was to assess a brief,

psychometrically sound measure that could capture overall mental health of students at the

school level, enabling comparisons across institutions as well as within them (with the latter

illuminating subgroups that might be especially vulnerable).

Schools’ Assessments of Students’ Mental Health

With the current generation of adolescents reporting higher levels of stress and

depression than those before them (APA, 2018; Twenge et al., 2019), schools today are

increasingly charged with helping ensure youths’ psychological well-being. In point of fact,

there is now a clearly articulated national emphasis on mental health literacy among teachers,

wherein they are charged with not only traditional teaching of academic subjects but also with

promoting their students’ good mental health (Lescheid, Saklofske, & Flett, 2018). Vigilance for

students’ mental health, in turn, requires regular assessments of adjustment levels in problem

areas that are common among teens, using measures that are psychometrically sound and yet

efficient to administer.
5
THE WELL-BEING INDEX
Extant Measures of Youth Mental Health

Considering existing self-reported measures of symptoms across multiple domains, some

of the best, most widely used instruments are lengthy. For example, the Youth Self-Report,

generally considered to be the gold-standard assessment tool, has a total of 112 items, measuring

multiple symptoms in both the internalizing and externalizing categories, as well as social

competence (Achenbach & Rescorla, 2001). Similarly, the Behavior Assessment System for

Children has a total of 160 items and was designed to measure the behavior and self-perceptions

of children and adolescents between four and 18 years of age (Reynolds & Kamphaus, 1992).

The Beck Youth Inventories, second edition (BYI-II), is a set of norm-referenced diagnostic

scales designed to assess children and adolescents between the ages of seven and 18 in areas of

Depression, Anxiety, Anger, Disruptive Behavior, and Self Concept. Each inventory of the BYI-

II takes 5-10 minutes to administer, but the full “combination” inventory, spanning both

internalizing and externalizing symptoms, takes between 30 minutes and one hour (Beck et al.,

2005).

Whereas these comprehensive measurements are invaluable when evaluating youth in

clinical settings (e.g., in assessments for treatment goals in psychotherapy), they tend to be less

practical for use in school-based studies on students’ well-being. The latter typically involve

assessment not just of symptoms, but also of multiple risk and protective factors that might affect

students’ adjustment levels (e.g., see Luthar & Kumar, 2018). In general, schools are reluctant to

allow more than one 45-50 minute class period to administer surveys on students’ psychological

and behavioral adjustment.

Considering, on the other hand, existing measures of symptoms that have the advantage

of being brief and thus more feasible for school-wide assessments, these tend to focus on limited
6
THE WELL-BEING INDEX
domains of youths’ maladjustment. To illustrate, the Center for Epidemiologic Studies

Depression Scale (CES-D; Radloff, 1977) is a 20-item scale designed to measure depressive

symptoms; this instrument, however, does not assess other areas that are important during

adolescence, including anxiety (APA, 2018), rule-breaking, and drugs and alcohol use (Moffitt,

2003). Similarly, there are several relatively brief measures that encompass only externalizing

problems. The Self-Report Delinquency Checklist (SRD; Huizinga & Elliot, 1986) contains 37

items assessing the seriousness of delinquent behaviors (including substance use), but no

internalizing symptoms. The widely used 53-item Youth Risk Behavior Survey (YRBS) was

designed to measure behaviors related to intentional and unintentional injury, tobacco use,

alcohol and other drug use, sexual activity, diet, and physical activity (Brener et al., 1995;

Centers for Disease Control, 1995). While the YRBS does measure substance use, it does not

have psychometrically validated subscales (with multiple, internally consistent items, as in the

YSR or BASC) that measure depression, anxiety, or rule-breaking.

Development and Validation of the Well-Being Index

In view of the literature presented, goals of this study were to test the psychometric

promise of a new measure of adolescents’ mental health – the Well-Being Index (WBI; Kumar,

2019) -- that assesses both internalizing and externalizing symptoms and yet is brief. This

measure was developed for use in school-based assessments of students’ psychosocial and

behavioral adjustment, keeping in mind the problem areas most commonly seen among

adolescents. Within the internalizing domain, these include depression and anxiety, and within

the externalizing domain, included are rule-breaking and substance use (Achenbach & Rescorla,

2001). Thus, the WBI contains five items each to assess depression, anxiety, and rule-breaking,

to be completed on a five-point Likert scale (Kumar, 2019). To measure substance use, the WBI
7
THE WELL-BEING INDEX
entails the use of five items, verbatim, from the publicly available Monitoring the Future Study

(Johnston, O’Malley, & Bachman, 2014).

With regard to procedures used to develop WBI items, the literature was reviewed to

identify symptoms that are (a) commonly reported among adolescents in distress (e.g., Hovens,

Cantwell, & Kiriakos, 1994; Roberts, Roberts, & Chan, 2007; Roberts, Roberts, & Xing, 2007),

and (b) are recurrently included in well-established measures of internalizing and externalizing

problems (Achenbach & Rescorla, 2001; Beck et al., 2005; Huizinga & Elliot, 1986; Radloff,

1977; Reynolds & Kamphaus, 1992). Examples of regularly referenced symptoms include

feelings of sadness, nervousness, and stealing, respectively, for depression, anxiety, and rule-

breaking, and drinking to intoxication for substance use. Having culled the symptoms most

commonly documented as signaling each of the four problem areas of interest for this new

measure, the top five were identified, and brief statements were composed to capture each (see

Methods for more details).

To document the psychometric properties of the WBI in this study, analyses were done

on both reliability as well as different types of validity. With regard to reliability, alpha

coefficients of internal consistency were examined for all subscale scores. These were computed

separately among boys and girls.

As a primary test of validity, subscale scores on the WBI were compared against those on

the YSR – which, as previously noted, is widely regarded as the gold standard of assessments of

adolescent symptoms. Simple correlations were examined to assess convergent and discriminant

validity, with expectations that WBI subscale scores would show strong links with the

conceptually parallel dimensions on the YSR, and lower correlations with those conceptually
8
THE WELL-BEING INDEX
distinct (e.g., WBI Depression and Anxiety would both have stronger links with YSR Anxious-

depressed, than with YSR Rule-breaking).

In a second set of analyses on validity, associations were examined between WBI (and

YSR) subscales and other aspects of adolescents’ psychosocial adjustment with which

conceptually, they should be related; the hope was that the magnitude of associations for the

subscales of the WBI would be at least comparable to (if not greater than) those of the YSR. In

these analyses, the central focus here was on dimensions of relationships, which we know are

critical in relation to exacerbating both internalizing and externalizing symptoms (Luthar &

Eisenberg, 2017). Specifically, links were examined with feelings of alienation from both parents

as well as from teachers, feelings of being bullied or victimized at school, and overall

relationship stress; each of these is conceptually linked with both internalizing and externalizing

problems (for reviews, see Flett, 2018; NASEM, 2019a; 2019b). In addition, we examined teens’

reports of discipline at school, i.e., the degree to which they felt that students who broke rules

were treated fairly, with the assumption that those high on rule breaking (or other symptoms)

would be less likely to see school rules as generally fair and appropriate.

In a third set of validity analyses, the focus was on the degree to which the WBI and YSR

each classified youth as having symptom levels surpassing the normal range, i.e., greater than 2

SD’s and 1.5 SDs from the sample mean, respectively labelled “much above average” and

“above average” on the YSR. The goal here was to identify percentage of agreement across the

two measures, as well as the percentage of WBI “false positives” and “false negatives” vis-à-vis

classifications by the YSR. The percentage of “false positives” reflected situations in which the

WBI might classify a student as falling above cutoffs on symptoms, but the YSR did not.
9
THE WELL-BEING INDEX
Conversely, the percentage of “false negatives” would reflect situations in which the WBI did

not classify a student as being above cutoffs, but the YSR did.

Isolation at School

As a further gauge of overall student well-being – apart from symptoms – the WBI

contains five items pertaining to students’ sense of alienation or rejection at school as opposed

to belonging and acceptance; information on this construct could also be valuable for school

administrators. As we now know well, resilience rests centrally on relationships (Luthar &

Eisenberg, 2019; NASEM, 2019a) and school-based relationships can strongly affect students’

well-being (e.g., Millings et al., 2014; Vaz et al., 2014). Just as the quality of relationships with

parents is critical in determining how well children negotiate the ongoing challenges of

adolescence, research has demonstrated the high the potential for relationships at school to move

adjustment trajectories toward more positive (or negative) directions (for a review, see

Domitrovich, Durlak, Staley, & Weissberg, 2017).

