description: tags: REL 2007034 appab
-
Upload
anon-233946 -
Category
Documents
-
view
214 -
download
0
Transcript of description: tags: REL 2007034 appab
-
8/14/2019 description: tags: REL 2007034 appab
1/29
Appendix A 15
Appendix A
AnAlytic strAtegy
o describe the psychometric properties o the
secondary and elementary school resilience and
youth development modules, this report examines
Te dimensionality o scales by using explor-
atory and conrmatory actor analysis models.
Measurement equivalence across demo-
graphic subgroups by estimating conrma-
tory actor analysis models with covariates
(such as multiple indicator, multiple cause
structural equation models).
Scale reliability by estimating internal consis-
tency and test-retest reliability coecients.
Construct validity by examining the relation-
ship o scales to other theoretically related
constructs and mean dierences across demo-
graphic subgroups.
Data
Statewide data rom the local administration o the
Healthy Kids Survey.Te data or the analyses inthis report are rom local administration o the
Healthy Kids Survey (HKS) in elementary, middle,
and high schools. Tese data were drawn rom a
database o all local HKS data processed between
1998 and spring 2005 by WestEds Health and
Human Development Program (approximately
2.1 million observations). Analyzing such a large
sample size would, however, make almost every
parameter estimate statistically signicant, would
inate chi-square values o model t, and would
make assessing substantive signicance moredicult. Tus, two mutually exclusive analytic
samples were used in the analyses: a main sample
and a validation sample. Te samples were drawn
rom the aggregate data le that included all HKS
data processed between the spring 2003 and the
spring 2005 administrations o the Healthy Kids
Survey. For the secondary school analysis, separate
samples were drawn or each grade (7, 9, and 11),
gender, and ethnicity (Chinese American, Arican
American, Mexican American, and white Euro-
pean American)with 500 respondents randomly
sampled per cell (12,000 total). Equal numbers
were used or each gender and ethnic group so
that models that do not adjust or gender and/orethnicity would not be aected by gender/ethnic
dierences in the sample.
Te elementary school Healthy Kids Survey is ad-
ministered only to fh graders and does not ask
students about their ethnic/racial group. Random
samples o 1,000 males and 1,000 emales (2,000
total) were drawn rom the aggregated HKS data
le. Tus, or the elementary school resilience and
youth development module, only gender dier-
ences in measurement structure were examined.Respondents with missing data on more than
hal the resilience items were excluded rom the
analysis. For estimating models with missing data,
maximum likelihood estimation with missing at
random (MAR) assumptions were used, which
assume that values are missing at random con-
ditional on the other observed items in the data
(Little & Rubin, 2002; Muthn & Muthn, 2006).
(See section on missing data patterns.)
Te same procedures were used to draw thevalidation samples or both the secondary school
and elementary school samplesexcept that
respondents included in the main sample were ex-
cluded rom the validation sample. Te data were
weighted by grade, race/ethnicity, and gender to
represent the characteristics o HKS respondents
surveyed rom spring 2003 to spring 2005.
Local evaluation HKS data.Statewide data was
supplemented with two sets o HKS data originally
collected or local evaluation. Data collected in2006 rom a large urban school district in South-
ern Caliornia were used to describe the temporal
stability o the derived scales (test-retest reliability).
Te elementary school Healthy Kids Survey and the
secondary school core module and resilience and
youth development module were administered two
times in two weeks to 132 fh-grade students and
90 ninth-grade students. Data collected in 2004/05
-
8/14/2019 description: tags: REL 2007034 appab
2/29
16 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
rom students in a large county in Southern Cali-
ornia were used to examine the relationship be-
tween the RYDM constructs and standardized test
scores. Standardized test score and school/com-
munity asset data were available or 2,898 students,
while test score and home and internal asset datawere available or 651 students.6 English Language
Arts and Mathematics Caliornia Standards est
scale scores were used as criterion variables.
Missing data patterns. Approximately 0.5 percent
o respondents in the elementary and secondary
modules were excluded rom the sampling pool
because o missing data on more than hal the
resilience items (table A1). In the secondary school
samples, approximately 65 percent o respondents
provided answers to all the survey items in theresilience and youth development module; an ad-
ditional 18 percent had missing values on one or
two items; 8 percent had missing values on 3 to 10
items; and 8 percent had missing values on 11 or
more items. Respondents with missing values on
11 or more items had lower scores on about one-
quarter o the secondary RYDM itemsscoring
approximately 912 percent o a standard devia-
tion lower on these items. Tese results held or
both the main and validation samples. Dierences
in item means were diminished signicantly afercontrolling or one or two o the remaining items,
suggesting that the missing at random assumption
is reasonable.
Approximately 81 percent o elementary students
provided valid answers to all the RYDM items
and 15 percent answered all but one or two items.Respondents with missing values on two or more
items had lower scores on seven o the elementary
RYDM items (averaging 0.24 standard deviations).
Tese dierences were no longer apparent afer
controlling or any two o the remaining items,
again suggesting that maximum likelihood esti-
mation with missing at random assumptions will
yield unbiased parameter estimates.
Exploratory and conrmatory actor analyses
Analyses were conducted to test empirically
whether the actor structure o the resilience in-
strument is consistent with current usage and with
its underlying conceptual model. For each sample
and subsample (grade, gender, ethnicity), the mea-
surement structure o the resilience instrument
was established by tting a series o exploratory
and conrmatory actor analysis models. Explor-
atory actor analysis (EFA) models were estimated
to determine roughly the number o actors under-
lying the data and the measurement structure othe latent actors. A combination o actors was
tAble A1
M aa a o oa a ma am om
h a oh vom mo
nm
m
m
sa ema
Ma am vaa am Ma am vaa am
nm
p
nm
p
nm
p
nm
p
0 7,819 65.2 7,865 65.5 1,627 81.4 1,622 81.1
1 1,634 13.6 1,615 13.5 266 13.3 249 12.5
2 585 4.9 545 4.5 55 2.8 59 3.0
35 497 4.1 539 4.5 33 1.7 45 2.3
610 445 3.7 437 3.6 15 0.8 14 0.7
11 m 1,020 8.5 999 8.3 4 0.2 11 0.6
ta 12,000 100 12,000 100 2,000 100 2,000 100
Note: Analytic samples randomly drawn rom students surveyed between spring 2003 and spring 2005. Secondary school resilience and youth development
module has 51 survey items. The elementary school module has 21 survey items.
-
8/14/2019 description: tags: REL 2007034 appab
3/29
Appendix A 17
used to determine the number o actors to retain
in the EFAs, including t indices, scree plots, the
number o eigenvalues greater than 1, conceptual
clarity, and simplicity. Models with the ewest
possible actors and models with no cross-loadings
were avored over more complex models.
Te results o the exploratory actor analysis mod-
els were then used as a starting point or a series o
nested conrmatory actor analysis (CFA) models.
Measures o model t, correlations among the
latent constructs (actors), and actor-loading pat-
terns were used to make decisions about models.
Tis process was replicated or each grade, gender,
and ethnic group, and or the main sample and the
validation sample.
o derive estimates or the EFA and CFA models,
Muthn and Muthns (2006)Mplus statistical
modeling program was used. Because all the items
used to measure resilience assets are ordinal,
Muthns (1984) approach to exploratory and
conrmatory actor analysis with ordinal indica-
tors was used.
In the general actor analysis model, the relation-
ship between the indicators (y*) and the under-
lying constructs () can be represented by:
(A1) y*= + +
where is a vector o measurement intercepts, is
a matrix o measurement slopes (actor loadings),
and is a matrix o residuals, assumed to be inde-
pendent o and with zero expectation. Te model
implies the ollowing covariance matrix oy*:
(A2) = +
where is the covariance matrix o and is the
covariance matrix o (see Long, 1983).
In general, the indicatorsy*are assumed to be
normally distributed, latent continuous variables.
A persons observed score on itemy depends on
her/his position ony*. I the observed item is con-
tinuous,y*is directly observed (y =y*). However,
i the observed item is dichotomous or ordinal, the
observed categorical variable (y) is linked to the
latent continuous variable (y*) in a nonlinear way
through a model o thresholds (see Muthn, 1984).
Te relationships between an observed ordinal or
dichotomous itemy with c categories to y* can beexpressed as:
(A3) y = c, ic
-
8/14/2019 description: tags: REL 2007034 appab
4/29
18 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
emale to school meaningul participation indi-
cates that the means o the underlying construct
are allowed to be dierent or males and emales.
Te actor loadings are not allowed to be dier-
ent or males and emales, and there is no direct
eect o emale on the individual items. Te model
assumes that the items unction identically or
males and emales in measuring school meaning-
ul participation.8
Te measurement model in panel B allows or
emale/male nonequivalence in the measurement
intercept or item R14. Tat is, it allows or a direct
eect o emale on R14 that is not dependent on
the underlying construct. Tis is indicated by the
arrow going directly rom emale to R14. A sig-
nicant emale/male dierence in measurement
intercept indicates that the item unctions di-
erently or emales and males in measuring the
underlying construct. For example, i the measure-
ment intercept or R14 is 25 percent o a standard
deviation (emale R14) lower or emales than
males, then or a given level o school meaningulparticipation, emales score 25 percent o a standard
deviation lower on R14. In this example, a given
score on item R14 does not mean the same thing or
males and emalesat least not with reerence to
the school meaningul participation construct.
