CHAPTER 3
METHODOLOGY
The current study attempted to determine the relationship between adversities and resilience
through the mediating variables (protective and compensatory factors) by examining the effects
of adversities on children's resilience and the following behavioral outcomes (emotional
awareness, self-esteem, and academic achievement). Well-being and resilience were protective
factors, whereas parental support and adult mentoring were compensatory factors.
The methodology is explained in the chapter under various section headings. Research design is
covered in Section 3.1. Section 3.2, titled Sample, explains the study's participants and sampling
procedure. The details of the tools used to measure the study's variables were described in
Section 3.3, Tools. Section 3.4 of the Procedure describes the researcher's exact steps to
complete this study. Section 3.5, titled Statistical Analysis, introduces and explains the statistical
computations in the study to manage the quantified data.
Further, this chapter also includes one figure explaining the sampling frame. Figure 1 describes
the sampling frame of students from districts of Punjab, India.
3.1 Research Design
This study used a descriptive research design. The exogenous variable is adversity. Well-being
and hardiness (protective factors) and parental support and adult mentoring (compensatory
factors) are the endogenous (mediating) variables, and resilience and behavioral outcomes such
as self-esteem, emotional awareness, and academic achievement are the endogenous variables.
54
Questionnaires were carried out independently on each subject to measure the above variables.
The art technique was also used to decrease the element of social desirability.
3.2 Sample
The target group for the present study is children facing adversities. Data was collected from 400
students (B: 200; G: 200) of the age range 9 to 12 years (M = 10.59, SD = 1.07) from the grade
5th to 7th (M = 5.63, SD = 1.07) of public and private schools across districts of Punjab, India:
Jalandhar, Patiala, Sangrur, and Ludhiana. The districts were randomly selected from 22 districts
of Punjab; out of the 453 schools of each four districts, only 100 schools (Jalandhar: 33, 33%;
Patiala: 30, 30%; Sangrur: 10, 10%; & Ludhiana: 27, 27%) gave consent for the data collection,
out of which four private and four public schools were randomly selected from the four districts.
The data collected was primary, qualitative, and quantitative. The final sample comprised 50
students from each school, with an equal number of boys (25) and girls (25). The sample of the
parents who filled out the adverse childhood experiences questionnaire on behalf of the children
was 400 (M = 37.73 yrs., SD = 3.23). All the parents were in 12th grade and above. The
questionnaire was filled by 80% of mothers and 20% of fathers. The socio-economic status of the
parents varied from Lower Income Group (30%), Middle Income Group (45%), and Upper
Income Group (25%). Figure 4 presents a graphic view of the sampling procedure and the final
sample selected for the study.
55
Schools selected from Districts of
Punjab
Jalandhar Patiala Sangrur Ludhiana
27% 33%
10%
30%
Figure 4. The selection of the schools on the basis of 22 districts of Punjab, India
DISTRICTS OF PUNJAB
22
DISTRICTS
DISTRICTS
1 2 3 4
A B A B A B A B
25 25 25 25 25 25 25 25 25 25 25 25 25 25 25
25
BOYS GIRLS BOYS GIRLS BOYS GIRLS BOYS GIRLS BOYS GIRLS BOYS GIRLS BOYS GIRLS BOYS GIRLS
Figure 5. Sampling frame of students from districts of Punjab, India
56
3.3 Tools used
The study employed a variety of tools, including a demographic sheet, student academic records,
self-report questionnaires, and an art technique. The demographic sheet gathered information
about the parents' attributes, like gender, income bracket, and educational attainment. All the
questionnaires used in the study were standardized psychological tests. The original
questionnaires were in English. The author employed the back translation technique to ensure
that the participants could understand the Hindi version of the questionnaires. For the Hindi
edition, the author used the help of a bilingual Hindi-English teacher. The demographic sheet,
standardized questionnaires, and art technique (Mandala Sheet) are attached in the Appendices.
3.3.1 Adverse Childhood Experiences Questionnaire (ACE-Q)
The Adverse Childhood Experiences Questionnaire (ACE-Q) (Felitti et al., 1998) was used to
determine whether or not the child has experienced adversities in their lives. The questionnaire
consists of 10 items (e.g., “Were your parents ever separated or divorced? Was a household
member depressed or mentally ill, or did a household member attempt suicide?”). The
parents/caregivers completed this questionnaire. The respondent must tick either ‘Yes’ (adversity
faced) or ‘No’ (adversity not faced) for each item. The possible scores are 0–10. The total scores
for the items measure the questionnaire. The calculated realibility of the sample studied was
(Cronbach α) of 0.75.