Another reason for school administrators to measure isolation at school would be to

evaluate the success of integrating new sets of students, especially at grade levels that represent

major transition points. For example, middle schools often bring together, in the 6th grade,

elementary school students from a number of feeder schools, leading to the formation of new

groups of friends and the possible isolation of some, as peer groups coalesce. Similarly, many

high schools include only grades 9 through 12; it can be helpful to keep track of the degree to

which incoming 9th graders feel a sense of belonging to the institution. Routine assessments at

such transitional points can also be critical for identifying any subgroups who might need

additional support integrating into the school, such as international students joining a new

school, even as they are transitioning to life in a new country.

10
THE WELL-BEING INDEX
Along with the four symptom domains, therefore, this study also entailed examination of

the psychometric properties of the five-item WBI scale on Isolation at School. As with symptom

scales, alpha coefficients were computed separately for girls and boys to document internal

consistency. To assess validity of this scale, of central interest were associations with measures

of alienation from both adults and peers at school, i.e., teacher alienation and peer victimization.

Summary

To summarize, goals in this study were to examine the reliability and validity of the WBI

as a useful potential measure of well-being of students in school-wide assessments. Specific

goals were to (1) Evaluate the reliability or internal consistency of each subscale; two on

internalizing symptoms (Depression and Anxiety), two on externalizing symptoms (Rule-

breaking and Substance Use), and a separate scale on Isolation at School; (2) Validate the

internal structure of distinct symptom subscales, i.e., empirically demonstrating the

conceptualized four-factor structure of the two internalizing and two externalizing symptoms

assessed; (3) Examine the convergent and discriminant validity of WBI symptom subscale

scores, via correlations with YSR subscales; (4) Assess criterion-related validity by examining

associations of the WBI subscales with conceptually-related, established measures; and (5)

Measure concurrent validity, that is, concordance in the percentage of youth falling above

clinical cutoffs on the WBI and on the YSR.

Method

Participants and Procedures

Data were obtained from 2,444 high school adolescents, grades 9 (n = 628), 10 (n = 678),

11 (n = 635), and 12 (n = 486), and postgraduates (n = 17). Of the sample, 49.8% were male (n =

1217) and 50.2% were female (n = 1227). Participants were from two high school samples, one
11
THE WELL-BEING INDEX
independent (which some refer to as ‘private’) and one public, respectively from the Northeast

and Southwest regions of the United States. Students described themselves as primarily

Caucasian (61.5%), with other ethnicities represented as follows: African American/Black

(4.4%), Latinx/Hispanic (15.5%), Asian/Asian American/Pacific Islander (8.2%), American

Indian/Native American (1.1%), Middle Eastern (0.6%), and Biracial/Multiracial/Other (9.8%).

Participants’ reports indicated that 72.8% of the parents were married, with about half having

graduated from college (53% of mothers and 50% of fathers) and a majority working full-time

jobs (61% of mothers and 83% of fathers).

As part of their ongoing work to foster positive youth development, school officials

administered a survey on well-being and adjustment in the Spring semester of 2019. Parents

were informed about the nature of the survey, given the option to decline their children’s

participation, and assured of the confidentiality and anonymity of the data. Adolescents

completed the questionnaires in their classrooms during regularly scheduled class time, using

computer-based surveys. Only 8% of the student body across both schools were either denied

parent consent to participate in the study or were absent on the day of data collection. A total of

2,546 students completed the survey with a rejection rate of 4% due to incomplete data, resulting

in a final sample of 2,444 adolescents. The present study was granted “exempt” status by the IRB

committee at Columbia University’s Teachers College, protocol number 20-161, as it involved

analyses of pre-existing, anonymous data.

Measures

Measures used in this study were part of a larger battery of instruments assessing

students’ well-being and the quality of their relationships with parents, peers, and adults at

school. The order of questionnaires was the same in all administrations, beginning with
12
THE WELL-BEING INDEX
demographic data, and measures with a negative valence (e.g., symptoms or difficulties in

relationships) interspersed with those of a positive valence (e.g., perceived support across

different relationships). The measures were all psychometrically sound; alpha reliability

coefficients, shown in Table 1, indicated acceptable levels of internal consistency (e.g., α ≥ .70;

Nunnally & Bernstein, 1994).

Well-Being Index (WBI)

The WBI was developed by the third author, drawing upon her background in psychology

and expertise with school-based assessments (Kumar, 2019; Luthar & Kumar, 2018; Luthar,

Kumar, & Zillmer, 2019; 2020). Three 5-item subscales were drafted to capture problem areas

common among adolescents (depression, anxiety, and rule-breaking; see below for details), plus

one subscale that would assess feelings of isolation at school. As is common for the creation of

new measures in psychology (see DeVellis, 2016), items were created by first generating

statements that paraphrased each category of symptoms to be captured, relying on existing,

relevant theoretical and empirical literatures. After refining the items to reflect specific and age-

appropriate wording, the initial pool of items was reviewed by other experts as is recommended

(DeVellis, 2016); each expert had experience in developmental psychopathology research and

clinical practice with adolescents (the first and second authors). Collectively, this team then

finalized WBI subscale items, with a goal of balancing both thoroughness and brevity in

representing distinct, yet conceptually related symptoms, and ensuring face validity.

The three newly created symptom subscales included Depression (e.g., “I am sad or

depressed”), Anxiety (e.g., “I worry or obsess”) and Rule-breaking (e.g., “I cheat on exams or

tests”). For each of these, subjects were asked to indicate the extent to which the items were true

for them within the past 6 months on a 5-point scale (0 = never, 1 = rarely, 2 = sometimes, 3 =
13
THE WELL-BEING INDEX
often, and 4 = very often). To assess the fourth mental health dimension, Substance Use,

questions from the Monitoring the Future (MTF) study (Johnston, O’Malley, & Bachman, 2014)

asked about substance use behaviors most common among youth – drinking alcohol, getting

drunk, using marijuana, smoking cigarettes, and vaping. The stem for these questions was,

“During the last 30 days, on how many occasions (if any) have you...” and sample items include

“Smoked cigarettes” and “Drank alcohol (including beer, wine, and liquor) -- more than just a

few sips”. Ratings were rescaled from a 7-point to 5-point scale (0 = never and 4 = 40 or more

times) to match the other WBI subscales. The reliability and validity of this type of self-report on

substance use have been amply documented (Johnston, Bachman, & O’Malley, 2014; Wallace &

Bachman, 1991). Finally, the supplementary Isolation at School subscale also contained five

items (e.g., “At my school, I feel isolated or like I don’t belong”). These items too were rated on

how true they were for the individual within the past 6 months, on a 5-point scale (0 = never, 1 =

rarely, 2 = sometimes, 3 = often, and 4 = very often).

Alpha coefficients of all WBI subscales are shown in Table 1. Among boys and girls

respectively, values were as follows: Depression, .82 and .84; Anxiety, .82 and .82; Rule-

breaking, .78 and .78, Substance Use, .83 and .76; and Isolation at School, .74 and .73.

(Note: When the WBI was administered within the school-based assessment battery, the five

subscales were presented in the same order that they are described here, and the order of all items

was fixed, as is true when administering the Youth Self Report described in the section that

follows.)

Measures Used to Validate the WBI Subscales

Symptoms: The Youth Self-Report (YSR)

14
THE WELL-BEING INDEX
The YSR (Achenbach & Rescorla, 2001) is a 112-item index designed to measure

internalizing and externalizing symptoms across several subscales. Participants are asked to

indicate frequencies of experiencing these within the past six months on a 3-point scale (0 =

never, 1 = sometimes, 2 = often). Although the entire YSR was included with standard

administration, in this study, we examined data on two internalizing subscales directly relevant

to WBI subscales, i.e., Anxious-depressed and Somatic Complaints, and one externalizing

subscale, Rule-breaking. The Anxious-depressed scale consists of 13 items, which describe

symptoms of feeling anxious (e.g., “I worry a lot”’) and depressed (e.g., “I feel worthless or

inferior”). The Somatic Complaints scale consists of 10 items, which describe physical

symptoms of anxiety or depression (e.g., “I feel overtired without good reason”). Finally, the

Rule-breaking subscale consists of 15 items, which describe delinquent behaviors (e.g., “I break

rules at home, school, and elsewhere”). In the present study, reliability coefficients among boys

and girls respectively were as follows: Anxious-depressed, .95 and .94; Somatic Complaints, .82

and .94; and Rule-breaking: .95 and .95 (see Table 1).