An applied strategy was used to ascertain whether
group dierences in measurement intercepts have
implications or evaluation research. Recommen-
dations or item changes are made only when the
measurement intercepts are substantively dierentacross groups ( 0.20 standard deviations) in both
the main sample and the validation sample.
Fit indices
A mean- and variance-adjusted 2 test o model t
is obtained by multiplying the minimum unc-
tion by twice the total sample size and dividing by
a scaling correction actor (or more details, see
Muthn, 1984, 1987; Muthn & Muthn, 2006).
Afer adjusting or the scaling correction ac-tor (see Satorra, 2000; Satorra & Bentler, 1999;
Muthn & Muthn, 2006), the dierence in 2 tests
or two nested models ollows a 2 distribution
and can be used to test whether a model results
in a statistically signicant improvement in t.
However, 2 dierence tests are sensitive to sample
size and can be inuenced by substantively mean-
ingless parameter dierences in large samples.
For this reason, the analysis also relied on several
other indices o model t.
For EFA models, root mean square residual
(RMSR) and root mean square error o approxi-
mation (RMSEA) values were used to assess model
t (Hu & Bentler, 1999). RMSR is the square root
o the mean o the squared residuals and indexes
the dierence between the sample variance/covari-
ance matrix and the variance/covariance matrix
predicted by the model. Hu and Bentler (1999)
R12
R13
R14
School
meaningful
participation
Panel A MIMIC modelingno measurement invariance
Female
R12
R13
R14
School
meaningful
participation
Panel B MIMIC modelinghypothetical gender
measurement intercept invariance (differential item
functioning for R14)
Female
igure A1
Hoha am o MiMic aoah
o o mam qva
Note: MIMIC reers to multiple indicators multiple causes structuralequation models.
-
8/14/2019 description: tags: REL 2007034 appab
5/29
Appendix A 19
suggest that RMSR values less than 0.05 indicate
good t. Te RMSEA is also based on dierences
between the observed and predicted variance/
covariance matrices, but penalizes or model com-
plexity. RMSEA is computed by:
(A5) RMSEA = 2
(n*df)1// ( )n
where 2 is the model chi-square value, n is the
total sample size, and dis the degrees o ree-
dom. RMSEA penalizes or model complexity
by dividing 2 by (n*d). Hu and Bentler (1999)
recommend RMSEA values o 0.06 or less as
the cut-o or good model t. Based on Hu and
Bentlers recommendations, more emphasis is
placed on RMSEA than on RMSR in EFA modelselection.
In addition to RMSEA, several additional t
indices were used to assess CFA models, includ-
ing Bentlers comparative t index (CFI), the
ucker-Lewis index (LI), and Muthn and
Muthns (2006) weighted root mean square
residual (WRMR). As implemented in Mplus, both
the CFI and LI compare estimated CFA models
to baseline models with uncorrelated variables
(independence model). CFI and LI are calculatedas ollows:
(A6) CFI =1max 2
Hodf
Ho, 0
max 2Ho
dfHo
, 2Bdf
B, 0
(A7)TLI =
2B
dfB
2Ho
dfHo
2B
dfB 1
where 2 and dHo denote the chi-squared value
and degrees o reedom o the estimated model
and 2 and dB denote the same or the baseline
model. Both CFI and LI are not appreciably
inuenced by sample size. By convention, CFI and
LI values greater than 0.95 indicate good t (Hu
& Bentler, 1999).
Yu and Muthn (2001) recently developed WRMR
to identiy good-tting models with categorical
outcomes. It is dened as ollows:
(A8) WRMR =e
(sr r)
vr
e
r
where sr is an element in the sample variance/
covariance (or probit threshold/polychoric cor-
relation) matrix, r is the element in the variance/
covariance matrix predicted by the model, r
is an estimate o the variance osr, and e is the
number o elements in the variance/covariance
matrix. According to Muthn, WRMR is suitable
or models where sample statistics have widely
varying variances, when sample statistics are on
dierent scales, and in models with categoricaloutcomes. Yu and Muthn (2001) suggest WRMR
values less than or equal to 1.00 or good models
with categorical outcomes. Because WRMR has
been tested or models with categorical outcomes,
greater weight is placed on this index in CFA
model selection.
Modication indices and 2 dierence testing
were also used to compare nested conrmatory
actor analyses models, particularly or testing
measurement intercept invariance.
Additional reliability and validity analyses
Internal consistency estimates o reliability o the
derived scales were calculated using Cronbachs
alpha or each grade, gender, and ethnic group in
both the main sample and the validation sample.
Nunnalys (1978) criterion o 0.70 was used as the
cuto or determining acceptable internal consis-
tency reliability or the secondary school survey.
Because o the notoriously low internal consis-tency evident in surveys o elementary school
students, this criterion was relaxed slightly to 0.60
or the elementary school resilience and youth
development module. o examine test-retest reli-
ability, RYDM survey data collected rom a small
sample o fh and ninth graders who took the
resilience and youth development module twice in
two weeks was used.
-
8/14/2019 description: tags: REL 2007034 appab
6/29
20 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
Dierences in resilience scale scores across the
demographic subgroups were also examined.
o make demographic dierences in the resil-
ience scales more interpretable, eect sizes were
calculated to represent the magnitude o such
dierences (Cohen, 1988). With two groups (male/emale), the dierence in scale means between
each group was divided by the pooled standard
deviation (Cohens d). Tus, the standardized
dierence represents the dierence between each
group in standard deviation units. With more
than two groups (race/ethnicity), the standard-
ized dierences were represented by multiplying
Cohensby 2which is roughly equivalent to the
standardized dierence calculated or two groups
when the number o observations in each cell is
equal (Cohen, 1988). Cohenswas calculated by
(A9) f=
(mim)2
kk
i=1
where mi represents the mean or each subgroup i,
m represents the population mean, k the number
o subgroups, and the pooled standard deviation.
Construct validity was assessed by examining
the relationship o the derived resilience scales toother theoretically related constructsincluding
substance use, school violence, school-related
behavior, and standardized test scores. o exam-
ine these relationships using a common metric,
correlations between resilience constructs and
criterion variables rom conrmatory actor analy-
sis models were estimated using the main and
validation samples. Latent constructs represent
continuous variables, while the criterion variables
are either dichotomous or ordinal. Tus, poly-
serial correlations are presented, which representthe correlation between a continuous variable and
a dichotomous or ordinal variable that reects
an underlying continuous variable (Bedrick &
Breslin, 1996).
-
8/14/2019 description: tags: REL 2007034 appab
7/29
Appendix b 21
Appendix B
results
Tis appendix presents the results o the analyses
conducted to evaluate the psychometric properties
o the resilience and youth development module.
Secondary school environmental resilience assets
Exploratory actor analysis results. EFA models were
estimated or each subpopulation and or the main
and validation samples to determine the number
o actors underlying the items. Te EFA models
suggested that the environmental resilience assets
items measure eight actors.9 Te actor pattern and
loadings or the main sample and cross-validation
sample are displayed in tables B1 and B2, respec-tively. Te 8-actor EFA solutions show conceptu-
ally clear actor-loading patterns that are mostly
consistent with the underlying theory guiding the
development o the instrument. Te pattern o ac-
tor loadings across all the demographic subgroups
is consistent with those displayed in tables B1 and
B2.10 Distinct actors are apparent or support and
meaningul participation in the school, community,
and home environments, as well as caring and pro-
social relationships in the peer environment.
However, the actor pattern evident in the 8-actor
solution is inconsistent with how the instrument
currently is being used in Caliornia because the
results suggest that caring relationships and high
expectations at school, in the home, and in the
community are notdistinct actors.
Confrmatory actor analysis results. A CFA model
equivalent to the 8-actor EFA models in tables
B1 and B2 was estimatedexcept that all but the
highest magnitude loadings rom the EFA modelwere constrained to be zero.11 Tat is, each item
was orced to load on only one actor. As with the
EFA models, the results were consistent across
each sample. Te CFA models indicated that item
R45 (My riends get into a lot o trouble) has a
relatively small actor loadingsuggesting that an
association with peers who get into a lot o trouble
is a less sensitive indicator o pro-social peers
than the other two items assessing this construct.
Moreover, there was a relatively high correlation
between home support and home meaningul
participation (0.78 and 0.79), which suggests that
these two constructs may not be distinct.
Te CFA models were re-estimated to include covari-
ates to detect dierences in measurement intercepts
across demographic subgroups. Several measure-
ment intercepts diered by demographic subgroup:
Te results or R23 (I help other people)
suggest that or a given level o community
meaningul participation, emale and Mexi-
can American youth report between one-fh
and one-third o a standard deviation higher
or helping other people. Te item thus has adierent meaning or these two populations.
