Group Assignment:

Data Analysis Report

No. Name Student ID

1 Eddie Yap 1002267012

2 Chew Hooi Yin 1002267005

3 Joserine Chong Zoe Ling 1002370728

4 Guo Jijie 1002161231

Ms. Syarifah Maisarah Syed Alwi

SP110 Statistics for Psychology
(September – October 2023)
Faculty of Social Sciences and Liberal Arts
UCSI University, Kuala Lumpur
[email protected]

18 OCTOBER 2023
 1

ACADEMIC INTEGRITY ASSIGNMENT DECLARATION FORM

I am submitting the assignment for:

ㅁ an individual project or
✓ a group project on behalf of all members of the group. It is hereby confirmed that the submission is
authorized by all members of the group, and all members of the group are required to sign this declaration.

I/We declare that: (i) the assignment here submitted is original except for source material explicitly
acknowledged/all members of the group have read and checked that all parts of the piece of work,
irrespective of whether they are contributed by individual members or all members as a group, here submitted
are original except for source material explicitly acknowledged; and (ii) the piece of work, or a part of the
piece of work has not been submitted for more than one purpose(e.g. to satisfy the requirements in two
different courses) without declaration. I/We also acknowledge that I am/we are aware of the University’s
policy and regulations on honesty in academic work, and of the disciplinary guidelines and procedures
applicable to breaches of such policy and regulations, as contained in the University website
https://www.ucsiuniversity.edu.my/sites/default/files/policy_for_undergraduate_programme.pdf .
In the case of a group project, we are aware that all members of the group should be held responsible and
liable to disciplinary actions, irrespective of whether he/she has signed the declaration and whether he/she has
contributed, directly or indirectly, to the problematic contents.

I/we declare that I/we have not distributed/ shared/ copied any teaching materials without the consent of the
course teacher(s) to gain unfair academic advantage in the assignment/ course. I/We also understand that
assignments without a properly signed declaration by the student concerned and in the case of a group
project, by all members of the group concerned, will not be graded by the teacher(s).

Eddie Yap
Chew Hooi Yin
Joserine Chong Zoe Ling
Guo Jijie 18 OCTOBER 2023
______________________________ ______________________________
Signature(s) Date

Eddie Yap 1002267012

Chew Hooi Yin 1002267005
Joserine Chong Zoe Ling 1002370728
Guo Jijie 1002161231
______________________________ ______________________________
Name Student ID

SP110 Statistics for Psychology

______________________________ ______________________________
Course Code Course Title
 2

Turnitin Report
 3

Introduction

Adolescence represents a crucial period in human development marked by significant

emotional and psychological growth (Backes & Bonnie, 2019). This transformative phase

introduces numerous adjustments and challenges influenced by a multitude of factors.

Understanding the complex interplay among these factors is vital for gaining insights into the

emotional well-being and family dynamics of adolescents. In this research, we undertake a

comprehensive exploration of these factors, with a specific focus on the roles of adolescent

age, parental warmth, parental income, family size, gender, and race in shaping adolescents'

experiences and self-esteem.

Our analysis commences with initial assessments, where we employ correlation

analyses to scrutinize connections between the variables. We examine the links between

adolescent age and paternal warmth, father's age and paternal warmth, father's monthly

income and paternal warmth, family size and paternal warmth, adolescent age and adolescent

self-esteem, father's age and adolescent self-esteem, father's monthly income and adolescent

self-esteem, family size and adolescent self-esteem, and the correlation between paternal

warmth and adolescent self-esteem. These preliminary analyses lay the groundwork for a

more detailed investigation of group disparities via t-tests and ANOVA tests, encompassing

comparisons of gender and race groups, along with an exploration of the relationship between

gender and race through a Chi-square test.

The final piece of this comprehensive study involves our multivariate analyses. Here,

we delve deeper into the interactions among these diverse factors. Our objective is to gain a

deeper understanding of how the combined influence of adolescent age, parental warmth,

parental income, family size, gender, and race shapes adolescents' emotional well-being and

self-esteem. By means of multiple regression analyses, we evaluate the collective impact of

 4

these variables and highlight the most influential predictors concerning paternal warmth and

adolescent self-esteem.