Conceptually Related Constructs

Adolescents’ feelings of alienation from mothers and fathers were assessed with the

Alienation subscale of the Short Form of the Inventory of Parent and Peer Attachment (IPPA;

Armsden & Greenberg, 1987). The IPPA Short Form consisted of 24 items with 12 pertaining to

each parent, each rated on a 5-point scale from 1 (almost never or never true) to 5 (almost

always or always true). The Alienation scale consisted of 8 items (4 for each parent) that

assessed the youth’s feelings of anger, isolation, and mistrust in relating to each parent (e.g.,

“Talking over my problems with my mother/father makes me feel ashamed or foolish,” “I feel

angry with my mother/father”), with reliability coefficients ranging from .77 to .82.
15
THE WELL-BEING INDEX
Also considered were dimensions from the School Climate and Connectedness Survey

(SCCS; American Institutes for Research, 2011), which evaluates student attitudes toward their

school and dimensions of connectedness. The following two school climate subscales were

included with three items each: Fairness of Discipline, e.g., “When students break rules, they are

treated fairly,” and Teacher Alienation, e.g., “My teacher has made me feel inadequate or

inferior to others.” For each dimension of school climate assessed, participants responded to

items on a 5-point scale from 1 (strongly disagree) to 5 (strongly agree). In the present study,

internal consistencies were well above 0.70, ranging from .79 to .94 (see Table 1).

Peer victimization was measured by the Revised Peer Experiences Questionnaire (RPEQ)

(Prinstein, Boergers, & Vernberg, 2001) which taps into overt, relational, and reputational

victimization rated on a 5-point scale from 1 (never) to 5 (a few times a week). The RPEQ has

demonstrated good validity and reliability (Prinstein et al., 2001), and in this sample, alpha

coefficients ranged from .86 to .92. Finally, students were assessed on overall stress from

relationships, responding to this prompt: “Please indicate how much stress the people in the

following relationships cause you (e.g., parents, teachers, coaches, friends in general, friends due

to competition, and significant others). For each relationship, students rated their response on a

5-point scale from 1 (not at all) to 5 (a great deal) with alpha coefficients ranging from .73

to .81.

Statistical Analyses

Analyses for this study were conducted using SPSS software (IBM SPSS Statistics,

Version 25.0), Mplus (Muthén & Muthén, Version 7.11). To test the equality of correlation

coefficients specifically, an interactive, online calculator was used (Lee & Preacher, 2013).

Results
16
THE WELL-BEING INDEX
Descriptive Data

Table 1 presents means and standard deviations on all variables in this study, separately

for boys and girls. As would be expected, girls had significantly higher levels of both WBI

internalizing symptoms, whereas boys had substantially higher externalizing problems, and

slightly higher substance use. Similar patterns were seen on YSR subscales. On all the validating

indices – negative aspects of relationships – girls reported higher levels than boys.

Internal Structure: Factorial Validity of WBI Symptom Subscales

After randomly splitting the sample in half, an Exploratory Factor Analysis (EFA) was

conducted on one half of the data to identify the underlying factor structure of items in the two

sets each of internalizing versus externalizing domains, Depression and Anxiety, versus Rule-

breaking and Substance Use. Analyses used maximum likelihood (ML), a common model-fitting

method that estimates factor loadings and unique variances, and CF-Quartimax rotation (i.e., an

oblique rotation method that uses the overall variance of the squared factor structure, based on

the assumption that the subscales are correlated). Results of the EFA model indicated the four

conceptually distinct subscales of symptoms (see Table 2); these findings provided the basis for

specifying a Confirmatory Factor Analysis (CFA) model that was fit to the other half of the data.

Evaluating Model Fit

In order to determine whether the specifications of the estimated CFA model were

consistent with the data, the goodness of fit of the model was evaluated. Using combinational

rules based on a two-index presentation strategy (Hu & Bentler, 1999), the standardized root

mean square residual (SRMSR) and the root mean square error of approximation (RMSEA;

Steiger, 1990) were considered. A model with relatively good fit would be evidenced by a
17
THE WELL-BEING INDEX
SRMR value of close to .08 (or less) and a RMSEA value of close to .06 (or less) (Hu & Bentler,

1999). RMSEA values close to .08 indicate fair fit (Browne & Cudek, 1993).

In the present study, a review of the related goodness-of-fit indexes revealed relatively

good-fitting EFA models as indicated by the following evaluative criteria for boys: SRMR = .04;

and RMSEA = .08. For girls, another good-fitting model was indicated by the following

evaluative criteria: SRMR = .04; and RMSEA = .07. Additionally, combinational rules based on

SRMR and RMSEA fit indexes revealed relatively good-fitting CFA models as indicated by the

following evaluative criteria for boys and girls, respectively: SRMR = .07 and .06; and RMSEA

= .08 and .07. CFA factor loadings are shown in Table 2.

Convergent, Discriminant, and Criterion-related Validity

Table 3 presents simple correlations across all variables in the study and shows that WBI

scores were significantly related to conceptually related YSR variables. Coefficients for the

WBI Depression in relation to YSR Anxious-depressed were .43 and .60 for boys and girls,

respectively, and those for the WBI Anxiety and YSR Anxious-depressed were .43 and .56.

Correlations between WBI Rule-breaking and the YSR Rule-breaking were .36 and .40 for boys

and girls, respectively. The WBI Isolation at School subscale was significantly related to school

climate subscales from pre-existing measures. Correlations among girls ranged from .22 to .44,

and those for boys ranged from .20 to .40 (please see Table 3 for more details).

Table 3 also shows expected associations between WBI subscales and conceptually

related constructs. WBI Depression and Anxiety subscales were significantly correlated with

alienation from mothers, with coefficients of .38 and .33, respectively, for boys, and .45 and .34

for girls. Parallel values for alienation from fathers were .36 and .33 for boys and .43 and .33 for

girls. Similarly, correlations with relationship stress were significant; for boys and girls,
18
THE WELL-BEING INDEX
respectively, values for WBI Depression were .36 and .41 and for WBI Anxiety, .34 and .37. In

relation to peer victimization, the values for the two internalizing subscales, respectively,

were .35 and .29 for boys, and .39 and .35 for girls. Finally, WBI Rule-breaking had pronounced

associations with substance use, with correlation coefficients of .42 and .41 for boys and girls,

respectively; it was also significantly related to all relationship variables.

To examine convergent and discriminant validity in greater depth, we compared the

magnitude of coefficients where the WBI subscale was conceptually (a) similar to the YSR

subscale vs. (b) dissimilar, using asymptotic z tests (Lee & Preacher, 2013; Steiger, 1980).

Results are shown in Table 4, with values involving conceptually similar pairs shown first in

boldface, followed by those involving dissimilar pairs. For example, coefficients for WBI

Depression and YSR Anxious-depressed / YSR-Rule-breaking were .43 / .28 and .60 / .31, for

boys and girls, respectively. As shown in Table 4, all six comparisons showed that the

coefficients for WBI scores in relation to similar YSR ones were significantly higher than values

involving conceptually dissimilar ones, at p < .001.

WBI versus YSR Symptom Subscales in Relation to Validating Predictors

Correlation Coefficients

Table 5 shows associations between all validating predictor variables and (a) the three

WBI symptom subscales versus (b) the parallel YSR subscales. Again, the pairs of correlation

coefficients were compared to determine if they differed in magnitude. For example, in relation

to Mom Alienation, correlations compared included WBI Depression vs. YSR Anxious-

depressed, which were .38 vs. .31 for boys, and .45 vs. .33 for girls.