For R54 (I do un things or go un places
with my parents), 11th graders report
substantially lower levels o participation in
un activities with parents or a given level o
home meaningul participation than do sev-
enth and ninth graders (0.29 to 0.33 standard
deviations). Tis represents a developmental
dierence in the appropriateness o this item.
Female and Chinese American youth report
lower requencies on R45 (My riends get into
a lot o trouble) or a given level o pro-social
peersreecting the dierent meaning at-
tached to this item by these populations.
Each o these measurement intercept dierences is
substantively signicant. Tat is, these particular
items assess the underlying constructs dierently
or demographic subgroups and thus should not be
used as indicators. Dropping these items, however,leaves three subscales with only two items, which is
ar rom ideal. able B3 presents revised CFA mod-
els afer dropping the items with non-invariant
measurement intercepts. able B4 reports latent
actor correlations.12 Note that the correlations
between home support and home meaningul par-
ticipation remain relatively high (0.73), indicating a
high degree o overlap between these two actors.
-
8/14/2019 description: tags: REL 2007034 appab
8/29
22 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
tAble b1
soa hoo voma a oao ao
aa , ma am, 8-ao oo
oa
a
im im 1 2 3 4 5 6 7 8
r6 sca sa w a a a m. 0.75 0.08 0.02 0.02 0.07 0.03 0.06 0.01
r8 sca sa w w im . 0.79 0.02 0.01 0.03 0.03 0.04 0.04 0.06
r10 sca sa w m
w i a m . . . 0.86 0.02 0.01 0.01 0.02 0.04 0.02 0.00
r7 sh sa w m w i a j. 0.82 0.02 0.00 0.01 0.02 0.01 0.02 0.02
r9 sh sa w awa wa m m . 0.92 0.05 0.02 0.03 0.05 0.06 0.03 0.02
r11 sh sa w a i w a . 0.83 0.01 0.05 0.00 0.03 0.01 0.05 0.04
r12 spa si a. 0.08 0.57 0.01 0.19 0.08 0.06 0.01 0.01
r13 spa si k a a 0.02 0.91 0.02 0.09 0.01 0.02 0.00 0.00
r14 spa si a mak a . 0.04 0.79 0.04 0.01 0.02 0.05 0.00 0.04
r15 cmca cmma w a a a m. 0.04 0.05 0.95 0.03 0.04 0.04 0.02 0.00
r17 cmca cmma w w i am a . . . 0.02 0.03 0.90 0.05 0.01 0.07 0.05 0.04
r20 cmca cmma wm i . 0.02 0.04 0.82 0.02 0.03 0.08 0.00 0.00
r16 cmh cmma w m w i a j. 0.01 0.01 0.90 0.02 0.03 0.04 0.01 0.01
r18 cmh cmma w a i w a . 0.02 0.05 0.90 0.02 0.10 0.05 0.02 0.03
r19 cmh cmma w awa wa m m . 0.04 0.01 0.95 0.00 0.05 0.08 0.03 0.04
r21 cmpa i am a , am, /
m, . . . 0.03 0.06 0.02 0.82 0.01 0.03 0.04 0.03
r22 cmpa i am ak m, a, a . . . 0.02 0.07 0.03 0.97 0.00 0.01 0.01 0.06
r23 cmpa i . 0.05 0.10 0.09 0.46 0.09 0.19 0.08 0.07
r49 hmca hm a w m wk . 0.02 0.07 0.03 0.00 0.86 0.01 0.01 0.02
r51 hmca hma w ak w m a m m. 0.03 0.08 0.02 0.12 0.77 0.27 0.01 0.10r53 hmca hma w m
w i a m . . . 0.02 0.01 0.03 0.12 0.76 0.32 0.03 0.06
r48 hmh hma w m w . 0.01 0.02 0.01 0.13 0.76 0.18 0.06 0.09
r50 hmh hma w a i w a . 0.04 0.02 0.07 0.03 0.83 0.02 0.00 0.03
r52 hmh hma w awa wa m m . 0.03 0.08 0.05 0.09 0.89 0.08 0.01 0.06
r54 hmpa i a w
m a . . . 0.01 0.08 0.01 0.04 0.30 0.63 0.02 0.04
r55 hmpa i a m a mak a . 0.02 0.11 0.00 0.08 0.09 0.68 0.00 0.08
r56 hmpa i mak w m am. 0.03 0.02 0.02 0.01 0.23 0.70 0.03 0.01
r42 pca A w a a a m. 0.02 0.04 0.04 0.05 0.04 0.04 0.83 0.06
r43 pca A w ak w m a m m. 0.02 0.03 0.01 0.03 0.01 0.01 0.96 0.00
r44 pca A w m w im a a a m. 0.00 0.00 0.00 0.03 0.02 0.02 0.92 0.02
r45 ph M a . 0.05 0.05 0.03 0.03 0.01 0.01 0.02 0.45
r46 ph M wa . 0.02 0.02 0.02 0.07 0.03 0.05 0.04 0.92
r47 ph M w . 0.02 0.03 0.01 0.01 0.05 0.05 0.02 0.68
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Weighted data. Loadings with largest absolute values bolded.
-
8/14/2019 description: tags: REL 2007034 appab
9/29
Appendix b 23
tAble b2
soa hoo voma a oao ao
aa , vaao am, 8-ao oo
oa
a
im im 1 2 3 4 5 6 7 8
r6 sca sa w a a a m. 0.76 0.03 0.03 0.01 0.10 0.06 0.05 0.03
r8 sca sa w w im . 0.78 0.03 0.02 0.03 0.04 0.04 0.03 0.04
r10 sca sa w m
w i a m . . . 0.85 0.02 0.02 0.03 0.04 0.04 0.01 0.01
r7 sh sa w m w i a j. 0.82 0.02 0.01 0.01 0.01 0.01 0.01 0.02
r9 sh sa w awa wa m m . 0.90 0.07 0.01 0.03 0.05 0.05 0.04 0.01
r11 sh sa w a i w a . 0.84 0.07 0.01 0.03 0.06 0.05 0.01 0.02
r12 spa si a. 0.11 0.59 0.01 0.18 0.06 0.07 0.04 0.03
r13 spa si k a a 0.03 0.88 0.03 0.09 0.01 0.01 0.00 0.00
r14 spa si a mak a . 0.02 0.80 0.04 0.00 0.02 0.04 0.01 0.03
r15 cmca cmma w a a a m. 0.02 0.06 0.95 0.03 0.08 0.01 0.03 0.02
r17 cmca cmma w w i am a . . . 0.01 0.01 0.89 0.06 0.02 0.08 0.03 0.03
r20 cmca cmma wm i . 0.00 0.02 0.83 0.01 0.04 0.06 0.02 0.03
r16 cmh cmma w m w i a j. 0.03 0.01 0.89 0.02 0.01 0.07 0.00 0.01
r18 cmh cmma w a i w a . 0.02 0.08 0.89 0.02 0.11 0.07 0.01 0.00
r19 cmh cmma w awa wa m m . 0.04 0.02 0.95 0.01 0.07 0.12 0.03 0.01
r21 cmpa i am a , am, /
m, . . . 0.03 0.06 0.02 0.83 0.02 0.01 0.02 0.03
r22 cmpa i am ak m, a, a . . . 0.00 0.07 0.01 0.97 0.02 0.03 0.02 0.05
r23 cmpa i . 0.04 0.13 0.08 0.47 0.08 0.16 0.09 0.05
r49 hmca hm a w m wk . 0.03 0.05 0.03 0.03 0.85 0.03 0.01 0.00
r51 hmca hma w ak w m a m m. 0.05 0.08 0.04 0.12 0.74 0.30 0.02 0.09r53 hmca hma w m
w i a m . . . 0.02 0.03 0.06 0.12 0.73 0.32 0.02 0.07
r48 hmh hma w m w . 0.01 0.03 0.00 0.11 0.75 0.20 0.07 0.12
r50 hmh hma w a i w a . 0.06 0.02 0.07 0.03 0.81 0.04 0.01 0.03
r52 hmh hma w awa wa m m . 0.08 0.08 0.05 0.08 0.85 0.05 0.03 0.03
r54 hmpa i a w
m a . . . 0.04 0.07 0.01 0.05 0.23 0.67 0.06 0.06
r55 hmpa i a m a mak a . 0.03 0.15 0.03 0.10 0.06 0.68 0.04 0.05
r56 hmpa i mak w m am. 0.02 0.02 0.00 0.01 0.16 0.77 0.02 0.02
r42 pca A w a a a m. 0.04 0.05 0.04 0.04 0.03 0.02 0.83 0.05
r43 pca A w ak w m a m m. 0.02 0.03 0.02 0.03 0.00 0.02 0.96 0.01
r44 pca A w m w im a a a m. 0.01 0.01 0.01 0.03 0.03 0.01 0.91 0.03
r45 ph M a . 0.07 0.07 0.04 0.01 0.02 0.03 0.09 0.42
r46 ph M wa . 0.02 0.02 0.01 0.04 0.01 0.02 0.07 0.85
r47 ph M w . 0.01 0.04 0.04 0.05 0.02 0.05 0.01 0.77
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Weighted data. Loadings with largest absolute values bolded.