3.3.2 Stirling Children’s Well-Being Scale (SCWBS)
Well-being was assessed using the Stirling Children's Well-being Scale (SCWBS), developed by
Liddle and Carter (2015). Twelve self-reporting items comprise the scale; answers are given on a
five-point Likert scale from 1 to 5 (1: Never; 5: All the time). The average score ranges from 1 to
57
5. The scale has two subcomponents, which consist of 6 items, each naming a positive emotional
state (e.g., “I think there are many things I can be proud of”) and a positive outlook (e.g., “I can
find lots of fun things to do”). It can be applied as a quick screening tool for psychological well-
being. Summated scores over the items provide a measure of well-being. The scale has a
reported reliability (Cronbach α) of 0.95.
3.3.3 Dispositional Hardiness Scale
Hardiness was measured using the Dispositional Hardiness Scale developed by Bartone and Paul
(2007). A four-point Likert scale, ranging from 0 to 3, is used to collect responses to the 15 self-
reporting questions (0: Not at all true; 3: Completely true). The items include commitment (e.g.,
"I look forward to my daily activities"), challenge (e.g., "Changes in routine are interesting to
me"), and control (e.g., "By working hard, you can nearly always achieve your goals"). Each
subscale covers the overall hardiness construct and has five items each. Summated scores over
the items provide a measure of hardiness. The scale has a reported reliability (Cronbach α) of
0.92.
3.3.4 Perceptions of Parents Scale (POPS) Child Version
Parental support was measured using the Grolnick, Ryan, and Deci (1991) Perceptions of Parents
Scale (POPS). The scale is a 22-item self-reporting questionnaire comprising 11 items for
mothers and 11 for fathers. To respond, participants must circle a letter in front of one of four
descriptions of a parent, most likely their parent (e.g., “Some fathers always like to talk to their
children’s teachers about how they are doing in school,” or “Some mothers usually don’t ask
their children what they did in school that day” and so on). It is a Likert scale with response
options ranging from 1 to 4. The range of average scores is 1 to 4. For each item, the participants
58
rated the degree to which the behavior or attitude described fit the item. The summation of scores
across the items provides a measure of parental support. The scale has a reported reliability
(Cronbach α) of 0.91.
3.3.5 The mentorship Relationship scale
The adult mentoring of children was measured using the mentorship relationship scale developed
by Jakacki and Pola (2011). To determine whether a mentor helps to improve academic
performance, attitude, behavior, or self-esteem, respondents must answer eight questions on a 5-
point Likert scale that ranges from 1 (strongly disagree) to 5 (strongly agree). The scale has two
parts. Part 1 describes the relationship with the mentor (e.g., “What was the role of your mentor?
How often did you communicate? How long have you had this relationship?”). For qualitative
data, students were asked to share any information about mentors in their lives in the comment
box that followed the mentor questions. Part 2 of the survey contains Likert scale items
describing the relationship with mentors, such as "Improve school performance (grades, test
scores); Improve self-esteem/self-confidence." The scale has the reported reliability (Cronbach
α) of 0.91.
3.3.6 Resilience Level
The resilience level in the children was estimated in two parts: part 1 was the self-reported
questionnaire, which was assessed using Ungar's (2009) Child and Youth Resilience Measure
(CYRM), and Part 2 was the art technique (Mandala Session) by South Lakes Federation (2014).
A positive correlation existed between CYRM and an art technique (r=.27, p<.01).
59
[Link] Child and Youth Resilience Measure (CYRM) (Part 1)
CYRM measures the resources—individual, relational, communal, and cultural—people have at
their disposal to support their resilience. The measure was created at the Resilience Research
Centre (RRC) in 14 communities worldwide as part of the International Resilience Project (IRP).
The CYRM was first intended for use with young people between 9 to 23 years old. Since its
original validation, the measure has also been modified into a 26-item variant for children ages 5
to 12. This 26-item self-report questionnaire asks respondents to rate their answers on a three-
point Likert scale from 1 to 3. (1: No; 2: Sometimes; 3: Yes). There are three possible scores: 1
to 3. It was used with resilience monkeys’ emojis. The items assess three subscales: context (e.g.,
"Do you participate in religious activities?"), relationship with the primary carer (e.g., "Do you
talk to your family about how you feel?"), and individual (e.g., "Do you share with people
around you?"). Resilience is measured using the sum of the scores across the items. The scale
has a reported reliability (Cronbach α) of 0.86.
[Link] The art technique (Part 2)
Resilience can be identified at four levels: individual, family, school, and friends/community.