This study endeavors to provide an all-encompassing viewpoint on the intricate

network of factors that define the adolescent experience. Our hope is that the discoveries will

not only enrich our comprehension of the adolescent psyche but also inform policies and

interventions tailored to support and nurture the emotional well-being of this pivotal

demographic group.
 5

Preliminary Step

The preliminary phase of the research was deemed essential as it contributed to the

overall smoothness of the process and enhanced the quality, validity, and reliability of the

research findings. This phase involved several crucial steps. Initially, data was imported from

existing sources into the file, followed by the definition of variables in the SPSS data, with

attributes such as name, type, decimals, labels, values, and the handling of missing data being

specified.

Subsequently, a comprehensive review and potential re-editing of the questionnaire

were undertaken to ensure that no responses were incomplete or unsatisfactory.

Unsatisfactory responses were managed accordingly, typically involving their exclusion from

the dataset. Codes were then assigned to each potential response, indicating the respective

category for each question. These codes played a central role in the categorization of data in

subsequent analyses.

Upon completion of the coding process, the coded data were subsequently transferred

from the questionnaires to the SPSS software using keypunching. Data cleaning was

performed using the pairwise deletion method, restricting the analysis solely to complete

questionnaires. For instance, due to 13 instances of missing values in 'Father’s Age,' the

analysis was focused on the 373 completed cases. The following step-by-step procedure was

implemented as explained below to guide the data cleaning process.

Within the "Variable View" tab at the bottom of the SPSS interface, missing data were

meticulously identified by visual inspection of the variables. Cases featuring missing data

were readily distinguishable through the presence of '999' in the variable labelled "FaAGE."

Specifically, cases 15, 17, 44, 48, 59, 66, 93, 170, 188, 192, 225, 244, and 260 were

acknowledged as having missing data. The choice was made to employ the pairwise deletion
 6

method, a common practice that exclusively takes into account cases with complete data. It's

worth noting that this strategy was selected based on its suitability for the specific research

requirements. Under the "Data" menu, the "Select Cases" option was accessed. Within the

ensuing dialogue box, the "If condition is satisfied" option was chosen. To define the criteria

for data retention, the "If" button beneath this condition was activated. Specifically, the

criterion was configured to retain only complete cases without any missing data, as defined

by the condition `NOT MISSING(FaAGE).` Subsequent to the application of the condition,

the "Filter" option within the "Data" menu was activated, indicating the use of a filter in the

analysis. Further data analysis procedures were executed, accompanied by the application of

relevant statistical techniques. The analysis solely considered cases meeting the specified

condition, i.e., those without any missing data in the 'Father's Age' variable.

Anomalies, such as the case in the frequency table for the statement "I am able to do

things as well as most people," where 33 and 44 were replaced with 3 and 4 respectively,

were resolved to ensure data consistency.

Data grouping was a critical procedure employed in this research to categorize

variables, including the father's age, father's income, and family size. Father's age data was

effectively organized into four distinct groups: individuals below 40, those aged 41-50,

between 51-60, and individuals aged 60 and above. The procedures below were performed to

categorise the variable "Father’s Age" into four categories.

In SPSS, the "Transform" menu was navigated, and "Recode into Different Variables"

was selected. The variable "Father’s Age" was inserted into the "Input Variable" section and

transferred to the "Output Variable" list, renamed as "Father_Age_4categories." The process

was continued by selecting "Old and New Value" in the dialogue box. In the "Old Value"

section, "Range, LOWEST through value" was chosen, and "40" was input as the lowest
 7

value. In the "New Value" section, "1" was assigned as the value for this range and "Add"

was clicked. Next, in the "Old Value" section, "Range" was selected, and "41" was input as

the lowest value and "50" as the highest value. In the "New Value" section, "2" was assigned

as the value for this range, and "Add" was clicked. This process was repeated for the

remaining age groups: "51 to 60" and "60 and above." For the "51 to 60" range, "Range" was

chosen, "51" was input as the lowest value, and "60" as the highest value. In the "New Value"

section, "3" was assigned as the value for this range, and "Add" was clicked. For "60 and

above," "Range, value through HIGHEST" was selected, "60" was input as the value, "4" was

assigned as the new value, and "Add" was clicked. After defining these ranges, "Continue"

was clicked to return to the "Numeric Variable -> Output Variable" interface. Finally, "OK"

was clicked to confirm the changes made.

The same meticulous grouping procedures were applied to variables "Father's

Income" and "Family Size," with specific class intervals ensuring meaningful and relevant

categorizations. For instance, "Father’s Income" was divided into 13 categories, while

"Family Size" was categorized into four meaningful groups.