-
8/14/2019 description: tags: REL 2007034 appab
10/29
24 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
tAble b3
Fa oa hoo voma a mo, ma am
im
oa
c a aa m
ema
a
saa
a
School support
r6 sca sa w a a a m. 1 0.80
r8 sca sa w w im . 0.98 0.79
r10 sca sa w m w i a m . . . 1.08 0.86
r7 sh sa w m w i a j. 1.05 0.84
r9 sh sa w awa wa m m . 1.09 0.87
r11 sh sa w a i w a . 1.10 0.88
School meaningul participation
r12 spa si a. 1 0.78
r13 spa si k a a . 0.98 0.77
r14 spa si a mak a . 1.12 0.88
Community supportr15 cmca cmma w a a a m. 1 0.92
r17 cmca cmma w w i am a . . . 0.99 0.91
r20 cmca cmma wm i . 0.97 0.88
r16 cmh cmma w m w i a j. 1.03 0.94
r18 cmh cmma w a i w a . 1.04 0.95
r19 cmh cmma w awa wa m m . 1.04 0.95
Community meaningul participation
r21 cmpa i am a , am, /m, . . . 1 0.88
r22 cmpa i am ak m, a, a . . . 0.97 0.86
Home support
r49 hmca hma w m wk. 1 0.84
r51 hmca hma w ak w m a m m. 1.03 0.87
r53 hmca hma w m w i a m . . . 1.05 0.89
r48 hmh hma w m w . 0.93 0.78
r50 hmh hma w a i w a . 1.10 0.92
r52 hmh hma w awa wa m m . 1.10 0.92
Home meaningul participation
r55 hmpa i a m a mak a . 1 0.85
r56 hmpa i mak w m am. 1.02 0.86
Peer caring relationships
r42 pca A w a a a m. 1 0.92
r43 pca A w ak w m a m m. 1.01 0.92
r44 pca A w m w im a a a m. 1.03 0.94
Pro-social peers
r46 ph M wa . 1 0.86
r47 ph M w . 0.91 0.78
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
11/29
Appendix b 25
Elementary school environmental resilience assets
Exploratory actor analysis results. An identi-
cal strategy was used to analyze the elementary
school RYDM environmental resilience items.
EFA models suggested that a 4-actor model best
represents the environmental resilience items,
with distinct actors or school support (car-
ing relationships and high expectations), home
support, meaningul participation (in the school
and home domains), and pro-social peers (tables
B5 and B6). Tese results were ound or both the
main sample and the validation sample and or
both boys and girls.
tAble b4
coao amo oa hoo voma
a, a omao ao aa mo
a
Ma am (1) (2) (3) (4) (5) (6) (7) (8)
(1) s 1.00
(2) s ma aa 0.59 1.00
(3) cmm 0.54 0.42 1.00
(4) cmm ma aa 0.42 0.58 0.46 1.00
(5) hm 0.47 0.37 0.59 0.44 1.00
(6) hm ma aa 0.48 0.59 0.51 0.38 0.73 1.00
(7) p a a 0.41 0.35 0.46 0.34 0.46 0.44 1.00
(8) p-a 0.42 0.40 0.38 0.39 0.49 0.50 0.54 1.00
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Weighted data.
tAble b5
ema hoo voma a oao ao
aa , ma am, 4-ao oo
im
oa
im 1 2 3 4
10 sca d a . . . a a a ? 0.74 0.05 0.01 0.01
13 sca ta . . . w . . . a m a? 0.62 0.07 0.00 0.05
11 sh ta . . . w a j? 0.56 0.02 0.17 0.07
14 sh ta . . . a a a j? 0.67 0.10 0.02 0.03
52 hmca pa . . . a a wk? 0.00 0.81 0.01 0.01
55 hmca pa . . . w a m a? 0.06 0.51 0.20 0.01
53 hmh pa . . . a a a j? 0.11 0.79 0.00 0.02
54 hmh pa . . . a m wa ? 0.10 0.77 0.03 0.00
9 spa d mak a / a ? 0.14 0.16 0.48 0.05
15 spa d a ? 0.16 0.11 0.50 0.15
56 hmpa d a m? 0.17 0.21 0.48 0.03
56 hmpa d mak a m? 0.10 0.05 0.37 0.03
50 ph d ? 0.03 0.01 0.07 0.63
51 ph d ? 0.02 0.04 0.18 0.68
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
Loadings with largest absolute values bolded.
-
8/14/2019 description: tags: REL 2007034 appab
12/29
26 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
Confrmatory actor analysis results. Te CFA results
also supported the 4-actor model. Te analyses
o dierential item unctioning suggested that the
measurement intercepts or item 15 (Do you do
things to be helpul at school?) and item 51 (Do
your best riends try to do the right thing?) dieror boys and girls. For a given level o meaningul
participation, emales report between 20 and 36
percent o a standard deviation higher requen-
cies o doing things to be helpul at school or a
given level o meaningul participation. In addition,
emales are substantially less likely to report that
their best riends try to do the right thing (0.43
standard deviations). Because o the magnitude
o these measurement intercept dierences, these
items should not be used to measure the underlying
constructs. Because dropping item 51 means thatonly one item is lef to measure pro-social peers,
item 50 (Do your best riends get into trouble?)
should also be dropped. Te elementary school
module thus would not assess pro-social peer assets.
Afer dropping the pro-social peer items, a 3-actor
model is lefwith actors or school support,
home support, and meaningul participation.
Because meaningul participation is measured
with only three items, a 2-actor model was also
estimated by orcing the relevant meaningul
participation items to load on the school and home
actors. Te t o the 2-actor model is relatively
close to that o the 3-actor model, although thelatter resulted in a statistically signicant im-
provement in model t (see 2 values or Model
4 versus Model 3 in appendix C). Moreover, an
inspection o the standardized loadings in the
2-actor model or items 9, 56, and 57 indicates
that these meaningul participation items are only
weakly related to underlying school and home
actors (0.26, 0.37, and 0.24, respectively). Tus the
3-actor model has the most support.
able B7 presents the results or the nal CFAmodel. A look at the standardized actor loadings
reveals that the relationships between meaning-
ul participation and its items are still weak (0.40,
0.53, and 0.30 or items 9, 56, and 57, respectively),
suggesting that the items are insensitive indicators
o meaningul participation. Overall, however, the
results are consistent with those reported or the
secondary school resilience and youth development
tAble b6
ema hoo voma a oao ao
aa , vaao am, 4-ao oo
im
oa
im 1 2 3 4
10 sca d a . . . a a a ? 0.73 0.03 0.03 0.0113 sca ta . . . w . . . a m a? 0.65 0.02 0.03 0.01
11 sh ta . . . w a j? 0.51 0.05 0.18 0.02
14 sh ta . . . a a a j? 0.71 0.14 0.07 0.02
52 hmca pa . . . a a wk? 0.01 0.73 0.00 0.01
55 hmca pa . . . w a m a? 0.07 0.48 0.22 0.01
53 hmh pa . . . a a a j? 0.10 0.90 0.09 0.01
54 hmh pa . . . a m wa ? 0.07 0.81 0.01 0.04
9 spa d mak a / a ? 0.05 0.20 0.62 0.06
15 spa d a ? 0.10 0.05 0.50 0.18
56 hmpa d a m? 0.07 0.16 0.36 0.08
56 hmpa d mak a m? 0.16 0.16 0.38 0.0950 ph d ? 0.02 0.03 0.00 0.51
51 ph d ? 0.02 0.03 0.05 0.77
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
13/29
Appendix b 27
modulewith the exception that meaningul par-ticipation is global, rather than domain-specic,
or the elementary school items and that pro-social
peers cannot be measured adequately.
Secondary school internal resilience assets
Exploratory actor analysis results. Te EFA models
indicated that two o the three items used to assess
cooperation and communicationR36 (I enjoy
working together with other students my age)
and R37 (I stand up or mysel without puttingothers down)either load on more than one
actor or do not load signicantly on any actor.
For simplicity, these items were dropped rom the
analysis, and EFA and CFA models were estimated
on the remaining set o items. Te EFA results
suggested that ve actors underlie the observed
itemssel-ecacy, empathy, problem solving,
sel-awareness, and goals/aspirations (tables B8
and B9). Te 5-actor solution is conceptually clearand is consistent with how the instrument is cur-
rently used in Caliornia.
Confrmatory actor analysis results. CFA models
consistent with the 5-actor EFA model were esti-
mated, with all but the highest loadings rom the
EFA models constrained to be zero. Several consis-
tent, substantively signicant dierences in mea-
surement intercepts across racial/ethnic groups
were evident when covariates were included:
Female youth are between 0.22 and 0.34 o a
standard deviation less likely to endorse item
R27 (I know where to go or help with a prob-
lem) or a given level o problem solving.