The researcher employed the mandala, an artistic technique, in the current study to assess
resilience, keeping in mind the resiliency mandala (Wolin & Wolin, 1993). A mandala is a
symbol of realizing one's inner self or essence. This entails achieving serenity or peace and being
true or honest with oneself (Babouchkina & Robbins, 2015). Mandalas have a calming and
rejuvenating effect on their creator, according to Jung (1972). With mandalas, children can
acquire resilience, concentrate their attention on the drawing process, and put their current
difficulties behind them (Kim et al., 2017).
60
Children's resilience was assessed using a Mandala Session: Exploring Feelings via Colors. The
idea was to use color to investigate their feelings about various aspects of their lives. A mandala
sheet and colored pencils (Faber's castle coloring pens) were required. Part 1 of the mandala
session breaks down the mandala into four divisions: self, friends, school, and home. Children
must first indicate in one section how they feel about a specific aspect of their lives, such as
"self." For example, if they are content and at ease in that area, they may indicate it using yellow.
They may go with a darker hue if circumstances are difficult.
The young person will be free to select the color they feel best represents their emotions, yet the
researcher will still let them interpret the colors in any way they see fit. Part two states that after
the young person has colored in one portion, discuss their color choices with them. Explain the
reasons behind the young person's color choices using active listening techniques. In the third
part, the researcher will select a different area and ask the young person to color it, repeating part
one. Again, talk about the rationale behind their color selections after this. After that, you can do
this for every section. There were light and dark colors. A score of one was represented by a dark
color and a score of two by a light color. The mandala had an overall score of 8. Summated
scores measured resilience. The art technique has a reported reliability (Cronbach α) of 0.56. The
scoring of the mandala is provided in the appendices.
3.3.7 Emotion Awareness Questionnaire (EAQ 30)
Emotion awareness was measured using the Emotion Awareness Questionnaire (EAQ)
developed by Rieffe et al. (2007). The questionnaire is a 30-item self-reporting test with three
Likert scale response options: 1 for not true, 2 for sometimes true, and 3 for frequently true. The
possible scores are 1 to 3. The items measure Differentiating Emotions {e.g., “I never know
exactly what kind of feeling I am having” (reverse coded)}, Verbal Sharing of Emotions (e.g., “I
61
can easily explain to a friend how I feel inside”), Bodily Awareness of Emotions (e.g., “When I
am scared or nervous‚ I feel something in my tummy”), Acting Out Emotions (e.g., “It is
important to understand how I am feeling”), Analyses of Emotions {e.g., “When I am upset
about something, I often keep it to myself” (reverse coded)}, and Attention to Others' Emotions,
{e.g., “When I am angry or upset, I try to hide this” (reverse coded)}. Summated scores over the
items provide a measure of EAQ. The scale has a reported reliability (Cronbach α) of 0.91.
3.3.8 Hare Self-esteem Scale
Self-esteem is measured using the Hare Self-esteem scale, which was created by Hare in 1975. It
is a 30-item self-reporting questionnaire requiring responses on a four-point Likert scale varying
from 1 to 4 (1: strongly disagree; 4: strongly agree). The items consist of three domains: self-
esteem school (e.g., “Most of my teachers do not understand me” reverse coded), self-esteem
peer (e.g., “I am not popular as other people of my age” reverse coded), and self-esteem home
(e.g., “My parents expect too much of me” reverse coded); each subscale consists of 10 items.
The scores ranged from 1 to 4 per domain. A score for overall self-esteem is calculated as the
average of the three subscales. Higher scores correspond to greater self-esteem. The original
scale has a reported reliability (Cronbach α) of 0.97.
3.3.9 Academic Achievement
The children's academic achievement in the form of marks was directly obtained from the school
teacher/principal. An annual report of students was obtained. The marks system of all the schools
was the same. As per the CBSE guidelines, students' marks are determined by utilizing a
Continuous Comprehensive Evaluation (CCE) for every subject. Formative assessment (FA) and
summative assessment (SA) are the two parts of the CCE. Teachers might implement formative
62
assessments to track their students' progress continuously. Quizzes, projects, assignments, and
class exams could be a part. In a school year, four formative evaluations are completed.
Performance evaluation at the end of each term is the summative assessment. Summative
evaluations take the form of written exams and are given in September and March at the end of
each term. The percentage of students in each of the seven subjects was taken in the current
study. Table 1 below describes the assessment criterion.
Table 1. The assessment criteria of the subjects
Term Type of % of weight Term-wise weight Total
assessment
First term- Formative 10% Formative 20% Formative
(April to Assessment - 1 Assessment Assessment
September) Formative 10% (1+2) =40%
Assessment - 2
Summative 30% Summative 30%
Assessment - 1 Assessment
–1
Second Term – Formative 10% Formative 20% Summative
(October to Assessment – 3 Assessment Assessment
March) Formative 10% (3+4) =60%
Assessment - 4
Summative 30% Summative 30% Total = 100
Assessment - 2 Assessment marks
-2
Note: Guidelines from the CBSE for the Continuous and Comprehensive Evaluation (CCE)
system. All schools in India affiliated with the CBSE must meet the criteria listed above
([Link]).