Finally, the selection of the appropriate statistical approach for analyzing the data and

testing the research hypotheses was made. The chosen data analysis strategy encompassed

Univariate Analysis, which included descriptive statistics such as frequency, percentage,

mean, standard deviation, minimum, and maximum. Bivariate Analysis techniques, such as

correlation analysis, t-tests, ANOVA tests, and chi-square tests, were applied to explore

relationships and differences between variables. Additionally, Multivariate Analysis,

specifically multiple regression analysis, was employed to investigate complex interactions

and dependencies among multiple variables.

 8

Data Transformation

A1.

a) Recode the following items in self-esteem scale: Negative item: (se2, se5, se6, se8,

se9)

The initial step in this assignment involves data preparation, encompassing several

essential tasks. These tasks include the review and editing of the questionnaire, the

assignment of codes to each possible response, data transcription, dataset cleaning, and the

addressing of missing responses. Lastly, adjusting the data and selecting the data analysis

strategy. Following these preparatory steps, the calculation of total scores for parental warmth

and self-esteem is required as stated in the assignment handout. However, before the

computation of total scores for paternal warmth and self-esteem, it is essential to reverse

negatively scaled items within the questionnaire. Specifically, items se2, se5, se6, se8, and

se9 in the dataset needed to be recoded. Recoding entails changing specific values or ranges

of values on these variables to reflect different meanings.

Frequently, questionnaires incorporate a mixture of positively and negatively worded

items. For instance:

1. "I feel that I have a number of good qualities." (se3)

(Response options: Strongly Disagree 1 2 3 4 5 Strongly Agree)

2. "At times, I think I am no good at all." (se2)

(Response options: Strongly Disagree 1 2 3 4 5 Strongly Agree)

In such questionnaires, a respondent who rates question 1 with a "5" would likely rate

question 2 with a "1," and simply summing or averaging these responses would yield limited
 9

insight. Indeed, by reversing the response to the second question (recoding 1 as 5, 2 as 4, and

so forth), a meaningful total or average score can be computed. Recoding plays a pivotal role

in ensuring that all responses within a dataset consistently follow the same direction. This, in

turn, enhances the validity of the overall score by aligning responses to both positively and

negatively worded items. These adjustments not only enhance validity but also significantly

improve the interpretability of the data, making it easier to compare and analyze. As a result,

this transformed score serves as a valuable tool for facilitating subsequent analyses. To aid in

the practical application of this process, the following paragraphs offer a step-by-step guide

for performing this recoding using SPSS.

To recode the items se2, se5, se6, se8, and se9 in SPSS, the dataset that requires

transformation was opened and accessed. Next, the “Recode into Different Variables” option

within SPSS was navigated to by clicking on “Transform”. In the “Recode into Different

Variables” interface, the variable se2 was moved to the “Input Variable -> Output Variable”

list. The new variable (se2) was then created in the input variable section, and it was renamed

into REse2 by clicking on se2 using the arrow button, and then providing the ‘new name’

(REse2) under “Name” of the “Output Variable” section for the new variable. To initiate the

recoding process, the "Change" button was clicked. This action led to the transformation of

"se2" into "REse2." The newzs variable's name ("REse2") was subsequently added to the

"Input Variable -> Output Variable" list. The above procedures were then repeated for the

variables "se5," "se6," "se8," and "se9." These variables were similarly transformed into new

variables: "REse5," "REse6," "REse8," and "REse9," respectively.

During the recoding process, the specification of old and new values takes place. In

the “Old and New Values” dialogue of “Recode into Different Variables”, REse2, REse5,

REse6, REse8, and REse9 on the input variable were specified to changed (the Old Values),

and what they should be changed to (the New Values). In this case, the value is set to 0 under
 10

the “Value” section of “Old Variable”, and the new value is set to 4, then "Add" is clicked.

This recoding process results in the following changes: 0 becomes 4, 1 becomes 3, 2 remains

2, 3 becomes 1, and 4 becomes 0. "Continue" is clicked to proceed with the recoding, and the

process is finalised by clicking the "OK" button at the bottom of the “Recode into Different

Variables” dialogue to run the value changes. After these steps have been completed, the

changes in the new variables in the “Variable View” can be reviewed, where the recoded

variables, namely REse2, REse5, REse6, REse8, and REse9, can be found. This recoding

process ensures that the negative items are appropriately transformed for further analysis.

A2.