Arican American and Mexican American
youth report higher levels o having goals
and plans or the uture (R24) than white
tAble b7
Fa ma hoo voma a mo, ma am
im
oa
c a aa m
ema
a
saa
a
School support
10 sca d a . . . a a a ? 1 0.76
13 sca ta . . . w a m a? 0.90 0.68
11 sh d a . . . w a j? 0.79 0.60
14 sh d a . . . a a a j? 0.95 0.72
Home support
52 hmca d a a . . . a a wk? 1 0.78
55 hmca d a a . . . w a m a? 0.89 0.69
53 hmh d a a . . . a a a j? 1.11 0.86
54 hmh d a a . . . a m wa ? 1.01 0.79
Meaningul participation
9 spa d mak a a ? 1 0.4056 hmpa d a m? 1.36 0.53
57 hmpa d mak a m? 0.77 0.30
la a a
(1) (2) (3)
(1) s 1.00
(2) hm 0.64 1.00
(3) Ma aa 0.48 0.62 1.00
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
14/29
28 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
European American and Chinese American
youth, even afer accounting or ethnic groupdierences in the latent construct.
Chinese American youth report substantially
lower levels o having goals and plans or the
uture than the other ethnic groups or a given
score on the underlying construct. In addition,
Chinese American youth also are about 2530
percent o a standard deviation less likely to
report that they plan to go to college afer high
school (R26) or a given level on goals.
With such pronounced racial/ethnic group mea-
surement intercept dierences, items R24 and R26
should be dropped, and so goals would not be as-
sessed on the secondary school module. Item R27
should not be used to assess problem solving.
able B10 shows the nal recommended CFA
model or the secondary school internal resilience
items afer dropping items R24, R25, R26, and R27
rom the analysis. Overall, the latent constructsare consistent with current usage o the RYDM,
except that the communication and cooperation
construct is dropped because two o the items or
this scale did not uniquely load on one actor, the
goals construct is dropped because o measure-
ment slope invariance, and the problem-solving
construct is measured with just two items.
Elementary school internal resilience assets
Exploratory actor analysis results. Te elementaryschool resilience and youth development module
was designed to measure three internal resilience
traitsempathy, problem solving, and goals and
aspirationswith seven items. Although explor-
atory actor analyses o these items suggest that
a 2-actor solution was appropriate or both the
main and validation samples, the actor patterns
were dierent or the two samples as well as or
tAble b8
soa hoo a a oao ao aa , ma am, 4-ao mo
im
oa
im
a
1 2 3 4
r31 c i a wk w m w a a m. 0.39 0.16 0.08 0.23
r36 c i j wk w m a.im a -a
r37 c i a m w w.
r29 se i a wk m m. 0.66 0.09 0.15 0.11
r30 se i a m i . 0.50 0.11 0.16 0.35
r32 se t a ma a i w. 0.32 0.02 0.21 0.37
r33 ema i a w m . 0.06 0.71 0.02 0.15
r34 ema i a wa . 0.02 0.88 0.02 0.09
r38 ema i a wa a k. 0.11 0.70 0.12 0.03
r35 ps W i i m ak w. 0.64 0.33 0.07 0.18
r27 ps i kw w w a m. 0.68 0.01 0.07 0.11
r28 ps i wk m ak w a m. 0.80 0.24 0.14 0.12
r39 sAwa t a m . 0.12 0.06 0.46 0.28r40 sAwa i a m m a . 0.01 0.05 0.91 0.07
r41 sAwa i a w i wa i . 0.01 0.03 0.86 0.04
r24 ga i a a a a . 0.11 0.02 0.10 0.64
r25 ga i a aa m . 0.09 0.07 0.02 0.98
r26 ga i a m a . 0.01 0.09 0.11 0.88
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
15/29
Appendix b 29
males and emales. Te items measure empathy
and goals/aspirations, but item 40 (Do you tryto work out your problems by talking or writing
about them?) either cross-loads or does not load
signicantly on the two actors, depending on the
analytic sample (see tables E124aE132). Te EFA
actor patterns were still ambiguous afer drop-
ping item 40, most likely because so ew items re-
mained to be analyzed (tables B11 and B12). Afer
moving to a CFA ramework, two nested models
were estimateda 1-actor model measuring over-
all internal assets and a 2-actor model measuring
empathy and goals/aspirations.
Confrmatory actor analysis results. Te 2-actor
CFA modelwhich includes distinct actors or
empathy and goals/aspirationsexhibited a sig-
nicantly better t to the observed data than the
1-actor model. able B13 presents the actor load-
ings and actor correlations or this CFA model
based on the main sample. An examination o the
standardized item loadings or goals/aspirations
indicates that two o the our items are weakly as-sociated with the underlying construct. Although
goals/aspirations is poorly measured by the in-
cluded items, this scale should be retained so that
its reliability and relationship to other constructs
can be urther investigated.
Reliability o the secondary and elementary school scales
Internal consistency.Te internal consistency o
the RYDM scales was estimated using Cronbachs
alpha coecient or the main sample, the valida-tion sample, and each demographic subsample
(tables B14 and B15). Te secondary school RYDM
scales (table B14) demonstrate acceptable levels o
reliability, with all scales exhibiting reliabilities
greater than 0.70, and 11 o 13 scales demonstrat-
ing reliabilities greater than 0.75. Te school
support, community support, and peer caring
relationships scales exhibit the highest internal
tAble b9
soa hoo a a oao ao aa , vaao am, 4-ao mo
im
oa
im
a
1 2 3 4
r31 c i a wk w m w a a m. 0.45 0.11 0.06 0.22
r36 c i j wk w m a.im a -a
r37 c i a m w w.
r29 se i a wk m m. 0.65 0.06 0.13 0.12
r30 se i a m i . 0.55 0.10 0.10 0.36
r32 se t a ma a i w. 0.46 0.10 0.16 0.36
r33 ema i a w m . 0.12 0.68 0.03 0.14
r34 ema i a wa . 0.02 0.84 0.03 0.11
r38 ema i a wa a k. 0.09 0.70 0.12 0.09
r35 ps W i i m ak w. 0.63 0.31 0.10 0.21
r27 ps i kw w w a m. 0.66 0.02 0.09 0.07
r28 ps i wk m ak w a m. 0.81 0.26 0.15 0.14
r39 sAwa t a m . 0.17 0.05 0.45 0.27r40 sAwa i a m m a . 0.02 0.07 0.94 0.07
r41 sAwa i a w i wa i . 0.09 0.01 0.79 0.01
r24 ga i a a a a . 0.11 0.04 0.10 0.63
r25 ga i a aa m . 0.09 0.11 0.03 0.97
r26 ga i a m a . 0.01 0.11 0.08 0.85
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
16/29
30 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
consistency, with alphas all exceeding 0.90. Te
problem-solving (alpha = 0.73) and pro-social
peers (alpha = 0.74) scales exhibit moderate but
acceptable levels o internal consistency, especially
considering that the scales have only two items.
Internal consistency does not dier markedly by
student grade, gender, or race/ethnicity. However,
the problem-solving scale shows lower reliability
tAble b10
Fa oa hoo a a mo, ma am
im
oa
c a aa m
ema
a
saa
a
Sel-efcacy
r31 c i a wk w m w a . . . 1.00 0.77r29 se i a wk m m. 1.04 0.80
r30 se i a m i . 1.09 0.84
r32 se t a ma a i w. 1.04 0.80
Empathy
r33 ema i a w m . 1.00 0.82
r34 ema i a wa . 1.11 0.91
r38 ema i a wa a k. 1.09 0.90
Problem solving
r35 ps W i i m ak w. 1.00 0.85
r28 ps i wk m ak/w a m. 0.94 0.80
Sel-awarenessr39 sAwa t a m . 1.00 0.84
r40 sAwa i a m m a . 1.02 0.86
r41 sAwa i a w i wa i . 0.99 0.83
la a a
(1) (2) (3) (4)
(1) s-fa 1.00
(2) ema 0.73 1.00
(3) pm 0.78 0.82 1.00
(4) s-awa 0.82 0.69 0.62 1.00
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Weighted data.
tAble b11
ema hoo a a oao ao aa , ma am, 2-ao mo
im
oa
im
a
1 2
37 ema d a w ? 0.70 0.04
38 ema d a w m ? 0.73 0.03
39 ps d kw w w a m? 0.06 0.63
40 ps d wk m ak/w . . . ? 0.31 0.36
41 ga/A d ? 0.17 0.52
42 ga/A d a a a a ? 0.03 0.38
16 ga/A d a . . . a ? 0.07 0.34
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
17/29
Appendix b 31
or Arican American students than or other
ethnic groups.