Results collectively indicated that despite the relative brevity of WBI subscales, this had

not led to lower overall magnitude of correlations, relative to those of the YSR, with validating
19
THE WELL-BEING INDEX
predictors. In fact, the magnitude of associations was generally comparable and, in some

instances, larger for the WBI. As shown in Table 5, of the 36 comparisons (18 x 2, for boys and

girls), 23 of them (64%) showed that correlations for WBI subscales were significantly larger, in

the expected direction, than those for their YSR analogues. In an additional 9 cases, WBI values

were higher, but these differences were not statistically significant. There were only 4 instances

where the coefficients for WBI were lower than those for YSR, with differences statistically

significant in just one of the cases. It should be noted that in the cases involving perceived

Fairness of Discipline, which conceptually, should be inversely correlated with adolescents’

symptoms, two correlation coefficients for WBI were lower than those for YSR in absolute value

(see Footnote in Table 5). At the same time, the valence of these associations was in the

expected negative direction for WBI but not the YSR, which is why these comparisons were

counted as favorable for the former.

R2 Values in Regressions

As another simple gauge of overall validity as symptom measures, regression analyses

were conducted using all six of the conceptually related validating variables as predictors, to

ascertain the total variance explained in the symptom subscale scores. Results of these

regressions, again, yielded no evidence that any of the brief WBI subscales had lower magnitude

of associations with predictors collectively as compared to the longer YSR subscales; in fact,

variance accounted for was greater in predicting to WBI subscales. Specifically, total R 2 values

were as follows for pairs of outcome variables from the two measures: WBI Depression versus

YSR Anxious-depressed, R2 .25 and .21 (boys), .36 and .21 (girls); WBI Anxiety versus YSR

Anxious-depressed, R2 .23 and .21 (boys), .25 and .21 (girls); WBI Rule-breaking versus YSR-

Rule-breaking, .22 and .13 (boys), .23 and .13 (girls).

20
THE WELL-BEING INDEX
To ascertain whether the magnitudes of variance explained (R2) were significantly

different when outcome variables were based on the WBI versus the YSR measures, a repeated

measures model (GLM procedure) was used (see Wheeler, 2017). For example, the first pair of

R2 values compared, with WBI Depression versus YSR Anxious-depressed as outcomes, was .25

vs. .21 for boys. Across all six pairs of R2 coefficient delineated in the previous paragraph,

values were significantly higher for outcomes involving the WBI, at p < .001.

Concurrent Validity: Classification of Youth Above Clinical Cutoffs

The percentage of students falling above clinical cutoffs on the WBI and on the YSR

each were examined, using gender-specific norms for each; results are presented in Table 6.

When the WBI Depression classifications were compared with those on the YSR Anxious-

depressed scale, instances where both measures classified students in the clinically significant

range (over 2 SD’s from the mean), were 69% and 72% for boys and girls respectively. Parallel

values in classifying students in the range of borderline significance (above 1.5 SD’s from the

mean) were 67% and 65% for boys and girls respectively. Similarly, when WBI Anxiety and

YSR Anxious-depressed classifications were compared, agreement above 2 SDs was 70% for

both genders, while agreement on being above 1.5 SDs was 66% and 63% for boys and girls

respectively. For Rule-breaking, the four values for matching classifications were 70% and 74%

versus 68% and 73%.

Disagreements in classification occurred almost entirely because the YSR identified more

students as being above clinical cutoffs than did the WBI, rather than the reverse. In classifying

both clinically significant and borderline scores, between 23% and 36% of instances involved a

negative classification by the WBI but a positive one by the YSR. By contrast, “false positives”

based on the WBI (not indicated by the YSR) were very rare, ranging from 1-4%.
21
THE WELL-BEING INDEX
Discussion

The present results \provide evidence for the promise of a brief measure for school-based

assessments of youth mental health, the Well-Being Index (WBI; Kumar, 2019). The measure is

applicable to teens from various sociodemographic backgrounds, as it captures major

internalizing and externalizing symptoms that generally occur during the adolescent years.

Examination of psychometric properties showed good internal consistency of all scales as well as

high levels of validity. From a pragmatic standpoint, given an administration time of

approximately ten minutes, the WBI can be administered even on a monthly basis (e.g., when a

school may have experienced a traumatic event, or to periodically track the value of a given

intervention program for targeted subgroups). At a more macro level, annual administration of

this short measure would allow administrators to keep track of the overall well-being of their

students, much as they keep track of scores on standardized tests or passing percentages on

achievement tests. In discussions that follow, major findings on the WBI’s psychometric

properties are summarized first, followed by an appraisal of its applications in school-based

assessments tied to prevention efforts.

WBI as a Measure of Symptoms: Psychometric Properties

Results of this study indicated good psychometric properties of the WBI as a measure of

adolescents’ internalizing and externalizing symptoms. With regard to reliability, internal

consistency of the four symptom subscales – Depression, Anxiety, Rule-breaking and Substance

Use – was acceptable among both boys and girls. The eight values ranged from .72 to .84, with a

median of .82; these values fall well above the range that is considered satisfactory (α ≥ .70;

Nunnally & Bernstein, 1994).

22
THE WELL-BEING INDEX
In terms of the distinctiveness of the symptom areas measured, factor analyses supported

a four-factor model. As they had been conceptualized, the 20 items did map on the two

internalizing subscales of Depression and Anxiety and two externalizing subscales of Rule-

breaking and Substance Use. An evaluation of the goodness of fit model revealed an

exceptionally well-fitting model.

WBI subscales also showed strong levels of convergent and discriminant validity. With

regard to the former, one would expect stronger links between WBI subscales of internalizing

symptoms with YSR scores also of internalizing problems, as opposed to YSR externalizing

scores. In fact, correlations for both WBI Depression and WBI Anxiety with the YSR Anxious-

depressed subscale were higher than coefficients of each of these with YSR Rule-breaking (1.5 –

2 times as high). All comparisons were statistically significant.

Similarly, when considering the magnitude of correlations with conceptually linked

constructs – aspects of relationships with parents, peers, and teachers – WBI subscales’

correlations were clearly at least equivalent to those for YSR subscales, if not slightly higher, on

the whole. Specifically, of 36 comparisons conducted, differences were significantly higher for

the WBI in 64% of cases, and for the YSR in less than 3%. Furthermore, when all sets of

predictors were entered into regression analyses, the total variance explained was significantly

greater when outcomes were symptoms measured by the WBI, as opposed to based on the YSR.

In sum, despite the brevity of WBI subscales relative to YSR (five items each as opposed to 10-

15 items), findings of this study did not suggest any relative deficits in validity of measurement.

There was also generally good agreement in classifying youth as showing clinically

significant levels of symptoms, as measured by the WBI versus the YSR; this was true even

though descriptors of symptoms were similar but not identical. In these analyses, classification
23
THE WELL-BEING INDEX
based on the WBI Depression and Anxiety subscales were each compared against classification

via the single YSR subscale that subsumed both these dimensions, Anxious-depressed. Similarly,

the WBI separately measures Rule-breaking and Substance Use whereas the YSR Rule-breaking

subscale includes acts related to conduct disturbances as well as to substance use. These caveats

notwithstanding, levels of agreement in classifying youth in the clinically significant range

(“Much above average” or 2 SDs from the mean) ranged from 69% to 74%, with a median of

70%.

Discrepancies in classifying students above clinical cutoffs almost all occurred because

of “false negative”, wherein analyses using the YSR classified more students at the extremes

than did analyses using the WBI. One potential explanation for the difference in classification

might be that levels of clinically significant symptoms have shifted over the last couple of

decades in the United States. Norms on the YSR are based in assessments done around the year

2000 (Achenbach & Rescorla, 2001), before the creation of social media and tragedies such as

9/11, and multiple reports have shown overall increases in adolescent distress over time (APA,

2018; NASEM, 2019; Twenge et al., 2019). To the degree that teens as a group are more

troubled overall, this would imply a higher overall mean on any given scale and

commensurately, smaller numbers of youth falling under the “extremes” of +2 SDs. To illustrate,

if the population mean of adolescents’ symptoms on a given measure increased from 50 to 60

over the decades with SDs of 10 in both cases, individuals with scores of 79 would be below

clinical cutoffs using 2019 norms on this measure, but well above clinical cutoffs using norms

from the past. This suggestion is in fact supported by findings, in this study, of mean T scores all

greater than 60 among both boys and girls and across all three YSR subscales assessed. (In

essence, this implies that a standardized score of 75 would connote approximately +1.5 SD’s on
24
THE WELL-BEING INDEX
the WBI (this sample’s means of about 60 and SDs of 15), but equivalent to a YSR T score of +2

SD’s, given YSR’s mean T score of 50 and SD of 10).