Compute the total scores for the following scales:

a) Paternal warmth

The calculation of total scores in SPSS stands as an important step in research and

data analysis, offering substantial advantages. It simplifies data analysis by consolidating

multiple responses into a single variable, thereby streamlining the interpretation and further

analysis processes. Moreover, total scores play a more extensive role by encompassing

related items, thereby contributing to a comprehensive evaluation of a specific construct. In

the context of paternal warmth, they provide a holistic assessment of a father's warmth

toward their children. Additionally, total scores facilitate comparisons between individuals or

groups, making it more straightforward to assess overall performance and responses, which,

in turn, aids in conducting group-level comparisons and identifying underlying data patterns.

The process of deriving the total score for paternal warmth involves summing

individual scores from specific statements related to paternal warmth, specifically pw1, pw2,
 11

pw3, pw4, pw5, pw6, pw7, pw8, pw9, and pw10. This process unfolded through a series of

steps in SPSS as mentioned in the paragraph below.

The process was initiated with the selection of the 'Transform' option from the SPSS

menu. Subsequently, 'Compute Variable' was chosen from the dropdown menu, opening the

'Compute Variable' dialogue. Within this dialogue, the 'Target Variable' field was designated

with the name "Paternal_Warmth," signifying the combined score for paternal warmth. The

equation for calculating the total score was carefully formulated, involving the summation of

the individual items, as indicated by `pw1 + pw2 + pw3 + pw4 + pw5 + pw6 + pw7 + pw8 +

pw9 + pw10`. Rigorous verification measures were implemented to ensure the equation's

precision, aligning variable names (e.g., pw1, pw2, etc.) with the designated items for

summation. Once confidence in the equation's accuracy was established, the 'OK' button was

activated. SPSS executed the necessary computations, generating the total scores for paternal

warmth. The results were encapsulated in a new variable, "Paternal_Warmth," containing the

combined scores for each case (e.g., each participant) in the dataset.

b) Adolescent self-esteem

The procedure for calculating the total score for adolescent self-esteem was executed

following the same methodology as that used for paternal warmth. In this case, the statement

se1 + REse2 + se3 + se4 + REse5 + REse6 + se7 + REse8 + REse9 + se10 was combined to

represent adolescent self-esteem. To perform this computation in SPSS, the following steps

were followed.
 12

The 'Transform' option from the SPSS menu was selected. 'Compute Variable' was

chosen from the dropdown menu, which opened the 'Compute Variable' dialogue. Within this

dialogue, the 'Target Variable' field was designated with the name "Adolescent_Self_Esteem"

to indicate the combined score for adolescent self-esteem. The equation for calculating the

total score, which involved summing the individual items (se1 + REse2 + se3 + se4 + REse5

+ REse6 + se7 + REse8 + REse9 + se10), was formulated. To ensure the accuracy of the

equation, meticulous verification was conducted to confirm that the variable names (e.g., se1,

REse2, etc.) were aligned with the intended items for summation. Once confidence in the

correctness of the equation was established, the 'OK' button was clicked. SPSS then executed

the necessary calculations, resulting in the generation of the total scores for adolescent

self-esteem. The outcome was a new variable, "Adolescent_Self_Esteem," containing the

consolidated scores for each case (e.g., each adolescent) within the dataset.
 13

Reliability Test

B1.

Access the reliability of the following scales:

Reliability, as defined by Gravetter and Wallnau (2016), pertains to a research study's

ability to consistently produce identical results when measuring the same scenario with the

same measurement tool, whether it's performed repeatedly or across different conditions,

locations, and populations. Reliability tests are essential as they assess the degree to which

the items within a scale or instrument are interconnected, ensuring that they provide

consistent results over time, under varying circumstances, and among diverse populations.

This evaluation of internal consistency is fundamental in guaranteeing the validity and

reliability of the data collected.

Reliability tests serve a dual purpose; they not only evaluate internal consistency but

also aid in the identification of items within a scale that may exhibit reduced reliability.

Detecting these problematic items enables researchers to take necessary steps to either refine

or eliminate them, ultimately contributing to the overall quality and effectiveness of the

measurement instrument.

A common approach to assessing reliability is through the evaluation of internal

consistency, wherein researchers often employ statistical tools such as Cronbach's alpha

coefficients. These measures play a pivotal role in helping researchers ascertain whether the

items within a scale or instrument consistently measure the same underlying construct or

concept. In essence, this ensures that the items are interrelated, providing consistent results

under varying conditions and across diverse populations.

