ORIGINAL ARTICLE The Journal of Nursing Research h VOL. 00, NO.

0, MONTH 2017

Psychometric Evaluation of the Adolescent

Health Promotion Scale in Chile: Differences
by Socioeconomic Status and Gender
Cristian A. Rojas-Barahona1 & Jorge Gaete2* & Esterbina Olivares3 & Carla E. Förster4
Eugenio Chandia5 & Mei-Yen Chen6

1
PhD, Associate Professor, Faculty of Education, Pontificia Universidad Católica de Chile & 2MD, PhD,
Associate Professor, Faculty of Medicine, Universidad de los Andes, Chile; and Postdoctoral Research Fellow,
Centre for Global Mental Health, Department of Population Health, London School of Hygiene & Tropical
Medicine, London, UK & 3MSc, RN, Adjunct Professor, School of Nursing (Campus San Felipe),
Universidad de Valparaı́so, Chile & 4PhD, Associate Professor, Faculty of Education,
Pontificia Universidad Católica de Chile & 5MSc, Instructor Professor, Faculty
of Maths, Pontificia Universidad Católica de Chile & 6PhD, RN, Professor
and Dean, College of Nursing, Chang Gung University of Science
and Technology, Taiwan, ROC.

behaviors of early adolescents in Chile and for comparing results

ABSTRACT
Background: The promotion of healthy behaviors is a relevant
issue worldwide, especially among adolescent populations,
as this is the developmental stage where most unhealthy
behaviors become ingrained.
Purpose: The aim of this study was to analyze the psychometric
properties of the Spanish version of the Adolescent Health
Promotion Scale (AHPS) in a Chilean sample of early adolescents.
Methods: The sample was composed of 1,156 adolescents
aged 10Y14 years from schools in San Felipe, Chile. Item structure
was assessed using exploratory and confirmatory factor analyses;
reliability was measured using Cronbach_s alpha; and differences
in terms of gender, age, and socioeconomic status (SES) were
established using analysis of variance.
Results: The analyses of item structure identified all of the six
original factors (nutrition behaviors, health responsibility, social
support, life appreciation, stress management and exercise
behavior) as significant. However, eight items did not fit the
Chilean population well. Therefore, the AHPS in Chile has been
reduced to 32 items. The Cronbach_s alpha of the 32-item
Chilean AHPS was .95, with the subscale coefficients ranging
from .76 to .94. In addition, female subjects performed better
than male subjects, and individuals of higher SES scored higher
than the middle and lower socioeconomic groups. No differ-
ences on AHPS scores were found in different age groups.
with those from other countries.
KEY WORDS:
healthy behaviors, adolescents, AHPS, psychometric properties.
Introduction
In 2009, the World Health Organization stated that devel-
oping health-promoting behaviors is one of the great chal-
lenges of the 21st century (World Health Organization, 2009).
This statement is based mainly on the evidence that many
causes of death and comorbidities are directly related to
modifiable behaviors such as being overweight or obese and
practicing unhealthy eating habits (Han, Lawlor, & Kimm,
2010), physical inactivity (Lee et al., 2012), and tobacco
and alcohol use.
Worrisomely, in the last several decades, the percentage of
children and adolescents who are overweight or obese has
Accepted for publication: April 10, 2016
*Address correspondence to: Jorge Gaete, MD, PhD, Faculty of
Medicine, Universidad de los Andes, Monseñor Álvaro del Portillo
12455, Las Condes, Santiago, Chile.
Tel: +56 (2) 226181000; E-mail: jgaete@uandes.cl
The authors declare no conflicts of interest.

Conclusions: The AHPS appears to have good psychometric
properties in terms of item structure and reliability. Consistent
with studies carried out in other countries, health promotion
behavioral differences were observed in association with gender
and SES. The results support the Chilean version of the AHPS as
an appropriate instrument for measuring the health promotion
Cite this article as:
Rojas-Barahona, C. A., Gaete, J., Olivares, E., Förster, C. E.,
Chandia, E., & Chen, M. Y. (2017). Psychometric evaluation of
the adolescent health promotion scale in chile: Differences by
socioeconomic status and gender. The Journal of Nursing
Research, 00(0), 00Y00. doi:10.1097/jnr.0000000000000196

Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
The Journal of Nursing Research
risen worldwide (Lobstein et al., 2015). In high-income
countries such as the United States, in a 4-year time span,
the prevalence of being overweight (including obese) among
school children (12Y17 years old) rose by 6% (to 35% in
2003Y2004) and by 3% (to 13% in 2003Y2004), respec-
tively (Lobstein & Jackson-Leach, 2007). Moreover, this
trend holds true for middle- and low-income countries as
well. For example, in Mexico, among adolescents between
12 and 19 years old, the prevalence of being overweight or
obese reached 35% in 2012 (Olaiz-Fernandez et al., 2006).
In Chile, the prevalence of obesity among children and
adolescents between 5 and 17 years old was 27% in female
students and 28% in male students (see Organization for
Economic Cooperation and Development, 2011).
There is good evidence of the beneficial effects of fruit
and vegetables consumption in reducing obesity, metabolic
diseases (diabetes), and the risk of cardiovascular illness
(Vatanparast, Baxter-Jones, Faulkner, Bailey, & Whiting,
2005). However, the recommendation to consume five or
more servings of fruit and vegetables a day is not followed
by most adolescents. For example, in the United States, only
one in five 9thY12th graders eat five or more servings of
fruits and vegetables every day (Eaton et al., 2008).
Tobacco and alcohol use represent another significant
global health problem, especially among adolescents. A
recent report places Chile at the top of the American con-
tinent with the highest tobacco use prevalence among 13- to
15-year-olds (32.8%) and the third highest in the world
(Page & Danielson, 2011). Regarding alcohol use, a 2013
survey found that 35.6% of Chilean adolescents admitted
to consuming alcohol during the previous 1-month periodVa
prevalence that is unchanged from that found in a 2003
survey (Servicio Nacional para la Prevención y Rehabilitación
del Consumo de Drogas y Alcohol - SENDA, 2013).
Most of these habits start early in life and especially
during adolescence (Blum, McNeely, & Nonnemaker, 2002).
Furthermore, their consolidation in this developmental
period compromises health during adulthood (te Velde,
Twisk, & Brug, 2007). Adolescence may be divided into
two stages: early (10Y14 years old) and late (15Y19 years
old; Sawyer et al., 2012). While each stage holds biological,
cognitive, social, and emotional challenges, early adoles-
cents experience several stressors and changes in their daily
lives at the personal and environmental levels that may need
special consideration. To get the most reliable information
from them, their cognitive development, higher interest
in the present rather than the future, and the increasing
influence of peer pressure must be taken into consideration
(Sawyer et al., 2012).
Given the evidence that the promotion of healthy be-
haviors in young populations may significantly reduce the
prevalence of noncommunicable diseases in adulthood and
the rates of mortality and morbidity, there is an urgent need
to have an instrument that is capable of assessing health-
promoting behaviors among early adolescents. The most
widely used tests are the Health-Promoting Lifestyle Profile
Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
Cristian A. Rojas-Barahona et al.
(Walker, Sechrist, & Pender, 1987), the Adolescent Lifestyle
Questionnaire (Gillis, 1997), the Adolescent Lifestyle Profile
(Hendricks, Murdaugh, & Pender, 2006), and the Adoles-
cent Health Promotion Scale (AHPS; Chen, Wang, Yang,
& Liou, 2003). The latter stands out because it has been
translated into several languages and used in a number of
cultures, including the United States, Iran, Taiwan, Turkey, and
Portugal (Aghamolaei & Tavafian, 2013; Chen, James, &
Wang, 2007; Musavian, Pasha, Rahebi, Atrkar Roushan,
& Ghanbari, 2014; Ortabag, Ozdemir, Bakir, & Tosun, 2011;
Temel, Basalan, Yildiz, & Yetim, 2011; Tomás, Queirós, &
Ferreira, 2015), in addition to having been employed to
assess the effects of health education (Hsiao et al., 2005).
The AHPS (Chen et al., 2003) is an instrument that was
designed to evaluate healthy practices in adolescents. This
scale considers the six healthy behavior dimensions of (a)
nutrition behaviors (NBs), (b) social support (SS), (c) life
appreciation (LA), (d) health responsibility (HR), (e) stress
management (SM), and (f) exercise behavior (EB). The AHPS
is an easy-to-use, self-report instrument that has been shown
to be highly reliable (alpha = .962; Chen et al., 2003). Thus,
its application in countries such as Chile may be very useful.
Nevertheless, some adaptations to Chilean culture may be
necessary. As noted by the author of the scale in a study
conducted with adolescents from Taiwan and the United
States, there are relevant cultural differences that should be
taken into account (Chen et al., 2007).
The objective of this study was to analyze the psychometric
properties of the AHPS in a sample of Chilean adolescents.
Methods
Design and Sample
The sample was composed of students from 5th to 8th
grades, from seven schools in San Felipe, Chile (Valparaiso
Region), whose authorities agreed to participate. These schools
were classified into three socioeconomic status (SES) groups:
low, middle, and high. This classification is based on the
criteria that were established by the 2009 National System
for the Measurement of Education Quality (Ministerio de
Educación, 2010), which was constructed based on family
income and the background information provided by stu-
dents’ parents. Three stratification variables (age, gender, and
school SES) were used to select the participants.
Participants
In this study, 1,156 students participated, 625 male students