An additional possibility that must be considered, in relation to inferences about false

negative classifications on the WBI versus YSR, is that there were differences in samples that

were used in defining normative values on the measures. The original YSR norms were based

on non-referred youth (Achenbach & Rescorla, 2001). By contrast, norms created for the WBI

were based on all students at the schools sampled, which presumably included some children

who had been referred for professional help.

In future research, administration of the WBI to additional samples of students will be

critical in indicating values that truly approximate the current “population” means and standard

deviations. By the end of 2019, WBI data had been obtained from almost 15,000 youth across

the United States (these additional samples were not assessed on the YSR, however, so are not

discussed in the present paper). Examination of the larger samples will also permit more fine-

grained analyses of clinical cutoffs separately by gender, grade level, and ethnicity. However,

before moving to any such large scale administration of the WBI, a critical prerequisite has to be

work such as that reported on in this study: Establishing good psychometric properties of this as

a stand-alone instrument, validated against the widely used, gold standard measure (the YSR),

and documenting strong relationships with other conceptually linked measures.

Applying the WBI Toward Prevention of Youth Psychopathology

Just as nations have a simple “happiness index,” we believe that it is critical that schools

have a simple well-being index of their overall student body. As noted earlier, adolescents are

increasingly experiencing mental health difficulties (APA, 2018). Today’s youth who are in the

range of 15-21 years, called Gen Z, are significantly more likely than prior generations to report
25
THE WELL-BEING INDEX
that their mental health is fair or poor, with 27 percent indicating this is true for them. By

contrast, ratings of fair or poor mental health were given by 15 percent of Millennials and 13

percent of Gen Xers, and only seven and five percent, respectively, of Boomers and older adults.

From a prevention standpoint, these data indicate that it would be useful to schools to be

able to regularly track where their student bodies stand on psychosocial adjustment, just as they

do with standardized test scores. In addition, administering the WBI measure could allow

administrators to see which symptom areas might need enhanced attention within their own

schools, as impressionistic evidence is not always borne out by the data. High school teachers are

often more aware of externalizing, acting-out behaviors, for example, than of students’

depression or anxiety; the latter problems are more covert in nature and can go undetected even

when at seriously high levels (Flett, Hewitt, Nepon, & Zaki-Azat, 2018; Luthar et al., 2020).

Additionally, the WBI can help identify subgroups of youth who are particularly

struggling in a given school. Students with symptoms falling in the troubling “red zone” can vary

across institutions, for example, by grade level (with freshmen standing out in some cases or in

others, juniors immersed in college applications), as well as across different demographic

subgroups (e.g., with distress elevated among particular sexual or ethnic minorities or

international students). In turn, identifying vulnerable subgroups can help target mental health

resources to where they are most needed. This can be particularly helpful in low-income schools,

where resources for mental health in are typically scant (Hoagwood et al. 2018), necessitating

efficient, targeted programs rather than “universal” interventions for all students.

There can also be much value in administering the WBI routinely in relatively well-

resourced schools, characterized by high levels of achievement; in fact, these students report

symptom levels commensurate with if not greater than those of their counterparts in poverty
26
THE WELL-BEING INDEX
(Luthar et al., 2019; NASEM, 2019). From the perspective of prospective parents, it could be

valuable to have a gauge of whether schools truly are invested in fostering the development of

the “whole child” (Luthar et al., 2019; Wilson & Marshall, 2019). In all likelihood, many

parents would appreciate having information on the students’ overall mental health across critical

domains of adolescent adjustment, in addition to stellar records on academics and extracurricular

activities.

Besides measuring adolescents’ symptoms in critical mental health domains, another

potentially useful application of the WBI lies the use of the scale capturing feelings of isolation

at school. Decades of research on resilience have shown that relationships are fundamental to

doing well in despite stress (Luthar, Crossman, & Small, 2015; NASEM, 2019a). Additionally,

negative relationship indices tend to be more powerful than positive ones in affecting one’s well-

being (i.e., “bad is stronger than good”; Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001).

Thus, feeling alienated by adults and peers at school, or simply dreading being at school, can

make a substantial difference in affecting students overall, including their academic performance

(Lescheid et al., 2018). This said, to our knowledge, there exists no brief measure capturing these

sentiments; the WBI subscale examined and validated here could help fill this gap.

Finally, there are three issues worth noting with regard to the use of the moniker “Well-

Being Index” for this measure that in fact encompasses adjustment problems. First, in

developmental research that is described as covering children’s “well-being” or “health”,

outcomes in fact commonly include symptoms such as depression, anxiety, and substance use, as

well as dimensions of social exclusion (Best, Manktelow, & Taylor, 2014; Patton et al., 2016;

Sieving et al., 2001; Vanassche, Sodermans, Matthijs, & Swicegood, 2013). Second, in

contemporary school-based surveys of students’ psychological or behavioral difficulties,

27
THE WELL-BEING INDEX
instruments often reference health rather than illness, e.g., the Global School-based Student

Health Survey (Center for Disease and Prevention, 2020) and the Independent School Health

Check (2020). Third, in the world of practice within K-12 education, enhancing overall “well-

being” of students is typically viewed as implying low levels of symptoms. To illustrate, Wilson

and Marshall (2019) exhort schools to appraise how they “conceptualize and adopt a definition

of health and wellness… and how they use it to inform their approaches to very real problems—

student suicide, depression, anxiety, self-harm, peer-to-peer sexual assault, self-medication/drug

and alcohol abuse.”

Limitations and Future Directions

Among the limitations of this study, the first is that the present sample included only high

school students (grades 9 to Postgrad), although the assessment tool is intended for use with

youth from grade 6 through grade 12. In the future, the WBI should be tested with middle

schoolers as well, aged 11-14 years. Second, data were only collected with students in two areas

of the country, Northeast and Southwest; the measure should be validated in different parts of the

country and with demographically heterogenous samples. With regard to psychometric

properties, future studies should also consider test-retest coefficients as indices of reliability, in

addition to levels of internal consistency.

There are several limitations associated with the sole reliance on one informant in this

study; scores on both the WBI and YSR were based in self-reports (as are many adolescent

symptom scales described at the outset of this article). The mono-method, mono-informant

approach could have led to under-reporting on some scales more so than others, with lower

acknowledgement of externalizing behaviors, for example, than internalizing symptoms.

Additionally, it could have led to some inflations in associations among measures. With regard
28
THE WELL-BEING INDEX
to the latter issue, it should be reiterated that in fact, of central interest in validating WBI

subscales are links between students’ subjective experiences of their own distress and the quality

of salient relationships in their lives, also as perceived by the teens themselves. This said, future

studies might usefully examine the degree to which assessments based on the WBI converge

with or differ from those based on reports from other informants, such as parents or teachers

(these adults’ ratings were not available here as the schools administered surveys anonymously,

allowing no identifiers for individuals’ data).

Summary and Conclusions

In conclusion, data presented here suggest the validity of the WBI as a brief, scientifically

sound measure to assess the adjustment of adolescents. Its use in school-based assessments can

help fill a critical need: Early and expedient detection of mental health issues, so that

interventions can be targeted to students who most need them, and in areas within which they are

most vulnerable (Leschied et al., 2018). The central role of mental health promotion in schools,

in fact, is now explicitly emphasized by researchers as well as practitioners. In their introduction

to the Handbook of School-Based Mental Health Promotion, the editors emphasized,

“(The) challenges and problems on the mental health front have become urgent enough

that . . . a focus on mental health promotion in children and adolescents must become part

of the regular school day, and this is just as important as the more traditional educational

learning that takes place in our schools. (Leschied et al., 2018, p. 1; see also Offner,

2018).