 14

DeVellis (2012) emphasizes that the ideal Cronbach's alpha coefficient for achieving

optimal reliability should exceed 0.7. It's important to note that the sensitivity of Cronbach's

alpha values to the number of items in a scale can influence the results. Generally, scales with

fewer than ten items might yield lower Cronbach's alpha values. When interpreting

Cronbach's alpha, values above .90 are classified as 'Excellent,' values above .80 as 'Good,'

those exceeding .70 as 'Acceptable,' values surpassing .60 as 'Questionable,' values above .50

as 'Weak,' and values over .40 as 'Unacceptable.'

The following procedures provide a step-by-step guide for assessing the reliability of

the Paternal Warmth scale in SPSS. The process was initiated by launching SPSS and loading

the dataset containing the variables relevant to the Paternal Warmth scale, specifically pw1 to

pw10. Next, the 'Analyze' option at the top of the SPSS window was selected. In the

'Analyze' dropdown, the 'Scale' option was chosen, leading to the selection of 'Reliability

Analysis.' Within the 'Reliability Analysis' dialogue, a list of variables from the dataset was

presented. The individual items that comprised the Paternal Warmth scale (i.e., pw1 to pw10)

were meticulously chosen. These selected items were then transferred into the designated box

labelled 'Items' using the provided arrow button. To measure internal consistency reliability,

the 'Alpha' option was confirmed as selected under the 'Model' section, representing

Cronbach's Alpha—a widely employed measure for this purpose. Subsequently, in the 'Scale

label' box, the name 'Paternal Warmth' was entered to identify the scale being assessed. To

define the specific statistics to be calculated, the 'Statistics' button was activated. Under

'Descriptives for,' the selections included 'Item,' 'Scale,' and 'Scale if item deleted,' providing

a comprehensive array of statistics for individual items and the overall scale. In the

'Inter-Item' section, 'Correlations' were chosen to gain insights into the relationships between

the different items that constituted the scale. Under the 'Summaries' section, 'Correlations'

were similarly selected to yield supplementary correlation data. After configuring these
 15

analysis preferences, the 'Continue' option was clicked, leading back to the primary

'Reliability Analysis' dialogue. Finally, the analysis was executed by clicking 'OK.'

Upon completion, SPSS generated an output that encompassed the reliability

coefficient (Cronbach's Alpha) and other pertinent statistics for the Paternal Warmth scale.

The resulting data was reviewed and organized into a tabulated format (Table 1) to facilitate

the assessment of the internal reliability of both the scale and its individual items.

The procedure for evaluating the reliability of the Adolescent Self-Esteem scale

closely mirrored the methodology employed for the Paternal Warmth scale. In this analysis,

the individual items encompassing the Adolescent Self-Esteem scale consisted of se1, REse2,

se3, se4, REse5, REse6, se7, REse8, REse9, and se10. Likewise, as in the preceding

procedure, the 'Scale label' box was utilized to designate the name 'Adolescent Self-Esteem,'

serving the purpose of clearly identifying and labelling the scale within the resulting analysis

output. Similar to the Paternal Warmth assessment, the data generated from this analysis was

reviewed and structured into a tabulated format (Table 1).

a) Paternal Warmth

The reliability of the Paternal Warmth scale was assessed using Cronbach's Alpha

coefficient (α), a measure of internal consistency. As presented in Table 1, Cronbach's alpha

for the 10-item Paternal Warmth scale was α = 0.90, indicating excellent internal consistency.

This impressive level of internal consistency, as indicated by Cronbach's Alpha coefficient,

underscores the strong interrelationships among the scale's items, resulting in consistent and

reliable outcomes. The α value of 0.90 significantly surpasses the commonly accepted

threshold of 0.70 for internal consistency, further validating the scale's reliability.

Consequently, a review of the scale items and their interitem correlations showed that all ten

items contributed positively to the overall internal consistency of the scale. Each item
 16

demonstrated a positive correlation with the total scale score, further confirming their

contribution to the coherent measurement of paternal warmth. The reliability test findings

support the conclusion that the Paternal Warmth scale is a reliable and internally consistent

instrument for measuring paternal warmth. This high level of internal consistency lends

credibility to the scale's ability to produce consistent and dependable results when assessing

paternal warmth in the following research contexts.

b) Self-esteem

The reliability of the Adolescent Self-esteem scale was assessed using Cronbach's

Alpha coefficient (α), a measure of internal consistency. As presented in Table 1, the 10-item