Internal consistency reliabilities or the elemen-
tary school RYDM scales are noticeably lowerthan those or the secondary school instrument
(see table B15). Tese low reliabilities are typical
o instruments administered to elementary school
students. Te school support, home support, and
empathy subscales demonstrate adequate reliabil-
itywith alphas ranging rom 0.63 to 0.65 or em-
pathy to 0.70 to 0.72 or school and home support.
Te elementary school meaningul participation
and goals/aspirations scales exhibit low levels
o reliability. Tese scales should not be used in
research or local evaluation activities requiring
precise measurement.
Stability.ables B16 and B17 show construct- and
item-level test-retest stability coecients or the
secondary school RYDM asset measures. Unlike the
internal consistency estimates, the resilience scales
evidence airly low levels o stability, with 8 o the 12
scales exhibiting pre-post correlations o less than
0.60. Only the community meaningul participa-
tion, home support, peer caring relationships, and
tAble b13
Fa ma hoo a a mo, ma am
im
oa
c a aa m
ema
a
saa
a
Empathy
37 ema d a w ? 1 0.71
38 ema d a w m ? 1.07 0.76
Goals/aspirations
39 ps d kw w w a m? 1 0.50
41 ga/A d ? 1.56 0.78
42 ga/A d a a a a ? 0.69 0.35
16 ga/A d a . . . a ? 0.50 0.25
la a a(1) (2)
(1) ema 1.00
(2) ga/aa 0.64 1.00
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
tAble b12
ema hoo a a oao ao aa , vaao am, 2-ao mo
im
oa
im
a
1 2
37 ema d a w ? 0.80 0.13
38 ema d a w m ? 0.80 0.06
39 ps d kw w w a m? 0.20 0.42
40 ps d wk m ak/w . . . ? 0.38 0.22
41 ga/A d ? 0.34 0.36
42 ga/A d a a a a ? 0.17 0.76
16 ga/A d a . . . a ? 0.08 0.56
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
18/29
32 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
sel-awareness scales demonstrate adequate stabil-ity. In the context o relatively high levels o internal
consistency, these comparatively low levels o stabil-
ity suggest that the resilience assets assessed by the
secondary school module demonstrate adequate
reliability at a single point in time.
A look at the item-specic stability coecients
in tables B16 and B17 shows the variation across
items. Although several are particularly unstable,the individual item test-retest reliabilities have
a negligible impact on the total scale test-retest
reliabilities. For example, the pre-post correlation
o item R8 (Tere is a teacher or some other adult
who notices when I am not there) is only 0.29.
However, dropping this item rom the school sup-
port scale does not markedly improve the stability
o the scale score (0.54 versus. 0.55).
tAble b14
soa hoo a o ab of b moah bo
ga g ra/
A 7 9 11 Ma ma
Aa
Ama
c
Ama
Ma
Ama W
Environmental resilience assets
s 0.90 0.89 0.91 0.92 0.90 0.91 0.89 0.90 0.90 0.91
s ma aa 0.78 0.76 0.77 0.80 0.77 0.78 0.74 0.79 0.78 0.79
cmm 0.95 0.94 0.95 0.95 0.95 0.95 0.95 0.95 0.95 0.95
cmm ma aa 0.75 0.73 0.76 0.77 0.76 0.75 0.76 0.72 0.77 0.73
hm 0.89 0.89 0.89 0.88 0.89 0.88 0.90 0.86 0.90 0.88
hm ma aa 0.78 0.76 0.78 0.79 0.78 0.78 0.75 0.79 0.77 0.79
p a a 0.90 0.88 0.90 0.92 0.89 0.90 0.89 0.90 0.91 0.89
p-a 0.74 0.73 0.75 0.74 0.72 0.74 0.72 0.71 0.75 0.73
Internal resilience assets
s-fa 0.82 0.81 0.82 0.82 0.83 0.80 0.82 0.82 0.82 0.81ema 0.85 0.86 0.85 0.85 0.85 0.83 0.84 0.85 0.85 0.86
pm 0.73 0.73 0.73 0.72 0.73 0.69 0.68 0.74 0.73 0.76
s-awa 0.81 0.80 0.81 0.82 0.82 0.80 0.81 0.82 0.80 0.81
Note: Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005.
Results are based on main sample. Cronbachs alpha coecients were almost identical in the validation sample.
tAble b15
ema hoo a o ab of b
Ma am vaa am
A Ma ma A Ma maEnvironmental resilience assets
s 0.71 0.70 0.72 0.70 0.71 0.69
hm 0.71 0.70 0.71 0.72 0.71 0.73
Ma aa 0.34 0.32 0.35 0.32 0.30 0.34
Internal resilience assets
ema 0.63 0.64 0.57 0.65 0.63 0.64
ga/aa 0.36 0.41 0.27 0.39 0.43 0.33
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
19/29
Appendix b 33
tAble b16
t- ab o oa hoo voma a o a m
im
oa
c a aa m
sa
f ()
School support 0.54
r6 sca sa w a a a m. 0.50
r8 sca sa w w im . 0.29
r10 sca sa w m w i a m . . . 0.51
r7 sh sa w m w i a j. 0.43
r9 sh sa w awa wa m m . 0.47
r11 sh sa w a i w a . 0.46
School meaningul participation 0.53
r12 spa si a. 0.33
r13 spa si k a a . 0.56
r14 spa si a mak a . 0.37
Community support 0.44r15 cmca cmma w a a a m. 0.33
r17 cmca cmma w w i am a . . . 0.41
r20 cmca cmma wm i . 0.53
r16 cmh cmma w m w i a j. 0.44
r18 cmh cmma w a i w a . 0.39
r19 cmh cmma w awa wa m m . 0.46
Community meaningul participation 0.82
r21 cmpa i am a , am, /m, . . . 0.83
r22 cmpa i am ak m, a, a . . . 0.64
Home support 0.68
r49 hmca hma w m wk. 0.57
r51 hmca hma w ak w m a m m. 0.62
r53 hmca hma w m w i a m . . . 0.60
r48 hmh hma w m w . 0.53
r50 hmh hma w a i w a . 0.52
r52 hmh hma w awa wa m m . 0.63
Home meaningul participation 0.49
r55 hmpa i a m a mak a . 0.52
r56 hmpa i mak w m am. 0.43
Peer caring relationships 0.73
r42 pca A w a a a m. 0.52
r43 pca A w ak w m a m m. 0.62
r44 pca A w m w im a a a m. 0.76
Pro-social peers 0.51
r46 ph M wa . 0.51
r47 ph M w . 0.46
Note: Results are based on a sample o 90 ninth-grade respondents rom seven classrooms in two schools in a large urban school district. Two weeks sepa-
rated the rst and second administrations o the survey instruments.
-
8/14/2019 description: tags: REL 2007034 appab
20/29
34 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
Te elementary school RYDM scales exhibit
higher stability than the secondary school scales
(table B18). Only two o the ve elementary school
scales exhibit pre-post correlations below 0.60.
Te stability coecients in table B18 are similaror higher than the internal consistency reliability
estimates presented in table B15 or elementary
school students.
Validity o the secondary and elementary school scales
Scale means.o assess construct validity, demo-
graphic dierences in resilience scale scores were
examined across grade, gender, and racial/ethnic
groups. Secondary school RYDM scale means,
standard deviations, and standardized dierencesacross groups are presented in table B19. With the
exception o caring relationships with peers, 9th and
11th graders report marginally lower environmen-
tal resilience assets than seventh graders. Caring
relationships with peers increases with school grade,
consistent with the notion that adolescents become
more involved with peers (although not necessar-
ily pro-social ones) as they age. Student internal
resilience asset scores do not dier markedly by
grade, although empathy increases with school
grade, and sel-awareness declines with grade.
Gender dierences in resilience assets gener-ally avor emales, who report marginally higher
school and community support and substantially
higher peer caring relationships and exposure to
pro-social peers. Females also report considerably
higher levels o empathy and problem solving.
Tese dierences are consistent with expecta-
tionsgirls ofen have more extensive social
support resources than boys (Colarossi & Eccles,
2000; Crosnoe, Johnson, & Elder, 2004; Frey &
Rthlisberger, 1996) and evidence higher empathy
(Eisenberg & Lennon, 1983).
White students generally report the highest envi-
ronmental assets in each area except or pro-social
peers. Chinese American students report the highest
aliation with pro-social peers. Mexican American
students report the lowest environmental resilience
assets in the school and peer domains and the low-
est meaningul participation in the community.
tAble b17
t- ab o oa hoo a a o a m
im c a aa m
sa
f ()
Sel-efcacy 0.58
r31 i a wk w m w a . . . 0.36r29 i a wk m m. 0.58
r30 i a m i . 0.37
r32 t a ma a i w. 0.50
Empathy 0.57
r33 i a w m . 0.44
r34 i a wa . 0.45
r38 i a wa a k. 0.45
Problem solving 0.52
r35 W i i m ak w. 0.48
r28 i wk m ak/w a m. 0.66
Sel-awareness 0.71r39 t a m . 0.59
r40 i a m m a . 0.48
r41 i a w i wa i . 0.66
Note: Results are based on a sample o 90 ninth-grade respondents rom seven classrooms in two schools in a large urban school district. Two weeks sepa-
rated the rst and second administrations o the survey instruments.