Similarly, in their report on the foundations of student success, Wilson and Marshall (2019)

noted that whereas schools’ central focus is on academics, students’ “current and long-term

mental, emotional, and physical well-being strongly contribute to their ultimate success in life”.
29
THE WELL-BEING INDEX
Thus, there is a critical need to carefully measure critical mental health indices and track these

over time (needless to say, such ongoing assessments could be still more useful in wake of

pandemic-related disruptions across schools). It is our hope that given the documented

psychometric properties and ease of administration given its brevity, the WBI can be applied in

future school-based assessments toward maximizing the wellness of a generation of highly

stressed youth.
30
THE WELL-BEING INDEX

References

Achenbach, T. M., & Rescorla, L. A. (2001). Manual for the ASEBA School-Age Forms &

Profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth, &

Families.

American Psychological Association (2018). Stress in America: Generation Z. Retrieved from

https://www.apa.org/news/press/releases/stress/2018/stress-gen-z.pdf

American Institutes for Research. (2011). 2011 School Climate and Connectedness Survey:

Statewide report. Available online at http://alaskaice.org/school-climate/survey.

Armsden, G. C., & Greenberg, M. T. (1987). The inventory of parent and peer attachment:

Individual differences and their relationship to psychological well-being in

adolescence. Journal of Youth and Adolescence, 16(5), 427-454.

Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than

good. Review of General Psychology, 5, 323–370. doi: 10.1037//1089-2680.5.4.323

Beck, J. S., Beck, A. T., Jolly, J. B., & Steer, R. A. (2005). Beck Youth Inventories Second

Edition for Children and Adolescents manual. San Antonio, TX: PsychCorp

Best, P., Manktelow, R., & Taylor, B. (2014). Online communication, social media and

adolescent wellbeing: A systematic narrative review. Children and Youth Services

Review, 41, 27-36.

doi:http://dx.doi.org.ezproxy1.lib.asu.edu/10.1016/j.childyouth.2014.03.001

Brener, N. D., Collins, J. L., Kann, L., Warren, C. W., & Williams, B. I. (1995). Reliability of

the youth risk behavior survey questionnaire. American Journal of Epidemiology, 141(6),

575-580. https://doi.org/10.1016/S1054-139X(02)00339-7
31
THE WELL-BEING INDEX
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen

& J. S. Long (Eds.), Testing structural equation models (pp. 136–162). Newbury Park,

CA: Sage.

Cederström, C., & Spicer, A. (2015). The Wellness Syndrome. John Wiley & Sons.

Centers for Disease Control and Prevention. (1995). Youth Risk Behavior Survey Data.

Retrieved from: www.cdc.gov/yrbs.

Center for Disease and Prevention (2020). Global School-based Student Health Survey (GSHS).

Retrieved from https://www.cdc.gov/gshs/questionnaire/index.htm

DeVellis, R. F. (2016). Scale development: Theory and applications (Vol. 26). Sage publications.

Domitrovich, C. E., Durlak, J. A., Staley, K. C., & Weissberg, R. P. (2017). Social‐emotional

competence: An essential factor for promoting positive adjustment and reducing risk in

school children. Child Development, 88(2), 408-416. https://doi.org/10.1111/cdev.12739

Ebbert, A.E., Kumar, N.L., & Luthar, S.S. (2019). Complexities in adjustment patterns among

the "best and the brightest": Risk and resilience in the context of high-achieving

schools. Research in Human Development, 16, 21-34.

https://doi.org/10.1080/15427609.2018.1541376

Flett, G. L. (2018). Resilience to interpersonal stress: Why mattering matters when building the

foundation of mentally healthy schools. In A. W. Leschied, D. H. Safloske, G. L. Flett

(Eds.), Handbook of school-based mental health promotion (pp. 383-410). Springer, Cham.

Flett, G. L., Hewitt, P. L., Nepon, T., & Zaki-Azat, J. N. (2018). Children and adolescents flying

under the radar: Understanding, assessing, and addressing hidden distress among

students. In A. W. Leschied, D. H. Safloske, G. L. Flett (Eds.), Handbook of school-

32
THE WELL-BEING INDEX
based mental health promotion (pp. 357-381): Springer International Publishing, Cham,

Switzerland.

Hoagwood, K. E., Atkins, M., Kelleher, K., Peth-Pierce, R., Olin, S., Burns, B., ... & Horwitz, S.

M. (2018). Trends in children’s mental health services research funding by the National

Institute of Mental Health from 2005 to 2015: A 42% reduction. Journal of the American

Academy of Child & Adolescent Psychiatry, 57(1), 10-13.

https://doi.org/10.1016/j.jaac.2017.09.433

Hovens, J. G., Cantwell, D. P., & Kiriakos, R. (1994). Psychiatric comorbidity in hospitalized

adolescent substance abusers. Journal of the American Academy of Child & Adolescent

Psychiatry, 33(4), 476-483.

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure

analyses: Conventional criteria versus new alternatives. Structural Equation Modeling: A

Multidisciplinary Journal, 6(1), 1-55. https://doi.org/10.1080/10705519909540118

Huizinga, D., & Elliot, D. S. (1986). Reassessing the reliability and validity of self-report

delinquency measures. Journal of Quantitative Criminology, 2, 293-327.

https://doi.org/10.1007/BF01064258

Independent School Health Check (2020). The Independent School Health Check

Administration Manual, 2019 – 2020, Revised July, 2019. Retrieved from

https://independentschoolhealth.com/services/

Johnston, L. D., O'Malley, P. M., Bachman, J. G., Schulenberg, J. E., & Miech, R.

A. (2014). Monitoring the Future national survey results on drug use, 1975–2013: Vol.

II. College students and adults ages 19–55. Ann Arbor, MI: University of Michigan,

Institute for Social Research.

33
THE WELL-BEING INDEX
Kumar, N. L. (2019). The Well-Being Index: A measure of youth mental health. Unpublished

document. Cambridge, MA: Authentic Connections.

Kutcher, S., Wei, Y., & Hashish, M. (2018). Schools and Mental Health: Is Some Necessary

Reexamining in Order?. In A. W. Leschied, D. H. Safloske, G. L. Flett (Eds.), Handbook

of school-based mental health promotion (pp. 83-101). Springer, Cham.

Lee, I.A., & Preacher, K.J. (2013, September). Calculation for the test of the difference between

two dependent correlations with one variable in common [Computer software].

Available from http://quantpsy.org.

Leschied, A. W., Saklofske, D. H., & Flett, G. L. (2018). Handbook of School-Based Mental

Health Promotion: An Evidence-Informed Framework for Implementation. Springer.

Luthar, S. S., Crossman, E. J., & Small, P. J. (2015). Resilience and adversity. In R. M. Lerner

and M. E. Lamb (Eds.). Handbook of Child Psychology and Developmental Science (7th

Edition, Vol. III, pp. 247-286). New York: Wiley.

Luthar, S. S., & Eisenberg, N. (2017). Resilient adaptation among at-risk children: Harnessing

science toward maximizing salutary environments. Child Development, 88, 337–349.

doi:10.1111/cdev.1273m

Luthar, S.S., & Kumar, N.L. (2018). Youth in high-achieving schools: Challenges to mental

health and directions for evidence-based interventions. In A. W. Leschied, D. H.

Saklofske, and G. L. Flett, Handbook of School-Based Mental Health Promotion: An

Evidence-Informed Framework (pp. 441-458). New York: Springer.

Luthar, S. S., Kumar, N. L., & Zillmer, N. (2019). High Achieving Schools connote

significant risks for adolescents: Problems documented, processes implicated, and

directions for interventions. American Psychologist. https://doi.org/10.1037/amp0000556

34
THE WELL-BEING INDEX
Luthar, S. S., Kumar, N. L., & Zillmer, N. (2020). Teachers’ responsibilities for students’ mental

health: Challenges in high achieving schools. International Journal of School &

Educational Psychology, 8, 119–130. https://doi.org/10.1080/21683603.2019.1694112

Millings, A., Buck, R., Montgomery, A., Spears, M., & Stallard, P. (2012). School

connectedness, peer attachment, and self-esteem as predictors of adolescent depression.