Adolescent Self-esteem scale demonstrated Cronbach's alpha of α = 0.42. This relatively low

value of Cronbach's Alpha (α = 0.42) suggests that the items within the Adolescent

Self-esteem scale displayed poor internal consistency. The α value falls significantly below

the commonly accepted threshold of 0.70 for internal consistency, indicating that the items

within the scale do not consistently relate to each other. This low level of internal consistency

raises concerns about the reliability of the scale. A closer examination of the scale items and

their interitem correlations indicated that the items did not contribute positively to the overall

internal consistency of the scale. Many of the items exhibited weak or negative correlations

with the total scale score, further undermining the scale's reliability. In summary, the findings

from the reliability test indicate that the Adolescent Self-esteem scale exhibits poor internal

consistency, with a Cronbach's Alpha coefficient below the accepted threshold. Consequently,

the scale may not be a reliable instrument for measuring adolescent self-esteem in the

research context.

Table 1
Reliability analysis of scales.
 17

Scale Number of items Cronbach's Alpha

Paternal warmth 10 0.90

Adolescent self-esteem 10 0.42

 18

Univariate Analysis

C1.

Describe the profile (e.g. frequency, percentage, mean, standard deviation, minimum,

maximum) of the respondents with respect to the following items:

a) Adolescent’s characteristics (age, sex and race)

Table 2 offers a comprehensive overview of the characteristics of the adolescent

participants, encompassing age, gender, and race. It provides a frequency distribution that

illustrates the distribution of frequencies across various values, offering valuable insights into

the sample composition. The frequencies ('n') represent the number of occurrences for each

characteristic and are additionally expressed as percentages (%).

The sample consisted of 373 adolescents, distributed across four age groups: 13, 14,

15, and 16. The majority of the respondents, 32.4%, were 13 years old, with 121 individuals

in this group. Similarly, 30.0% of the participants were 14 years old, totalling 112

respondents. Adolescents aged 15 accounted for 27.6% of the sample, with 103 participants.

The smallest age group was 16 years old, representing 9.9% of the respondents, comprising

37 individuals. Thus, the data indicates that the largest proportion of the participants fell

within the 13-year-old category, while the 16-year-old group had the fewest participants.

Descriptive statistics for age indicated a mean age of 14.15, a standard deviation of 0.99, and

an age range from (minimum = 13 years) to (maximum = 16 years).

In terms of gender distribution, the sample was divided into two categories: female

and male. The data showed that females constituted the majority, with 50.9% of the total

respondents, which corresponds to 190 individuals. Conversely, males made up 49.1% of the

respondents, totalling 183 individuals.

 19

Additionally, the racial composition of the respondents was assessed, with the three

main categories being Malay, Chinese, and Indian. The data revealed that Malays were the

dominant racial group, representing 74.0% of the respondents, with 276 individuals. Chinese

respondents constituted 16.4% of the total, totalling 61 individuals. The Indian community

comprised the minority, accounting for 9.7% of the respondents, with 36 individuals.

Figure 1
A histogram based on the frequency table of adolescence age.

In Figure 1, the histogram depicts the distribution of adolescent ages, revealing a

pattern reminiscent of a normal distribution. This distribution was conventionally compared

to the normal curve, known for its characteristic bell-shaped, unimodal, and symmetrical

form with a peak at the average. The frequency distribution indicated that age 13 had the

highest frequency, while a gradual decline in frequency was observed as we moved from age

14 to age 16 (see Figure 1).

Figure 2
 20

A histogram based on the frequency table of adolescent's gender.

In Figure 2, the histogram displayed the distribution of adolescents' gender, resulting

in a frequency distribution with a rectangular shape. This type of distribution is characterized

by nearly equal frequencies for all values. In this dataset, the number of male adolescents

closely corresponds to the number of female adolescents, indicating a near-even split between

the two genders. The rectangular shape of the distribution underscores the gender balance

within the sample (see Figure 2).

Figure 3

A histogram based on the frequency table of adolescent race

 21

In Figure 3, the histogram shows a distribution of adolescent race, resulting in an unimodal

frequency distribution where one particular category exhibited a significantly higher

frequency than the others. The histogram data analysis indicated that the Malay race was the

predominant category, displaying the highest frequency, while the frequencies for the other

racial categories were notably lower. This unimodal distribution highlights the substantial

prevalence of the Malay race within the sample.

Table 2
Characteristics of adolescents (n = 373).