-
8/14/2019 description: tags: REL 2007034 appab
21/29
Appendix b 35
Chinese American students exhibit the lowest envi-
ronmental resilience assets in the home domain and
the lowest reported community support.
White students also report the highest level o
internal assets in sel-ecacy, empathy, andproblem solving. Mexican American and Chinese
American students report the lowest sel-ecacy,
Arican American students exhibit the lowest
empathy scores.
able B20 presents elementary school RYDM scale
means or males and emales. Overall, the gender
dierences or elementary school students are
consistent with those or secondary school students.
Compared with boys, girls report marginally higher
school support, meaningul participation, and
goals/aspirations and substantially higher empathy.
Relationships with other constructs.o urtherassess construct validity, the relationship o each
resilience asset construct to other theoretically
related constructs assessed on the Healthy Kids
Survey was examinedincluding substance
use, violence, harassment, depression, and sel-
reported school grades and truancy. Te relation-
ships o resilience assets to Caliornia Standard-
ized English Language Arts and Mathematics
tAble b18
t- ab o ma hoo a o a m
im c a aa m
sa
f ()
Environmental resilience assets
School support
10 d a . . . a a a ? 0.53
13 d a . . . w a m a? 0.52
11 d a . . . w a j? 0.38
14 d a . . . a a a j? 0.39
Home support 0.70
52 d a a . . . a a wk? 0.56
55 d a a . . . w a m a? 0.65
53 d a a . . . a a a j? 0.53
54 d a a . . . a m wa ? 0.29
Meaningul participation 0.57 9 d mak a a ? 0.39
56 d a m? 0.34
57 d mak a m? 0.44
Internal resilience assets
Empathy 0.70
37 d a w ? 0.55
38 d a w m ? 0.56
Goals/aspirations 0.41
39 d kw w w a m? 0.30
41 d ? 0.49
42 d a a a a ? 0.04
16 d a . . . a ? 0.03
Note: Results are based on a sample o 136 th-grade respondents rom eight classrooms in three schools in a large urban school district. Two weeks sepa-
rated the rst and second administrations o the survey instruments.
-
8/14/2019 description: tags: REL 2007034 appab
22/29
36 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
test scores were examined using data previouslycollected by WestEd.
able B21 shows the relationships between envi-
ronmental resilience assets and theoretically re-
lated constructs or secondary school students. All
but one o the assessed dimensions o environmen-
tal assets are correlated with student substance use.
Students who report high environmental resilience
assets are less likely to report that they engage insubstance use. Te exception is peer caring rela-
tionships, which is weakly correlated with most o
the substance use indicators except substance use
on school property.
Environmental resilience assets are also nega-
tively associated with student depression and
truancy, and positively associated with students
tAble b19
soa hoo ba ma b moah bo
Grade Standardized
diferencea Male Female
Standardized
diferenceaArican
American
Chinese
American
Mexican
American White
Standardized
diferencea7 9 11
Environmental resilience assets
s 2.91
(0.80)
2.74
(0.82)
2.83
(0.81)
0.17 2.77
(0.83)
2.89
(0.80)
0.14 2.84
(0.85)
2.79
(0.76)
2.74
(0.82)
2.96
(0.80)
0.20
s
ma
aa
2.32
(0.86)
2.20
(0.84)
2.21
(0.87)
0.12 2.24
(0.85)
2.26
(0.85)
0.02 2.28
(0.88)
2.22
(0.81)
2.13
(0.84)
2.36
(0.86)
0.20
cmm
3.24
(0.92)
3.15
(0.94)
3.17
(0.95)
0.08 3.11
(0.96)
3.26
(0.90)
0.16 3.19
(0.97)
3.00
(0.96)
3.12
(0.95)
3.44
(0.81)
0.34
cmm
ma
aa
2.93
(1.10)
2.81
(1.12)
2.81
(1.12)
0.11 2.86
(1.11)
2.82
(1.12)
0.05 2.84
(1.13)
2.89
(1.05)
2.51
(1.16)
3.16
(1.01)
0.42
hm 3.45
(0.71)
3.36
(0.74)
3.33
(0.74)
0.14 3.35
(0.76)
3.40
(0.71)
0.07 3.35
(0.80)
3.27
(0.70)
3.34
(0.75)
3.55
(0.64)
0.28
hmma
aa
2.88(0.94)
2.72(0.94)
2.71(0.94)
0.17 2.73(0.96)
2.81(0.93)
0.09 2.76(0.98)
2.65(0.93)
2.72(0.95)
2.94(0.91)
0.23
p a
a
3.10
(0.93)
3.17
(0.91)
3.26
(0.89)
0.15 2.92
(0.96)
3.44
(0.78)
0.58 3.13
(0.95)
3.15
(0.87)
3.07
(0.96)
3.37
(0.84)
0.25
p-a 3.05
(0.85)
2.95
(0.82)
2.96
(0.78)
0.11 2.84
(0.83)
3.13
(0.77)
0.35 2.89
(0.85)
3.18
(0.73)
2.83
(0.84)
3.04
(0.79)
0.33
Internal resilience assets
s-fa 3.24
(0.72)
3.18
(0.72)
3.22
(0.72)
0.08 3.19
(0.74)
3.23
(0.67)
0.06 3.22
(0.76)
3.15
(0.68)
3.12
(0.73)
3.36
(0.64)
0.25
ema 3.10
(0.87)
3.17
(0.82)
3.23
(0.78)
0.13 2.97
(0.88)
3.36
(0.71)
0.46 3.03
(0.90)
3.22
(0.74)
3.10
(0.85)
3.31
(0.77)
0.26
pm 2.87
(0.99)
2.85
(0.96)
2.87
(0.95)
0.02 2.64
(1.00)
3.08
(0.88)
0.45 2.82
(1.00)
2.82
(0.93)
2.81
(0.99)
2.99
(0.94)
0.15
s-awa 3.31
(0.78)
3.21
(0.80)
3.21
(0.79)
0.12 3.22
(0.82)
3.27
(0.76)
0.06 3.27
(0.83)
3.16
(0.78)
3.23
(0.79)
3.32
(0.76)
0.14
a. Standardized diference represents the diference between groups divided by the pooled standard deviation (Cohens d). With more than two groups,
the standardized diference is represented by multiplying Cohens by 2which is generally equivalent to the standardized diference calculated or two
groups (see appendix A).
ns = not statistically signicant rom 0 (p < .05)
Note: Standard deviations in parentheses. Analytic samples consist o 12,000 7th-, 9th-, and 11th-grade respondents sampled rom surveys administered
between spring 2003 and spring 2005.
-
8/14/2019 description: tags: REL 2007034 appab
23/29
Appendix b 37
sel-reported school connectedness and grades.
Te environmental resilience asset scales are less
consistently related to indicators o violence, ha-
rassment, and perceptions o school saety.
Te criterion variablesCaliornia Standards est
(CS) English Language Arts and Mathematics test
scoresare associated with school and communityassets, as well as home support. Te associations
are weak, however, with school support showing
the strongest relationship to test scores. est scores
are not signicantly associated with meaningul
participation in the home environment, peer car-
ing relationships, and pro-social peers.
able B21 suggests that the secondary school RYDM
instrument provides a valid assessment o environ-
mental resilience assets because these constructs
are associated with student substance use, depres-sion, sel-reported grades, truancy, and test scores
in expected ways. Although the correlations with
school connectedness and sel-reported grades are
moderate and have medium eect sizes, the correla-
tions or most o the criterion variables are small.
able B22 shows correlations between internal
resilience assets and the criterion variables or
secondary school students. Te results are similar
to those or environmental assets. With the excep-
tion o standardized test scores, each dimension
o internal resiliencesel-ecacy, empathy,
problem solving, and sel-awarenessis correlated
with most o the considered criterion variables,
which supports construct validity.
able B23 presents correlations between the el-
ementary school resilience assets and the criterion
variables o substance use, aggression, perceived
saety, and sel-reported academic perormance.
Both the environmental resilience and internal
resilience scales are positively associated with most
o the criterion variables, which supports construct
validity. Although the criterion variables are dier-
ent in the two samples, the correlations are stron-
ger or the elementary school resilience instrument
than or the secondary school instrument.