Journal of Adolescence, 35(4), 1061-1067.

https://doi.org/10.1016/j.adolescence.2012.02.015

Moffitt, T. E. (2003). Life-course-persistent and adolescence-limited antisocial behavior: A 10-

year research review and a research agenda. In B. B. Lahey, T. E. Moffitt, & A. Caspi

(Eds.), Causes of conduct disorder and juvenile delinquency (pp. 49-75). New York, NY,

US: The Guilford Press.

Muthén, L. K., & Muthén, B. O. (2013). Mplus: Statistical analyses with latent variables. User’s

Guide (Version 7.11). Los Angeles, CA: Muthén & Muthén.

National Academies of Science, Engineering, and Medicine (2019a). Vibrant and Healthy Kids:

Aligning Science, Practice, and Policy to Advance Health Equity. Washington, DC: The

National Academies Press. https://doi.org/10.17226/25466

National Academies of Science, Engineering, and Medicine (2019b). The Promise of

Adolescence: Realizing Opportunity for All Youth. Washington, DC: The National

Academies Press.

Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw

Hill.
35
THE WELL-BEING INDEX
Offner, D. (2018). Mental health in the classroom: educators as the 'first responders'. Retrieved

from www.nais.org/learn/independent-ideas/april-2018/mental-health-in-the-classroom-

educators-as-the-first-responders

Patton, G. C., Sawyer, S. M., Santelli, J. S., Ross, D. A., Afifi, R., Allen, N. B., ... & Kakuma, R.

(2016). Our future: a Lancet commission on adolescent health and wellbeing. The

Lancet, 387(10036), 2423-2478. https://doi.org/10.1016/j.jadohealth.2016.08.002

Prinstein, M. J., Boergers, J., & Vernberg, E. M. (2001). Overt and relational aggression in

adolescents: Social-psychological adjustment of aggressors and victims. Journal of

Clinical Child Psychology, 30(4), 479-491.

https://doi.org/10.1207/S15374424JCCP3004_05

Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general

population. Applied Psychological Measurement, 1(3), 385-401.

https://doi.org/10.1177/014662167700100306

Reynolds, C. R. and Kamphaus, R. W. (1992) Behaviour Assessment System for Children

Manual. Circle Pines, MN: American Guidance Service.

Roberts, R. E., Roberts, C. R., & Chan, W. (2009). One‐year incidence of psychiatric disorders

and associated risk factors among adolescents in the community. Journal of Child

Psychology and Psychiatry, 50(4), 405-415.

https://doi.org/10.1111/j.1469-7610.2008.01969.x

Roberts, R. E., Roberts, C. R., & Xing, Y. (2007). Rates of DSM-IV psychiatric disorders among

adolescents in a large metropolitan area. Journal of Psychiatric Research, 41(11), 959-

967. https://doi.org/10.1016/j.jpsychires.2006.09.006

Ross, S. M., Morrison, G. R., & Lowther, D. L. (2010). Educational technology research past
36
THE WELL-BEING INDEX
and present: Balancing rigor and relevance to impact school learning. Contemporary

Educational Technology, 1(1), 17-35.

Sieving, R. E., Beurhing, T., Resnick, M. D., Bearinger, L. H., Shew, M., Ireland, M., & Blum,

R. W. (2001). Development of adolescent self-report measures from the national

longitudinal study of adolescent health. Journal of Adolescent Health, 28(1), 73-81.

doi:http://dx.doi.org.ezproxy1.lib.asu.edu/10.1016/S1054-139X(00)00155-5

Steiger, J. H. (1980). Tests for comparing elements of a correlation matrix. Psychological

Bulletin, 87, 245-251. https://doi.org/10.1037/0033-2909.87

Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimation

approach. Multivariate Behavioral Research, 25, 173–180.

https://doi.org/10.1207/s15327906mbr2502_4

Twenge, J. M., Cooper, A. B., Joiner, T. E., Duffy, M. E., & Binau, S. G. (2019). Age, period,

and cohort trends in mood disorder indicators and suicide-related outcomes in a

nationally representative dataset, 2005-2017. Journal of Abnormal Psychology, 128, 185-

199. https://doi.org/10.1037/abn0000410.

Vanassche, S., Sodermans, A. K., Matthijs, K., & Swicegood, G. (2013). Commuting between

two parental households: The association between joint physical custody and adolescent

wellbeing following divorce. Journal of Family Studies, 19(2), 139-158.

Vaz, S., Falkmer, M., Parsons, R., Passmore, A. E., Parkin, T., & Falkmer, T. (2014). School

belongingness and mental health functioning across the primary-secondary transition in a

mainstream sample: Multi-group cross-lagged analyses. PloS one, 9(6), e99576.

https://doi.org/10.1371/journal.pone.0099576

Wallace Jr, J. M., & Bachman, J. G. (1991). Explaining racial/ethnic differences in adolescent
 37
THE WELL-BEING INDEX
drug use: The impact of background and lifestyle. Social Problems, 38(3), 333-357.

https://doi.org/10.2307/800603

Wheeler, A. P. (2017, June 12). Testing the equality of coefficients same independent different

dependent variables. Wordpress. https://andrewpwheeler.com/2017/06/12/testing-the-

equality-of-coefficients-same-independent-different-dependent-variables/

Wilson, D. P., & Marshall, M. G. (2019). Reframing the foundation for student success.

Retrieved from

https://www.nais.org/magazine/independent-school/summer-2019/educating-the-whole-

student/

Table 1
Descriptive Statistics and Psychometric Properties of Study Variables
Boys Girls Boys Girls
Eta
WBI Subscales   Mean SD Mean SD F gender
sq
Depression .82 .84 5.60 4.33 7.50 4.48 113.75*** .05
 38
THE WELL-BEING INDEX

Anxiety .82 .82 5.85 4.33 8.95 4.71 288.23*** .11

Rule-breaking .78 .78 4.04 3.66 3.35 3.30 24.17*** .01
Substance Use .83 .76 1.56 3.29 1.30 2.63 4.47* 0
Isolation at School .74 .73 3.98 4.01 5.39 4.19 72.29*** .03
YSR Subscales a   Mean SD Mean SD
Anxious-depressed Raw score .95 .94 8.06 7.97 10.77 8.26 66.79*** .03
Anxious-depressed T score - - 63.43 14.91 64.72 15.28
Somatic Complaints Raw score .94 .92 5.69 5.80 7.32 5.98 45.76*** .02
Somatic Complaints T score - - 63.32 14.40 63.55 14.08
Rule-breaking Raw score .95 .95 8.85 8.11 8.03 7.57 6.49* 0
Rule-breaking T score - - 62.96 13.79 61.66 12.51
Validating Variables
Mom Alienation .77 .77 8.21 3.62 8.76 3.74 12.82*** .01
Dad Alienation .80 .82 8.33 3.84 8.90 4.07 11.65** .01
School – Fairness of Discipline .86 .79 2.87 1.04 3.02 0.87 12.81*** .01
School – Teacher Alienation .92 .92 1.83 1.02 1.97 1.03 10.87** .01
Peer Victimization .92 .86 4.50 2.17 4.59 1.72 1.29 0
Relationship Stress .81 .73 1.91 0.84 2.14 0.79 43.44*** .02
Note. n = 1217 for boys and 1227 for girls.
a
For the YSR, values are reported for both raw and T scores (boldface); the latter indicate that this 2019 sample
had higher mean T scores and SDs (T > 60; SDs > 12) than in 2001 norms (T = 50; SD = 10).
 39
THE WELL-BEING INDEX
Table 2
Exploratory and Confirmatory Factor Analyses Using Maximum Likelihood Estimation and CF-Quartimax Rotation
EFA CFA
Factor 1 Factor 2 Factor 3 Factor 4
WBI Subscales Depression Anxiety Rule-breaking Substance Use Factor Loadings
Boys Girls Boys Girls Boys Girls Boys Girls Boys Girls
Low life enjoyment .69 .74 .08 -.02 -.01 .06 0 -.05 .73 .75
Low self-worth .43 .59 .37 .24 -.01 .05 0 -.08 .75 .70
Suicidal ideation .79 .81 -.09 -.13 .07 .01 .01 .07 .69 .70
Sad .71 .72 .18 .19 -.02 -.05 .04 .05 .86 .85
Tired .39 .28 .25 .28 .06 .16 .03 -.03 .56 .61
Anxious .03 .03 .86 .85 -.03 -.07 .03 .07 .83 .85
Nervous .04 -.02 .80 .82 .02 .04 -.06 -.05 .81 .82
Worry .01 .05 .68 .63 .11 .12 .02 -.02 .70 .72
Headaches .24 .14 .27 .28 .05 .14 .09 .04 .44 .48
Nausea .23 .13 .36 .45 .18 .11 .05 .06 .64 .65
Breaks school rules -.09 -.01 .06 -.01 .72 .62 .06 .08 .65 .72
Breaks parents’ rules .00 .02 .00 -.02 .74 .72 .09 .06 .81 .82
Cheats .08 -.05 -.11 -.01 .52 .60 -.05 -.03 .59 .55
Lies -.01 .04 .08 .09 .71 .63 -.06 -.08 .64 .70
Steals .18 .07 -.08 0 .56 .45 -.07 .11 .51 .54
Cigarettes .07 -.04 .02 .11 .06 -.02 .29 .38 .47 .42
Alcohol -.02 -.06 .02 0 .02 .11 .80 .74 .80 .83
Drunk -.05 0 .03 .03 -.02 -.07 .91 .90 .87 .89
Marijuana .14 .10 -.07 -.05 .02 .05 .63 .62 .64 .65
Vaping .20 .05 -.13 0 .13 .10 .47 .58 .66 .59
Note. n = 604 males and 617 females for EFA Sample; n = 623 males and 610 females for CFA Sample.
Shading indicates items loading onto the designated factor.
 40
THE WELL-BEING INDEX