and 531 female students, aged between 10 and 14 years,
distributed into two age groups: 628 who were e12 years
(M = 10.91 years, SD = 0.67 years) and 528 who were
912 years (M = 12.99 years, SD = 0.61 years). Most students
(88%) did not work after school. Fathers and mothers had
an average of 13.5 and 13.0 years of education, respectively.
It should be pointed out that 25% of the students did not
know the educational level achieved by their parents. In
Psychometric Evaluation of the AHP Scale in Chile
terms of SES, 514 (45%) students were from low-SES
families, 382 (33%) were from middle-SES families, and
260 (22%) were from high-SES families. On the basis of the
most recently available statistics, the socioeconomic distri-
bution of the Chilean households is 40% for low SES, 51%
for middle SES, and 9% for high SES (Instituto Nacional de
Estadı́sticas, 2005). In this study, we oversampled students
from high-SES families to assist with comparisons of SES.
Instrument
The AHPS, developed by Chen et al. (2003, 2007), consists
of 40 items that are scored using a 5-point format that
represents the frequency of the reported behaviors (1 = never,
2 = sometimes, 3 = half of the time, 4 = often, and 5 = always).
The total score is calculated summing the score for each item
(total possible score = 40Y200 points). The items in the scale
are grouped into six subscales: NB, SS, HR, LA, EB, and SM.
The Cronbach’s alpha for the scale is .93 and ranges between
.76 and .88 for the subscales (Chen et al., 2003).
Translation Procedures
The original authors provided English and Spanish versions
of the scale. The Spanish version provided was not culturally
adapted to Latin America. Therefore, the Chilean team asked
two professional translators to produce a new Spanish version
and then asked other two professionals to back-translate the
new Spanish version into English. Both English versions were
compared and discussed with three expert raters, and some
items were adapted with minor changes in wording to pro-
duce a better Spanish version.
Data Collection and Ethical Considerations
The ethical committee of the Faculty of Education at Pontificia
Universidad Católica de Chile approved the study. After
obtaining authorization from the principal of each school,
we asked the pupils’ parents to sign and return an informed
consent, which gave permission for their children to par-
ticipate in the survey. We also asked students for their assent
before collecting data.
The questionnaires were completed individually during
a group session at each educational institution. At the begin-
ning of the session, the person in charge informed the par-
ticipants about the study and noted that participation was
voluntary. It took approximately 40 minutes for the par-
ticipants to complete the questionnaire.
Data Analyses
IBM SPSS Statistics Version 21 (IBM, Inc., Armonk, NY,
USA) and R (R Foundation for Statistical Computing, Vienna,
Austria) statistical packages were used to conduct the analyses.
Descriptive statistical analyses
Means, standard deviations, skewness, and kurtosis (Mardia,
1974) were calculated to assess the performance of the
AHPS items.
Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
VOL. 00, NO. 0, MONTH 2017
Dimensionality
The analyses were conducted in two phases. In the first
phase, an exploratory factor analysis (EFA) was per-
formed to determine the factor structure of the AHPS using the
40 items and the whole sample. Because the items in this
scale are categorical and represent a gradation in Likert-
type rating, a polychoric correlation matrix was used. To
assess the adequacy of the matrix for conducting an EFA,
the KaiserYMeyerYOlkin (KMO) test and Bartlett’s sphe-
ricity test (Hair, Black, Babin, & Anderson, 2010) were
calculated. For factor extraction, we used principal axis
factoring with the promax oblique rotation method be-
cause the variables were not normally distributed. The
selection of the number of factors was based on eigen-
values 9 1.0. The selection of the items into each factor was
based on having a factor loading greater than .32 (Costello
& Osborne, 2005). When variables loaded on more than
one factor (cross-loading), they were carefully examined,
and theoretical meaningfulness was assessed before being
considered for removal from further analyses (Simms &
Watson, 2007). When an item loaded onto more than one
factor, with a larger loading observed on the theoretical
factor, the item was preserved. However, when the loading
was larger on a factor other than the theoretical factor, the
item was removed.
In the second phase, we conducted a confirmatory factor
analysis (CFA). We used the weighted least-squares method
with mean and variance estimators. The techniques used
to evaluate the fit of the common factor and confirmatory
analyses included the comparative fit index (CFI), TuckerY
Lewis index (TLI), root mean square error of approximation
(RMSEA; Hu & Bentler, 1999), and weighted root mean
square residual (WRMR; Hu & Bentler, 1999). The fol-
lowing cutoff values for good model fit have been suggested:
CFI 9 .90, TLI 9 .90, RMSEA G .05, and WRMR G 1.0 (Hu
& Bentler, 1999; Yu, 2002).
Reliability
The internal consistency of the instrument and subscales was
established using Cronbach’s alpha.
Association
An analysis of variance was employed to determine possible
associations with gender, age, SES, and their interaction
effects in the resultant scale and in its subscales. The inter-
action effects analyzed were Gender  Age, Gender  SES,
Age  SES, and Gender  Age  SES.
Results
Descriptive Statistical Analyses
Descriptive statistics for all of the AHPS items are shown
in Table 1.
The Journal of Nursing Research Cristian A. Rojas-Barahona et al.