Variable n Percentage (%)

Age (Years)
13 121 32.4
14 112 30.0
15 103 27.6
16 37 9.9
Mean 14.15
Sd. 0.99
Min. 13
Max. 16

Sex
 22

Male 183 49.1

Female 190 50.9

Race
Malay 276 74.0
Chinese 61 16.4
Indian 36 9.7
Note./Remark. Sd = Standard deviation, Min. = Minimum, Max = Maximum
b) Father’s Characteristics (age, monthly income)

Table 3 presents a comprehensive overview of the demographic characteristics of the

fathers participating in the study, with a focus on their age and monthly income. The table

employs frequency distribution to provide insights into the distribution of fathers across

various categories relative to the total number of occurrences. The value "n" signifies the

number of respondents' fathers within each category, and these figures are also expressed as

percentages (%).

Age categories encompassed four groups: fathers below 40 years old, fathers aged 41

to 50, fathers aged 51 to 60, and fathers aged 61 and above. The frequency distribution,

expressed as 'n' (number of occurrences), reveals that out of 373 respondents, 46 fathers

(12.3%) were below 40 years old, 204 fathers (54.7%) were aged between 41 and 50, 115

fathers (30.8%) fell within the 51 to 60 age range, while a smaller group of 8 fathers (2.1%)

were aged 61 and above. Notably, the most substantial proportion of fathers fell within the 41

to 50 age group, whereas the fewest were in the 61 and above category. The age data is

summarized by a mean age of 48.01, with a standard deviation of 6.16, and an age range from

(minimum = 34 years) to (maximum = 67 years).

In addition to age, fathers' monthly income was assessed across 13 different income

categories, ranging from below 499 MYR to 6000 MYR and above. The income distribution

is as follows: 13 fathers (3.5%) reported monthly incomes below 499 MYR, 60 fathers
 23

(16.1%) had incomes ranging from 500 to 999 MYR, 78 fathers (20.9%) fell within the 1000

to 1499 MYR category, 50 fathers (13.4%) reported monthly incomes from 1500 to 1999

MYR, 60 fathers (16.1%) had incomes between 2000 and 2499 MYR, 19 fathers (5.1%)

reported incomes of 2500 to 2999 MYR, 36 fathers (9.7%) earned between 3000 and 3499

MYR, 12 fathers (3.2%) had incomes of 3500 to 3999 MYR, 21 fathers (5.6%) reported

incomes from 4000 to 4499 MYR, 3 fathers (0.8%) earned between 4500 to 4999 MYR, 17

fathers (4.6%) had incomes ranging from 5000 to 5499 MYR, 1 father (0.3%) reported an

income of 5500 to 5999 MYR, and 3 fathers (0.8%) earned 6000 MYR and above. Notably,

the most substantial proportion of fathers reported monthly incomes in the 1000 to 1499

MYR category, whereas the fewest fathers earned between 5500 to 5999 MYR. The income

data is summarized by a mean monthly income of 1981.56 MYR, with a standard deviation

of 1277.74 MYR, and an income range from (minimum = 80 MYR) to (maximum = 6000

MYR).

Table 3
Characteristics of fathers (n = 373).

Variable n Percentage (%)

Age (Years)
below 40 46 12.3
41 to 50 204 54.7
51 to 60 115 30.8
61 and above 8 2.1
Mean 48.01
Sd. 6.16
Min. 34
Max. 67

Monthly Income (In MYR)

below 499 13 3.5
500 to 999 60 16.1
1000 to 1499 78 20.9
1500 to 1999 50 13.4
2000 to 2499 60 16.1
2500 to 2999 19 5.1
 24

3000 to 3499 36 9.7

3500 to 3999 12 3.2
4000 to 4499 21 5.6
4500 to 4999 3 0.8
5000 to 5499 17 4.6
5500 to 5999 1 0.3
6000 and above 3 0.8
Mean 1981.56
Sd. 1277.74
Min. 80.00
Max. 6000.00
Note. Sd = Standard deviation, Min. = Minimum, Max = Maximum
Figure 4
A histogram based on the frequency table of father’s age.

Figure 4 illustrates a histogram depicting the distribution of fathers' ages, revealing a

frequency distribution pattern characteristic of a normal distribution. In this context, the

distribution displays a multimodal shape, indicating the presence of multiple high-frequency

regions separated by lower frequencies. This observation contrasts with the standard of

comparison, which is an unimodal, symmetrical curve known as the normal curve,

characterized by a singular peak resembling the shape of a bell. The frequency distribution
 25

within the age group of 40 to 50 years appears to be at its peak, while the frequencies in other

age groups exhibit fluctuating patterns of increase and decrease.