Table 3

Correlations Among Well-Being Index, YSR, and Validating variables, and Results of Regressions with Symptoms Predicted by Validators

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14

1. WBI Depression - .73** .41** .21** .67** .60** .51** .31** .45** .43** -.21** .28** .39** .41**

2. WBI Anxiety .71** - .33** .17** .64** .56** .54** .24** .34** .33** -.15** .30** .35** .37**

3. WBI Rule-breaking .42** .37** - .41** .37** .24** .26** .40** .38** .31** -.23** .21** .27** .32**

4. WBI Substance Use .24** .15** .42** - .20** .11** .19** .37** .22** .20** -.15** .10** .27** .20**

5. WBI Isolation at School .66** .59** .42** .10** - .52** .41** .26** .34** .31** -.22** .39** .44** .41**

6. YSR Anxious-depressed .43** .43** .14** .20** .38** - .82** .75** .33** .28** 0 .16** .28** .38**

7. YSR Somatic .34** .38** .16** .24** .26** .86** - .81** .27** .23** -.03 .16** .29** .37**

8. YSR Rule-breaking .28** .23** .36** .44** .23** .80** .76** - .25** .20** -.02 .06* .24** .31**

9. Mom Alienation .38** .33** .27** .18** .34** .31** .26** .25** - .44** -.14** .22** .27** .38**

10. Dad Alienation .36** .33** .32** .23** .35** .28** .25** .26** .51** - -.23** .21** .27** .35**

11. School–Fairness Discipline -.10** -.01 -.25** -.19** -.20** .06* .04 -.03 -.12** -.12** - -.25** -.19** -.17**

12. School–Teacher Alienation .28** .25** .28** .23** .35** .16** .17** .17** .22** .26** -.15** - .39** .34**

13. Peer Victimization .35** .29** .29** .27** .40** .31** .27** .26** .30** .27** -.12** .52** - .41**

14. Relationship Stress .36** .34** .22** .21** .37** .35** .29** .29** .32** .34** -.05 .37** .50** -

Note. Correlation coefficients for girls are in the top right of diagonal; those for boys are in the bottom left of the diagonal.

Dark and light shaded cells indicate WBI and YSR subscales, respectively, correlated with conceptually related constructs.
41
THE WELL-BEING INDEX
Table 4

Comparisons of Correlation Coefficients: WBI Subscales in Relation to Conceptually Similar/ Not Similar YSR Subscales

WBI Subscales WBI Subscale / YSR Subscale Conceptually similar / Not similar subscales

Boys r Girls r

Depression Anxious-depressed / Rule-breaking .43 / .28 *** .60 / .31 ***

Anxiety Anxious-depressed / Rule-breaking .43 / .23 *** .56 / .24 ***

Rule-breaking Rule-breaking / Anxious-depressed .36 / .14 *** .40 / .24 ***

Note. ***p < .001.

42
THE WELL-BEING INDEX

Table 5
Comparisons of Correlation Coefficients: Predictors in Relation to Conceptually Similar WBI / YSR Subscales

Predictors Correlations between WBI / YSR WBI / YSR

WBI subscale YSR subscale Boys r Girls r

Mom Alienation Depression / Anxious-depressed .38 / .31 ** .45 / .33 ***

Dad Alienation Depression / Anxious-depressed .36 / .28 ** .43 / .28 ***

Teacher Alienation Depression / Anxious-depressed .28 / .16 *** .28 / .16 ***

Fairness Discipline Depression / Anxious-depressed -.10a / .06 *** -.21a / .00 ***

Peer Victimization Depression / Anxious-depressed .35 / .31 .39 / .28 ***

Relationship Stress Depression / Anxious-depressed .36 / .35 .41 / .38

Mom Alienation Anxiety / Anxious-depressed .33 / .31 .34 / .33

Dad Alienation Anxiety / Anxious-depressed .33 / .28 * .33 / .28 *

Fairness Discipline Anxiety / Anxious-depressed -.01a / .06 ** -.15a / .00 ***

Teacher Alienation Anxiety / Anxious-depressed .25 / .16 ** .30 / .16 ***

Peer Victimization Anxiety / Anxious-depressed .29 / .31 .35 / .28 **

Relationship Stress Anxiety / Anxious-depressed .34 / .35 .37 / .38

Mom Alienation Rule-breaking / Rule-breaking .27 / .25 .38 / .25 ***

Dad Alienation Rule-breaking / Rule-breaking .32 / .26 ** .31 / .20 ***

Fairness Discipline Rule-breaking / Rule-breaking -.25 / -.03 *** -.23a / -.02 ***

Teacher Alienation Rule-breaking / Rule-breaking .28 / .17 *** .21 / .06 ***

Peer Victimization Rule-breaking / Rule-breaking .29 / .26 .27 / .24

Relationship Stress Rule-breaking / Rule-breaking .22 / .29 * .32 / .31

Note. *p < .05, **p < .01, ***p < .001. Instances where r for WBI subscale < r for YSR subscale are shown in
a
grey. As perceived Fairness of Discipline is conceptually expected to be inversely correlated with
symptoms, these comparisons were favorable for WBI subscales.
Table 6

Concordance Among Well-Being Index and YSR Subscales Using WBI Normative Sample Cutoffs

YSR Anxious-depressed

Clinically significant or Borderline or

“Much Above Average” (2 SD) “Above Average” (1.5 SD)

WBI Depression Boys Girls Boys Girls

Agreement on Both 69% 72% 67% 65%

WBI No / YSR Yes 30% 27% 31% 34%

WBI Yes / YSR No 1% 1% 2% 1%

YSR Anxious-depressed

Clinically significant or Borderline or

“Much Above Average” (2 SD) “Above Average” (1.5 SD)

WBI Anxiety Boys Girls Boys Girls

Agreement on Both 70% 70% 66% 63%

WBI No / YSR Yes 29% 30% 33% 36%

WBI Yes / YSR No 1% 0% 1% 1%

YSR Rule-breaking

Clinically significant or Borderline or

“Much Above Average” (2 SD) “Above Average” (1.5 SD)

WBI Rule-breaking Boys Girls Boys Girls

Agreement on Both 70% 74% 68% 73%

WBI No / YSR Yes 27% 23% 28% 25%

WBI Yes / YSR No 3% 3% 4% 2%

Note. n = 1217 for boys and 1227 for girls. Cutoffs of 2 and 1.5 SDs for the YSR were based on published norms; for the
WBI, these two cutoffs were based on means and standard deviations within this sample. Bolded values indicate the
percentage of agreement between the YSR and WBI on classification of clinically significant and borderline.