TABLE 1.
Descriptive Data for All Scale Items
Item Mean SD Skewness Kurtosis

1. I eat three meals daily. 3.25 1.62 j0.16 j1.63

2. I choose foods without too much oil. 3.15 1.52 j0.03 j1.51
3. Include dietary fiber (e.g., fruits or vegetables). 3.58 1.49 j0.53 j1.23
4. Drink at least 1.5 liters of water daily (or 6Y8 cups). 3.38 1.49 j0.27 j1.43
5. Each meal includes five food groups (e.g., bread, meat, milk,I). 3.33 1.41 j0.19 j1.35
6. Eat breakfast daily.a 3.89 1.51 j0.85 j0.98
7. I speak up and share my feelings with others. 2.81 1.50 .27 j1.40
8. I care about other people. 3.46 1.38 j0.33 j1.26
9. I talk about my concerns with others. 2.79 1.46 .24 j1.34
10. Make an effort to smile or laugh every day.a 3.82 1.42 j0.77 j0.92
11. Enjoy keeping in touch with relatives.a 4.18 1.28 j1.32 0.32
12. Make an effort to have good friendships.a 4.36 1.18 j1.75 1.75
13. Talk about my troubles to others. 2.80 1.50 .23 j1.41
14. Read food labels when I shop. 2.82 1.57 .23 j1.50
15. I watch my weight. 3.26 1.50 j0.15 j1.46
16. Discuss my health concerns with a doctor or nurse. 2.65 1.50 .37 j1.34
17. Observe my body at least monthly. 3.21 1.55 j0.14 j1.53
18. Brush my teeth and use dental floss after meals.a 3.77 1.35 j0.70 j0.88
19. Wash hands before meals. 4.20 1.27 j1.40 0.55
20. Read health information. 3.25 1.50 j0.18 j1.44
21. Make an effort to choose foods without preservatives. 3.06 1.50 j0.01 j1.44
22. Make an effort to like myself. 4.13 1.37 j1.28 0.11
23. Make an effort to feel happy and content. 4.30 1.21 j1.53 0.93
24. I usually think positively. 4.07 1.24 j1.12 j0.01
25. Make an effort to understand my strengths and weaknesses and accept them. 3.98 1.32 j1.03 j0.28
26. Make an attempt to correct my defects. 3.98 1.30 j1.03 j0.26
27. Make an effort to know what_s important for me. 4.09 1.27 j1.19 0.13
28. Make an effort to feel interesting and challenged every day. 3.96 1.31 j0.98 j0.36
29. Make an effort to believe that my life has purpose. 4.22 1.25 j1.43 0.70
30. Perform stretching exercise daily. 3.41 1.45 j0.32 j1.33
31. Exercise rigorously 30 minutes at least 3 times per week. 3.32 1.51 j0.21 j1.47
32. Participate in physical fitness class at school weekly.a 4.31 1.24 j1.59 1.08
a
33. Warm up before rigorous exercise. 3.89 1.41 j0.91 j0.67
34. Make an effort to stand or sit up straight.a 3.82 1.35 j0.76 j0.81
35. Make an effort to spend time daily for relaxation. 3.80 1.40 j0.79 j0.80
36. Make an effort to determine the source of my stress. 3.70 1.43 j0.68 j0.99
37. Make an effort to watch my mood changes. 3.74 1.38 j0.69 j0.90
38. Sleep for 6Y8 hours each night. 3.89 1.37 j0.86 j0.72
39. Make schedules and set priorities. 3.61 1.47 j0.55 j1.18
40. I try not to lose control when things happen that are unfair. 3.69 1.41 j0.67 j0.95
a
Items deleted from the final version.

Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
Psychometric Evaluation of the AHP Scale in Chile
Dimensionality
Phase 1: exploratory factor analysis
The KMO value was .97, whereas the result of Bartlett’s
sphericity test was significant (2 2 = 564.72, df = 78, p G .000).
The EFA displays a seven-factor solution with eigenvalues
9 1.0 that explained 61% of the variance, split into 15%,
10%, 9%, 7%, 7%, 7%, and 6%, respectively. Eight items
were removed for the following reasons: Three items loaded
onto two factors, and the highest loading was on a factor
other than the theoretical factor (6, 18, and 32); one item
(10) did not display any high loadings (all loadings below
.32); and four items displayed loadings on a single factor
other than the theoretical factor (11, 12, 19, and 34).
Afterward, and with these eight items removed, the EFA
was run again, resulting in a six-factor model (eigenvalues 9
1.0) with indicators that approximated the theoretical
structure of the original instrument, with 32 items explaining
62% of the variance, split into 19%, 11%, 9%, 9%, 8%,
and 6%, respectively. In the revised instrument, LA
explained 19% of the variance and was composed of Items
22, 23, 24, 25, 26, 27, 28, and 29; SM explained 11% of the
variance and was composed of Items 35, 36, 37, 38, 39, and
40; SS explained 9% of the variance and was composed of
Items 7, 8, 9, and 13; NB explained 9% of the variance and
was composed of Items 1, 2, 3, 4, and 5; HR explained 8%
of the variance and was composed of Items 14, 15, 16, 17,
20, and 21; and EB explained 6% of the variance and was
composed of Items 30, 31, and 33. The distribution of the
items in each factor is presented in Table 2.
The correlations between constructs ranged from .42 to
.84 (see Table 3).
Phase 2: confirmatory factor analysis
The EFA revealed that the questionnaire was multidimen-
sional (six factors), and the best item structure included 32
items. Subsequently, we performed a CFA to estimate the fit of
the six-factor model. Standardized factor loadings showed
that all of the items loaded well onto each latent construct (see
Table 2). The fit indices were good, considering the acceptance
intervals of the CFI (.96), TLI (.96), RMSEA (.04), and
WRMR (1.60) values (see Table 4).
Reliability
The internal consistency for the 32 items of the instrument
as determined by Cronbach’s alpha was .95. The alpha
coefficients of the six original subscales retained internal
consistencies of between .75 and .94 (Table 5).
Descriptive Statistical Values of the
Revised Scale
The range of total possible scores for the full scale is
32Y165. The total mean score was 112.58 (SD = 45.59).
The scores of distributions for each subscale and for the
full instrument are close to normal distributions (Table 5).
Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
VOL. 00, NO. 0, MONTH 2017
Association
Gender
Differences were observed in the 32-item scale (F(1147, 1)
= 5.07, p G .05), as the average for female students (M =
126.65, SD = 29.99) was higher than for male students
(M = 121.78, SD = 30.68). The analysis of the subscales
showed that, for some, the average for female students
was higher than for male students, with significant differ-
ences in LA (F(1147, 1) = 7.06, p G .05; female students:
M = 35.05, SD = 6.34; male students: M = 33.08, SD =
7.39), HR (F(1147, 1) = 10.03, p G .05; female students:
M = 19.06, SD = 6.62; male students: M = 17.57, SD =
6.67), and SS (F(1147, 1) = 21.60, p G .001; female students:
M = 16.74, SD = 5.56; male students: M = 14.77, SD = 5.70).
In EB, male students had a higher average than female
students (F(1147, 1) = 5.60, p G .05; male students: M =
14.74, SD = 4.52; female students: M = 14.04, SD = 4.61).
There were no differences in NB (F(1147, 1) = .078, p = .78)
or SM (F(1147, 1) = 1.39, p = .24).
Socioeconomic status
Differences were observed in the full scale (F(1147, 2) =
5.63, p G .05), with the high status group displaying the
highest average (M = 129.88, SD = 25.20). No significant
differences were present between the other two levels (middle:
M = 121.32, SD = 32.40; low: M = 123.05, SD = 31.02).
Regarding the subscales, differences were observed in NB
(F(1147, 2) = 4.78, p G .05; high: M = 17.14, SD = 8.73;
middle: M = 13.63, SD = 11.96; low: M = 14.85, SD =
11.03), LA (F(1147, 2) = 4.41, p G .05; high: M = 37.14,
SD = 8.73; middle: M = 33.63, SD = 11.96; low: M =
34.85, SD = 11.03), SS (F(1147, 2) = 18.78, p G .001;
high: M = 27.14, SD = 8.73; middle: M = 23.63, SD =
11.96; low: M = 24.85, SD = 11.03), and EB (F(1147, 2) =
5.45, p G .001; high: M = 7.14, SD = 8.73; middle: M = 3.63,
SD = 11.96; low: M = 4.85, SD = 11.03). No differences
were observed in HR (F(1147, 2) = .92, p = .40) or SM
(F(1147, 2) = 1.85, p = .16).
Age
No differences were observed in the average total scores,
regardless of participant age (F(1211, 3) = 1.37, p = .24).
No differences between age groups were found for any of
the subscales.
Interaction effects
No interaction effects were observed between the variables
(Gender  Age, Gender  SES, Age  SES, and Gender 
Age  SES), either in the full scale or in the subscales of
the 32-item model (p 9 .05).
Discussion
This study focuses on the psychometric properties of the
Spanish version of the AHPS for Chilean adolescents and
its implications and limitations for clinical use.
The Journal of Nursing Research Cristian A. Rojas-Barahona et al.