63
The criteria mentioned above apply to all the subjects: seven subjects, a total of 700. The
percentage was calculated for each student (Percentage = Marks obtained/700*100) and recorded
in the data.
3.4 Procedure
In order to acquire authorization to collect data on school premises, the researcher personally
approached the principals of the selected schools. The head was given a brief explanation of the
study's purpose by the researcher, along with an assurance that the data collection would not
interfere with the academic routines or teaching schedules of the participating students and the
respective teachers. Following approval, the researcher gave the demographic sheet, consent
form, and questionnaire on adverse childhood experiences to the parents of the children via the
school administration. The consent form stated that all data will be treated with the utmost
confidentiality and used exclusively for research purposes. Also, it indicated that their children’s
involvement was entirely optional and that they were free to end it at any time. After obtaining
consent from parents and administering the adverse childhood experiences questionnaire, the
researcher approached the children and administered the test individually. For each child, the
duration was roughly sixty minutes. Children colored the corresponding sections of the mandala
on mandala papers that they first received and then completed questionnaires. The researcher
provided reinforcement, in terms of some educational materials, to all children who had
participated. The data was scored and entered into the Excel sheet. Amos (version 23.0) and
SPSS (version 26.0) were used for the data analysis. The interpretation of the results was carried
out using the several conceptual frameworks provided here.
64
3.5 Statistical Analysis
The statistical methods employed to examine the data are described in this section. Descriptive
and inferential statistics were the statistical methods used in this research. Descriptive statistics
were produced to represent the data. Inferential statistics like correlation and structural equation
modeling (SEM) were utilized to examine the relationship between the variables. Amos (version
23.0) and SPSS (version 26.0) were used for the data analysis.
3.5.1 Descriptive Statistics
Descriptive statistics were calculated for behavioral outcomes, psychological variables, and
socio-demographic variables. Quantitative and qualitative research used descriptive statistics to
calculate percentages and a data summary (Kaliyadan & Kulkarni, 2019). It includes measures of
central tendency (mean, median & mode) and dispersion and variation (range, standard deviation
& variance). The current study computed the percentage, mean, and standard deviation for socio-
demographic information. The behavioral outcomes, psychological factors' means, and standard
deviations were also calculated.
3.5.2 Correlation
Correlation analysis was performed to ascertain the link between the variables. Analysis was
conducted to determine the degree to which variables like adversities, protective factors
(hardiness and well-being), compensatory factors (parental support and adult mentoring),
resilience, and behavioral outcomes (academic achievement, emotional awareness, and self-
esteem) are related to one another.
65
3.5.3 Structural Equation Modeling (SEM)
Structural equation modeling (SEM) was employed to investigate the relationship between the
variables. The primary goal of the investigation was to examine the conceptual models that
illustrate the connections between the independent (adversities, protective factors, and
compensatory factors) and dependent variables (resilience and behavioral outcomes).
In the present research, protective factors (well-being and hardiness) and compensatory factors
(adult mentoring and parental support) were the mediating variables between adversities,
resilience, and behavioral outcomes (emotional awareness, self-esteem, and academic
achievement).
SEM includes two processes: a validation of the measuring model applied by the
confirmation factor analysis and a second adjustment of the structural model performed by the
path analysis. In the current investigation, structural equation modelling, or SEM, was conducted
using Amos 23.0. Researchers can enter their model into Amos as a path diagram to show the
possible relationships between the different variables. If the CR is more than ± 1.96 in p =.05 and
± 2.58 in p = 0.01, the estimate is significant.
It provides various absolute model fit indices (Chi-square, GFI, NFI, and many others) to
gauge how well the model fits the information. The researchers who conduct analysis using SEM
and whose sample size is quite 200 should also consider other fit tests (ex., NNFI, CFI,
RMSEA to check the model (Shadfar and Iraj, 2013). The following critical indices (Table 2)
were employed in this study to assess the structural models' goodness of fit.
66
Table 2. Essential Indices
INDICES CRITERIA RESEARCHERS
CMIN/DF Between 2-5 Paswan, 2009
GFI >.95 OR >.90 Jöreskog and D. Sörbom. 1993
NFI >.90 Bentler & Bonett, 1980
CFI >.95 OR >.90 Hu, Li-Tze, and Peter Bentler.
1999
TLI >.95 OR >.90 Bentler, 1990
RMSEA <.08 Browne, Michael and Robert
Cudeck. 1992
67