Comparison o current and recommended
measures o resilience assets
ables B24B27 compare the current use o items
to measure resilience assets among secondary
and elementary students with this studys recom-
mended use.
tAble b20
ema hoo ba ma b
A Ma ma
saaz
a
Environmental resilience assets
s 3.32(0.61)
3.28(0.62)
3.37(0.59)
0.15
hm 3.72
(0.44)
3.70
(0.46)
3.74
(0.42)
0.07
Ma aa 2.50
(0.60)
2.46
(0.60)
2.54
(0.60)
0.12
Internal resilience assets
ema 3.01
(0.79)
2.84
(0.84)
3.18
(0.71)
0.42
ga/aa 2.24
(0.35)
2.21
(0.39)
2.26
(0.31)
0.15
a. Standardized diference represents the diference between groups divided by the pooled standard deviation (Cohens d).
ns = not statistically signicant rom 0 (p < .05)
Note: Standard deviations in parentheses. Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring
2003 and spring 2005. Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
24/29
38 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
tAble b21
coao bw oa hoo voma a a o vaab
s
s
ma
aa
cmm
cmm
ma
aa
hm
hm
ma
aa
p a
a
p-a
Substance uselm a 0.17 0.20 0.14 0.25 0.21 0.20 0.02 0.30
30-a a 0.17 0.18 0.15 0.27 0.24 0.23 0.02 0.33
ta a 0.18 0.23 0.25 0.26 0.33 0.29 0.16 0.37
lm a 0.16 0.15 0.07 0.16 0.20 0.20 0.03 0.28
30-a a 0.16 0.14 0.06 0.17 0.20 0.20 0.03 0.28
30-a k 0.13 0.14 0.05 0.16 0.19 0.18 0.03 0.29
A a 0.16 0.10 0.14 0.18 0.23 0.21 0.11 0.27
lm majaa 0.16 0.16 0.08 0.19 0.18 0.19 0.00 0.29
30-a majaa 0.20 0.19 0.10 0.22 0.20 0.20 0.04 0.32
Majaa a 0.19 0.18 0.13 0.22 0.20 0.24 0.14 0.33
Violence
b , , . 0.10 0.06 0.08 0.04 0.09 0.08 0.14 0.15
Aa a 0.08 0.08 0.06 0.08 0.10 0.08 0.07 0.08
pa 0.14 0.07 0.10 0.09 0.14 0.12 0.16 0.26
Ma m a 0.04 0.00 0.02 0.00 0.10 0.09 0.06 0.11
sa jk, mm 0.04 0.03 0.01 0.00 0.10 0.11 0.09 0.14
ha 0.06 0.02 0.06 0.02 0.09 0.07 0.07 0.11
o 0.12 0.10 0.05 0.10 0.16 0.17 0.00 0.33
dama 0.18 0.08 0.14 0.09 0.22 0.17 0.12 0.29
a a 0.12 0.00 0.09 0.01 0.15 0.12 0.09 0.21
Psychological well-being
d 0.16 0.16 0.10 0.15 0.21 0.23 0.05 0.17
School-related actors
s 0.49 0.38 0.30 0.26 0.32 0.29 0.28 0.30
s a (-) 0.24 0.26 0.13 0.29 0.20 0.19 0.13 0.29
ta 0.15 0.15 0.08 0.15 0.23 0.22 0.01 0.27
Standardized test scoresa
cst e laa A 0.15 0.04 0.12 0.12 0.12 0.00 0.02 0.02
cst Mama 0.11 0.04 0.09 0.10 0.09 0.03 0.02 0.01
a. Analytic sample or standardized test score results based on local evaluation data obtained rom a large county in Southern Caliornia. Standardized test
score and school/community asset data were available or 2,898 students, while test score and home and internal asset data were available or 651 students.CST = Caliornia Standards Test
ns = not statistically signicant rom 0 (p < .05)
Note: Analytic sample or substance use, violence, psychological well-being, and school-related actors based on 12,000 7th-, 9th-, and 11th-grade respon-
dents sampled rom HKS surveys administered between spring 2003 and spring 2005.
-
8/14/2019 description: tags: REL 2007034 appab
25/29
Appendix b 39
tAble b22
coao bw oa hoo a a a o vaab
s-fa ema pm s-awa
Substance use
lm a 0.22 0.14 0.17 0.24
30-a a 0.19 0.11 0.14 0.22
ta a 0.20 0.13 0.17 0.21
lm a 0.20 0.14 0.16 0.22
30-a a 0.20 0.13 0.12 0.20
30-a k 0.17 0.17 0.14 0.18
A a 0.22 0.20 0.22 0.25
lm majaa 0.29 0.23 0.14 0.19
30-a majaa 0.20 0.16 0.08 0.14
Majaa a 0.25 0.23 0.20 0.21
Violence
b , , . 0.13 0.14 0.12 0.11
Aa a 0.12 0.01 0.02 0.10
pa 0.16 0.25 0.22 0.13
Ma m a 0.11 0.03 0.02 0.12
sa jk, mm 0.09 0.04 0.01 0.15
ha 0.11 0.05 0.08 0.15
o 0.19 0.14 0.16 0.22
dama 0.23 0.26 0.26 0.23
a a 0.25 0.17 0.21 0.24
Psychological well-being
d 0.26 0.02 0.11 0.30
School-related actors
s 0.33 0.29 0.28 0.32
s a (-) 0.29 0.22 0.21 0.20
ta 0.20 0.15 0.17 0.19
Standardized test scoresa
cst e laa A 0.05 0.09 0.03 0.02
cst Mama 0.05 0.05 0.05 0.02
a. Analytic sample or standardized test score results based on local evaluation data obtained rom a large county in Southern Caliornia. Standardized test
score and school/community asset data were available or 2,898 students, while test score and home and internal asset data were available or 651 students.
CST = Caliornia Standards Test
ns = not statistically signicant rom 0 (p < .05)
Note: Analytic sample or substance use, violence, psychological well-being, and school-related actors based on 12,000 7th-, 9th-, and 11th-grade respon-
dents sampled rom HKS surveys administered between spring 2003 and spring 2005.
-
8/14/2019 description: tags: REL 2007034 appab
26/29
40 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
tAble b23
coao bw ma hoo a a o vaab
ema a ia a
s
hm
Ma
aa ema
ga a
aa
Substance use
lm a 0.25 0.31 0.20 0.20 0.28
lm a 0.26 0.21 0.23 0.18 0.28
lm majaa 0.12 0.15 0.13 0.01 0.14
Aggression victimization
b , , . 0.17 0.12 0.07 0.07 0.14
Ma m a 0.10 0.13 0.12 0.05 0.14
b a a 0.10 0.13 0.12 0.02 0.06
Aggression perpetration
p, , 0.28 0.23 0.22 0.33 0.34
sa ma m 0.22 0.21 0.12 0.22 0.31Perceived saety
a a 0.48 0.30 0.14 0.19 0.41
a a a 0.20 0.20 0.11 0.01 0.25
Academic perormance
s ma 0.17 0.20 0.14 0.09 0.25
ns = not statistically signicant rom 0 (p < .05)
Note: Analytic samples consist o 2,000 th-grade respondents sampled rom surveys administered between spring 2003 and spring 2005. Weighted data.
-
8/14/2019 description: tags: REL 2007034 appab
27/29
-
8/14/2019 description: tags: REL 2007034 appab
28/29
42 MeAsuring resilience And youth developMent: the psychoMetric properties o the heAlthy Kids survey
tAbleb25
currentandreommendedmeasuresofinternalresilienea
ssetsamongseondaryshoolstu
dents
cunuoonayoom
rom
mnuoonayoom
conu
im
conu
im
cooaonan
ommunaon
I
standupormyselwithoutputtingothersdown.
cooaonan
ommunaona
Itemdroppedcross-loadings
I
enjoyworkingtogetherwithotherstud
entsmyage.
Itemdroppedcross-loadings
i
anwokwomonwoa
nononanmn.
s-fay
ianwokwomonwoa
nononanmn.
s-fay
i
anwokoumyom.
ianwokoumyom.
i
anomoniy.
ianomoniy.
t
amanynaiow.
tamanynaiow.
emay
i
awnomonn
u.
emay
iawnomonnu.
i
younanwaooo
ou.
iyounanwaoooou.
i
younanwaoo
annk.
iyounanwaooannk.
pomovn
I
knowwheretogoorhelpwithaproblem.
pomovn
Itemdropped
unctionsdiferently
oremalesandmales
W
nininomonoakw
.
Wnininomonoakw.
i
yowokouomyakn
o
wnaoum.
iyowokouomyakn
ownaoum.
s-awan
t
auoomy.
s-awan
tauoomy.
i
unanmymooann.
iunanmymooann.
i
unanwyiowaio.
iunanwyiowaio.
goaanaaon
I
havegoalsandplansortheuture.
goaanaaona
Itemdropped
unctionsdiferentlyor
AricanAmericans/MexicanAmericans
I
plantograduateromhighschool.
Itemdroppedonlyoneitem
lettomeasureconstruct
I
plantogotocollegeorsomeother
s
choolaterhighschool.
Itemdropped
unctionsdiferently
orChineseAmericans
a.
Constructdroppedbecauseoins
ucientnumberoitems.
Note:Greenitemsdroppedrom
therecommendedmodelbecauseodiferentialit
em
unctioning,
inconsistentloadingpatterns,or
cross-loadingsorbecauseonlyasingleitem
remainstomeasure
construct.
-
8/14/2019 description: tags: REL 2007034 appab
29/29