The criteria for adequacy of EFA were found to be
similar to the original version of the AHPS, with a KMO
of .97 (Chen et al., 2003; KMO = .94) and a significant
result from Bartlett’s test of sphericity (p G .000 in both
cases). In general, most items behaved according to
expectations in terms of comprehension and response
categories. However, the grouping of the items that was
observed in this study differed from that proposed by
Chen et al. (2003) for an adolescent population in Taiwan.
Initially, the grouping of the items with eigenvalues above
1 indicated the existence of seven factors (accounting for
61% of the variance), which differs significantly from the
findings published by Chen et al. (2003), who proposed
that six latent factors explained 51.14% of the variance
with 40 items. However, after eight items were removed
because of inadequate performance (cross-loading and lack

TABLE 2.
Standardized Weights for Each Factor of the Six-Factor, 32-Item Model
% Variance
Item LA SM SS NB HR EB Explained

22. Make an effort to like myself. .82 j.06 j.04 .08 .03 .01 .19
23. Make an effort to feel happy and content. .92 j.07 .04 .03 j.05 .03
24. I usually think positively. .94 .01 j.02 j.11 .00 .05
25. Make an effort to understand my strengths, weaknesses .83 .11 j.03 .01 .00 j.05
and accept them.
26. Make an attempt to correct my defects. .72 .05 j.02 j.01 .18 j.03
27. Make an effort to know what_s important for me. .83 .05 j.05 .05 .03 j.03
28. Make an effort to feel interesting and challenged every day. .82 .02 .03 j.10 .06 .03
29. Make an effort to believe that my life has purpose. .85 .09 .02 .01 j.06 j.04
35. Make an effort to spend time daily for relaxation. .12 .60 .03 j.07 j.04 .19 .11
36. Make an effort to determine the source of my stress. .00 .81 .06 j.12 .08 .01
37. Make an effort to watch my mood changes. .02 .81 .02 j.01 .01 .00
38. Sleep for 6Y8 hours each night. .01 .69 j.03 .14 j.07 j.05
39. Make schedules and set priorities. j.05 .74 j.07 .07 .13 j.04
40. I try not to lose control when things happen that are unfair. .12 .67 j.03 .09 j.03 j.09
7. I speak up and share my feelings with others. .05 j.01 .75 .08 .00 j.01 .09
8. I care about other people. .13 j.02 .48 .25 .01 j.04
9. I talk about my concerns with others. j.09 .01 .96 j.04 .03 .01
13. Talk about my troubles to others. .00 .02 .84 j.19 .13 .05
1. I eat three meals daily .09 .03 .21 .53 j.26 j.05 .09
2. I choose foods without too much oil. j.10 j.03 j.09 .67 .16 .06
3. Include dietary fiber (e.g., fruits or vegetables). j.07 .00 j.02 .77 .11 .01
4. Drink at least 1.5 liters of water daily (or 6Y8 cups). j.01 j.03 j.06 .61 .08 .10
5. Each meal includes five food groups (e.g., bread, meat, milk,I) .05 .04 j.09 .68 .05 j.04
14. Read food labels when I shop. .01 .03 .00 .06 .62 .03 .08
15. I watch my weight. .06 j.15 j.02 .33 .51 .08
16. Discuss my health concerns with a doctor or nurse. j.12 .07 .14 .04 .69 j.05
17. Observe my body at least monthly. .05 .05 .14 .09 .51 j.11
20. Read health information. .10 .07 j.03 j.07 .66 .05
21. Make an effort to choose foods without preservatives. .02 .01 j.02 .10 .67 j.02
30. Perform stretching exercise daily. .05 .01 j.04 .06 .00 .85 .06
31. Exercise rigorously 30 minutes at least 3 times per week. j.07 .01 .08 .04 .01 .84
33. Warm up before rigorous exercise. .07 .25 .03 .07 j.02 .42
Total .62

Note. LA = life appreciation; SM = stress management; SS = social support; NB = nutrition behavior; HR = health responsibility; EB = exercise behavior.

Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
Psychometric Evaluation of the AHP Scale in Chile
TABLE 3.
Estimated Correlations Among the Six
Constructs of the AHPS (Final Model)
LA SM NB HR SS
LA
SM .62*
NB .76* .84*
HR .42* .55* .45*
SS .67* .76* .69* .57*
EB .65* .55* .63* .46* .81*
Note. AHPS = Adolescent Health Promotion Scale; LA = life appreciation; SM =
stress management; NB = nutrition behavior; HR = health responsibility; SS =
social support; EB = exercise behavior. *p G .001.
of theoretical meaningfulness), we found a six-factor model
with 32 items that explained 62% of the variance.
Few other studies have explored the item structure of the
AHPS. For instance, in Portugal, among 1,213 adolescents,
a factor analysis of the scale revealed a six-factor structure
using all 40 items, with a total variance of 45.6% (Cardoso
et al., 2015). However, when we examined the factor load-
ings presented in that publication, several items did not
load appropriately on some of the factors. For example,
four of the six items (2, 3, 4, and 5) in the NB factor loaded
at less than .32. Moreover, three of the eight items (10, 11,
and 12) in the SS factor were reported with low factor
loadings (G.32). It is interesting to note that six of the eight
items (10, 11, 12, 18, 19, and 34) that were removed in this
study, which focused on a population of Chilean adoles-
cents, displayed very low loadings in the Portuguese study
(Tomás et al., 2015), suggesting that these perform poorly
in several settings and that their removal is justified.
In all of the countries where the AHPS has been used, the
order of the factors differed from each other in terms of the
amount of variance explained. This may be considered as a
cultural difference; that is, adolescents from different coun-
tries assign different levels of importance to the same aspects.
For instance, in Taiwan, the main factor was SS (28.8%),
and the least important factor was EB (3.5%). In Chile, we
found that the most important factor was LA (19.0%) and
the least important factor was EB (6.0%). In Portugal, LA
(10.1%) and NB (4.3%) were the most and least important
factors, respectively, whereas in Turkey, these factors were
TABLE 4.
Goodness-of-Fit Indices for the AHPS (Phase 2)
# 2 (df) p CFI
Six-factor model 1078.708 (449) G.001
Note. AHPS = Adolescent Health Promotion Scale; CFI = comparative fit index; RMSEA = root mean square error of approximation; TLI = TuckerYLewis
index; WRMR = weighted root mean square residual.
Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
VOL. 00, NO. 0, MONTH 2017
LA (17.0%) and NB (3.4%), respectively. Thus, when plan-
ning promotion interventions, adapting content according
to the weight that various types of behaviors have in the
overall concept of health promotion among adolescents is
an important consideration. It is also necessary to stress
that, in all of the studies mentioned, including the original
study (Chen et al., 2003), the LA factor explained a large
percentage of the overall variance. It appears to be a factor
that must be regarded as essential for health promotion in
adolescents, regardless of their culture.
When performing the EFA, we decided to use an oblique
method to rotate assumed correlations between factors. We
found support for this assumption, considering that most of
the correlation coefficients for the six-factor model were
higher than .5 (all with ps G .001). Two correlations had
values 9 .8: SM with NB and SS with EB. The first as-
sociation may be due to the fact that both factors involve
making an effort to achieve daily routines. The second as-
sociation may be connected to the fact that, at least in
Chile, adolescents tend to prefer physical exercise in group
settings (e.g., playing football) rather than individually.
The reliability of the scale with 32 items displayed a
Cronbach’s alpha of .95, whereas the subscales ranged
from .75 to .94, which is considered very good (Hogan,
2013). These values are higher than those reported by Chen
et al. (2003). The internal consistency values for all sub-
scales were good (9.75). The SS scale had remarkable internal
consistency, with four items and a coefficient of .85. The EB
scale is also noteworthy for this reason, with three items
and a coefficient of .80.
With respect to the removed items, several reasons may
explain the lack of association with the theoretical subscale.
For example, Item 6 of the NB subscale ‘‘Eat breakfast
daily’’ may not be considered as part of that factor because
Chilean people may not think that breakfast is an impor-
tant meal, unlike lunch. For Items 10, 11, and 12, which
are part of the SS subscale (10: ‘‘Make an effort to smile
or laugh every day’’; 11: ‘‘Enjoy keeping in touch with
relatives’’; 12: ‘‘Make an effort to have good friendships’’),
the reason may be that these three items may not indicate
actual actions to the respondents as the other items do but
rather be a proposition for making an effort to perform
these actions. It is important to address this issue in other
studies, as the one conducted in Portugal found very low
loadings for these three items. The opposite may be the case
for Items 18 and 19, which are originally associated with
the HR subscale (18: ‘‘Brush my teeth and use dental floss
TLI RMSEA (90% CI) WRMR
.96 .96 .04 (.03, .04) 1.60
The Journal of Nursing Research Cristian A. Rojas-Barahona et al.

TABLE 5.
Descriptive Data and Alpha Coefficients for Each Subscale and the Full Scale
Scale Number of Item Mean SD Min Max Asymmetry Kurtosis Alpha

Nutritional behaviors 5 16.69 7.53 5 25 j0.24 j1.53 .75

Social support 4 11.86 5.84 5 25 0.10 j1.352 .85
Life appreciation 8 32.73 10.23 8 40 j1.20 0.12 .94
Stress management 6 22.43 8.46 6 30 j0.71 j0.923 .85
Health responsibility 6 18.25 9.12 6 30 0.02 j0.145 .83
Exercise behavior 3 10.86 4.37 3 15 j0.48 j1.15 .80
Full scale 32 112.58 45.59 32 165 j0.50 j0.916 .95

Note. Min = minimum; Max = maximum.

after meals’’; 19: ‘‘Wash hands before meals’’), as these
items are the only two that refer to specific, everyday actions.
In addition, both of these actions are regarded as basic be-
havioral habits in Chile that are learned from an early age
and thus may not be considered to be a health responsibil-
ity matter but rather a matter of good manners. In the case
of Items 32 and 34, which are associated with the EB
subscale (32: ‘‘Participate in physical fitness class at school
weekly’’; 34: ‘‘Make an effort to stand or sit up straight’’),
different reasons may be involved. On the one hand, Item
32 loads on the theoretical factor as well as on another
factor (more on the latter). The reason for this may be that
physical education lessons do not depend on the student
and are often suspended because other academic activities
are prioritized. On the other hand, in the case of Item 34,
pupils may not make a connection between paying atten-
tion to one’s posture and one’s EBs and SM (the highest
loading is on this latter factor), suggesting that this item
may be a factor on its own, which is supported by the vari-
ability of loadings for this item as shown in several studies
discussed previously.
Regarding gender, differences may be observed in three
of the six subscales, where the average is higher for female
students than for male students. These results are consis-
tent with a study by Sjøberg and Schreiner (2010), which
investigated more than 30 European and Asian countries.
They found adolescent girls to be more interested in body
and health issues than boys, with boys earning higher scores
than girls only for the EB factor. These results may be ex-
plained because girls may prefer not to exercise because
of misconceptions about the meaning of ‘‘femininity’’ or
the perception that sweat spoils their appearance (Dwyer
et al., 2006).
Differences also emerged when analyzing the sample by
SES. High-SES respondents gave average scores that were
significantly higher than those of the other two groups in
four of the six subscales. This may reflect their greater
sociocultural capital, which entails more awareness of their
health and well-being, better access to knowledge, and more
resources to adopt healthy behaviors. Therefore, it seems
necessary to take this aspect into account when applying
this scale and interpreting the results of this study.
Finally, the age of the respondents did not seem to in-
fluence their answers. The age range of the participants in
this study (10Y14 years) was more limited than that in the
study conducted by Chen et al. (2003), which included in-
dividuals who were between 12 and 22 years old, with an
average age of 16 years. Our results support the idea that
early adolescents, as a group, share similar characteristics.
Responses on this scale reflect the behavior of a sample of
students from central Chile, which may differ in the case of
students from other parts of the country. Chen et al. (2007)
found differences between the behavior of Taiwanese and
American adolescents, a situation that may also apply to
different regions within Chile.
This study was affected by several limitations. Cultural
differences may exist between students from different geo-
graphical regions within Chile. Our study was conducted in
a small city in the center of the country. Therefore, it will be
important to explore the performance of this scale in other
parts of the country. In addition, we did not perform a CFA
on a separate sample of adolescents, which is sometimes
recommended to test the generated hypotheses. Moreover,
we were unable to provide information regarding how this
scale performs in adolescents who are older than 14 years.
Further research is recommended to address these limitations.
The strengths of this study include the addition of analyses
that consider gender and SES differences, two variables that
influence healthy behaviors in adolescents directly. In ad-
dition, this study allows researchers to contrast their results
with this reference point.
Conclusions and Recommendations
The AHPS is a valid and reliable instrument for measuring
the health-promoting behaviors of Chilean adolescents. Eight
items were eliminated from this scale to maintain the original
six-factor structure (NBs, HR, SS, LA, SM, and EB). This
change from the original scale is justified because of the
different cultural contexts of the original and current studies.

Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
Psychometric Evaluation of the AHP Scale in Chile
Finally, in interpreting test results, the gender ratio and
SES status of participants must be taken into account, as
indicated by studies that have been conducted in various
countries and cultural settings (Chen et al., 2007; Sjøberg &
Schreiner, 2010).
The promotion of healthy behaviors is of global impor-
tance. Therefore, having an instrument that is adapted to
specific cultural and linguistic settingsVin this case, SpanishV
and provides psychometric indices that are consistent
with the original scale makes it possible to obtain data from
school-aged adolescents to perform follow-up studies,
focused interventions, diagnoses to help set public policy,
and so on. Concurrently, having an internationally valid
and reliable instrument facilitates the development of
comparative studies.
References
Aghamolaei, T., & Tavafian, S. S. (2013). Health behaviors of a
sample of adolescents in Bandar Abbas, Iran. International
Journal of High Risk Behaviors & Addiction, 2(1), 34Y38.
doi:10.5812/ijhrba.8842
Blum, R. W., McNeely, C., & Nonnemaker, J. (2002). Vulnerability,
risk, and protection. The Journal of Adolescent Health, 31(1,
Suppl.), 28Y39. http://dx.doi.org/10.1016/S1054-139X(02)00411-1
Cardoso, C., Pina, P. J., Rodrı́guez, T. J. (2015). Health-promoting
behaviors: psychometric properties of an assessment tool.
Text Context Nursing, Florianópolis, 24(1), 22Y9. http://dx.doi.
org/10.1590/0104-07072015000700014
Chen, M. Y., James, K., & Wang, E. K. (2007). Comparison of
health-promoting behavior between Taiwanese and American
adolescents: A cross-sectional questionnaire survey. Interna-
tional Journal of Nursing Studies, 44(1), 59Y69. doi:10.1016/
j.ijnurstu.2005.11.015
Chen, M. Y., Wang, E. K., Yang, R. J., & Liou, Y. M. (2003).
Adolescent health promotion scale: Development and psy-
chometric testing. Public Health Nursing, 20(2), 104Y110. doi:10.
1046/j.1525-1446.2003.20204.x
Costello, A. B., & Osborne, J. W. (2005). Best practices in
exploratory factor analysis: Four recommendations for getting
the most from your analysis. Practical Assessment Research
& Evaluation, 10(7), 173Y178.
Dwyer, J. J., Allison, K. R., Goldenberg, E. R., Fein, A. J.,
Yoshida, K. K., & Boutilier, M. A. (2006). Adolescent girls’
perceived barriers to participation in physical activity. Ado-
lescence, 41(161), 75Y89.
Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Ross, J.,
Hawkins, J., I Centers for Disease Control and Prevention
(2008). Youth risk behavior surveillanceVUnited States, 2007.
Morbidity and Mortality Weekly Report, 57(4), 1Y131. Re-
trieved from http://www.cdc.gov/mmwr/preview/mmwrhtml/
ss5704a1.htm
Gillis, A. J. (1997). The Adolescent Lifestyle Questionnaire:
Development and psychometric testing. The Canadian Journal
of Nursing Research, 29(1), 29Y46.
Hair, J. F., Black, W. C., Babin, B. J. & Anderson, R. E. (2010).
Multivariate data analysis. Upper Saddle River, NJ: Pearson.
Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
VOL. 00, NO. 0, MONTH 2017
Han, J. C., Lawlor, D. A., & Kimm, S. Y. (2010). Childhood obesity.
Lancet, 375(9727), 1737Y1748. doi:10.1016/S0140-6736(10)
60171-7
Hendricks, C., Murdaugh, C., & Pender, N. (2006). The Adoles-
cent Lifestyle Profile: Development and psychometric char-
acteristics. Journal of National Black Nurses’ Association,
17(2), 1Y5.
Hogan, T. (2013). Psychological testing: A practical introduction.
New York, NY: Wiley.
Hsiao, Y. C., Chen, M. Y., Gau, Y. M., Hung, L. L., Chang, S. H.,
& Tsai, H. M. (2005). Short-term effects of a health promotion
course for Taiwanese nursing students. Public Health Nurs-
ing, 22(1), 74Y81. doi:10.1111/j.0737-1209.2005.22111.x
Hu, L.-T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in
covariance structure analysis: Conventional criteria versus new
alternatives. Structural Equation Modeling: A Multidisciplinary
Journal, 6(1), 1Y55. doi:10.1080/10705519909540118
Instituto Nacional de Estadı́sticas. (2005). Socioeconomic classi-
fication of Chilean households. Santiago, Chile: Instituo
Nacional de Estadı́sticas. (Original work in Spanish)
Lee, I. M., Shiroma, E. J., Lobelo, F., Puska, P., Blair, S. N.,
Katzmarzyk, P. T., & Lancet Physical Activity Series Working
Group. (2012). Effect of physical inactivity on major non-
communicable diseases worldwide: An analysis of burden of
disease and life expectancy. Lancet, 380(9838), 219Y229. doi:10.
1016/S0140-6736(12)61031-9
Lobstein, T., & Jackson-Leach, R. (2007). Child overweight and
obesity in the USA: Prevalence rates according to IOTF defini-
tions. International Journal of Pediatric Obesity, 2(1), 62Y64.
doi:10.1080/17477160601103948
Lobstein, T., Jackson-Leach, R., Moodie, M. L., Hall, K. D.,
Gortmaker, S. L., Swinburn, B. A., I McPherson, K. (2015).
Child and adolescent obesity: Part of a bigger picture. Lancet,
385(9986), 2510Y2520. doi:10.1016/S0140-6736(14)61746-3
Mardia, K. (1974). Applications of some measures of multivar-
iate skewness and kurtosis in testing normality and robust-
ness studies. Sankhyā: The Indian Journal of Statistics, Series
B, 36(2), 115Y128.
Ministerio de Educación. (2010). SIMCE: Results 2009. Santiago,
Chile: Gobierno de la República de Chile, MINEDUC. (Original
work in Spanish)
Musavian, A. S., Pasha, A., Rahebi, S. M., Atrkar Roushan, Z., &
Ghanbari, A. (2014). Health promoting behaviors among
adolescents: A cross-sectional study. Nursing and Midwifery
Studies, 3(1), e14560. doi:10.17795/nmsjournal14560
Olaiz-Fernandez, G., Rivera-Donmarco, J., Shamah-Levy, T.,
Rojas, R., Villalpando-Hernandez, S., Hernandez-Avila, M., &
Sepulveda-Amor, J. (2006). National survey on health and
nutrition 2006 (Vol. MX). Cuernavaca, Mexico: Instituto Nacional
de Salud Pública. (Original work in Spanish)
Organization for Economic Cooperation and Development. (2011).
Health at a glance 2011: OECD indicators. Retrieved from http://
dx.doi.org/10.1787/health_glance- 2011-1-en
Ortabag, T., Ozdemir, S., Bakir, B., & Tosun, N. (2011). Health
promotion and risk behaviors among adolescents in Turkey.
The Journal of School Nursing, 27(4), 304Y315. doi:10.1177/
1059840511408322
Page, R. M., & Danielson, M. (2011). Multi-country, cross-national
comparison of youth tobacco use: Findings from global
The Journal of Nursing Research
school-based health surveys. Addictive Behaviors, 36(5),
470Y478. doi:10.1016/j.addbeh.2011.01.008
Sawyer, S. M., Afifi, R. A., Bearinger, L. H., Blakemore, S. J.,
Dick, B., Ezeh, A. C., & Patton, G. C. (2012). Adolescence: A
foundation for future health. Lancet, 379(9826), 1630Y1640.
doi:10.1016/S0140-6736(12)60072-5
Servicio Nacional para la Prevención y Rehabilitación del
consumo de drogas y alcohol - SENDA. (2013). Décimo
Estudio Nacional de Drogas en Población Escolar de Chile,
8-sico a 4- medio. Santiago, Chile: Ministerio de Interior de Chile.
Simms, L. J., & Watson, D. (2007). The construct validation
approach to personality scale construction. In R. W. Robins,
R. C. Fraley, & R. F. Krueger (Eds.), Handbook of research
methods in personality psychology (pp. 240Y258). London,
England: Guilford Press.
Sjøberg, S., & Schreiner, C. (2010). The ROSE project. Overview
and key findings. Retrieved from http://eacea.ec.europa.eu/
education/eurydice/
te Velde, S. J., Twisk, J. W., & Brug, J. (2007). Tracking of fruit
and vegetable consumption from adolescence into adult-
hood and its longitudinal association with overweight. The
British Journal of Nutrition, 98(2), 431Y438. doi:10.1017/
S0007114507721451
10
Copyright © 2017 Taiwan Nurses Association. Unauthorized reproduction of this article is prohibited.
Cristian A. Rojas-Barahona et al.
Temel, A. B., Basalan, F., Yildiz, S., & Yetim, D. (2011). The
reliability and validity of adolescent health promotion scale in
Turkish community. Journal of Current Pediatrics, 9(1), 14Y22.
Tomás, C. C., Queirós, P. J. P., & Ferreira, T. D. J. R. (2015).
Health-promoting behaviors: Psychometric properties of an
assessment tool. Texto & Contexto-Enfermagem, 24(1), 22Y29.
http://dx.doi.org/10.1590/0104-07072015000700014
Vatanparast, H., Baxter-Jones, A., Faulkner, R. A., Bailey, D. A.,
& Whiting, S. J. (2005). Positive effects of vegetable and
fruit consumption and calcium intake on bone mineral
accrual in boys during growth from childhood to adolescence:
The University of Saskatchewan Pediatric Bone Mineral
Accrual Study. The American Journal of Clinical Nutrition, 82(3),
700Y706.
Walker, S. N., Sechrist, K. R., & Pender, N. J. (1987). The
Health-Promoting Lifestyle Profile: Development and psy-
chometric characteristics. Nursing Research, 36(2), 76Y81.
World Health Organization. (2009). Milestones in health promotion
statements from global conferences. Geneva, Switzerland: Author.
Yu, C. Y. (2002). Evaluating cutoff criteria of model fit indices
for latent variable models with binary and continuous out-
comes (Unpublished doctoral dissertation). University of
California, Los Angeles.