Figure 5
A histogram based on the frequency table of father’s monthly income.

In Figure 5, the histogram illustrates the distribution of fathers' monthly income,

showcasing a frequency distribution pattern consistent with a normal distribution. This

distribution aligns with the standard of comparison, characterized by a bell-shaped curve

commonly referred to as the normal curve. The normal curve is an unimodal, symmetrical

distribution with a central peak resembling the shape of a bell. In this context, the frequency

distribution demonstrates a similar bell-shaped pattern, with the highest frequency occurring

within the 40 to 50 age group. Subsequently, frequencies in other income groups exhibit

fluctuating patterns of increase and decrease.

Table 4
Family social context (n = 373).

Variable n Percentage (%)

 26

Family Size (Member)

below 3 143 38.3
4 to 7 201 53.9
8 to 11 28 7.5
12 and above 1 0.3
Mean 4.54
Sd. 1.784
Min. 3
Max. 14
Note. Sd = Standard deviation, Min. = Minimum, Max = Maximum
Table 4 provides a comprehensive view of the frequency distribution concerning the

family social context, specifically in terms of family size. The frequency distribution is a

representation of the ratio of occurrences of each category to the total number of respondents,

which in this study amounted to 373 participants. These frequencies (n) are subsequently

converted into percentages (%) to facilitate a better understanding of the data.

The families of the respondents were categorized into four groups based on their size:

those with below 3 members, those with 4 to 7 members, those with 8 to 11 members, and

those with 12 members and above. The distribution reveals that 143 respondents come from

families with fewer than 3 members, constituting 38.3% of the sample. A larger segment,

comprising 201 respondents (53.9%), hails from families with 4 to 7 members. A smaller

group of 28 respondents (7.5%) belongs to families with 8 to 11 members. Only one

respondent (0.3%) reported being part of a family with 12 members or more.

Notably, the category with 4 to 7 family members represents the majority, while the

category with 12 or more members is the least prevalent. The statistical summary of the

family social context provides a mean value of 4.54, a standard deviation of 1.784, a

minimum family size of 3, and a maximum family size of 14.

Figure 6
A histogram based on the frequency table of family size.
 27

In Figure 6, the depicted frequency distribution exhibits a positively skewed shape

which means that the mean holds a higher value in contrast to both the mode and the median.

This skewness results from the influential presence of high scores within the distribution's

extended tail. Notably, the most prominent peak in the family size category occurs at 3

members, followed by a gradual decline in frequencies as family size increases.

 28

BIVARIATE ANALYSIS

D1.

Perform correlation analysis for the following pairs of variables:

a) Adolescent’s age and paternal warmth

The statistical analysis in this study employed Pearson's correlation test to assess the

relationship between adolescents' age and paternal warmth. The primary objective of this test

was to determine whether there is a significant correlation between these two continuous

variables. Table 5 presents the results of this analysis, which investigated the association

between adolescents' age and paternal warmth.

The null hypothesis posited that there is no significant relationship between

adolescents' age and paternal warmth. To examine this relationship, a Pearson's correlation

test was conducted. The results revealed a statistically significant negative correlation

between adolescent age and paternal warmth, denoted by r = -0.12*, with a p-value of 0.02

(p<0.05). Consequently, the initial hypothesis was rejected, indicating that adolescent age is

inversely correlated with paternal warmth. In summary, this finding suggests that there is a

statistical relationship between adolescent age and paternal warmth, and this relationship is

characterized by a decrease in paternal warmth as adolescents get older.

b) Father’s age and paternal warmth

The analysis in this study utilized Pearson's correlation test to investigate the

relationship between father's age and paternal warmth. The main aim of this test was to

ascertain if there exists a significant correlation between these two continuous variables. The

results of measuring the correlation between the father's age and paternal warmth are shown

in Table 5.
 29

The null hypothesis is that there is a significant relationship between the father's age

and paternal warmth. Pearson's correlation test was used to measure the relationship between

the father's age and paternal warmth. The results indicated that there was a significant

correlation between the father's age and paternal warmth, with r = -0.10 and p = 0.06 (p >

0.05). Therefore, the original hypothesis was rejected, suggesting that the father's age is

negatively correlated with paternal warmth. This statistical relationship suggests that there

might be factors associated with younger fathers that contribute to higher levels of paternal

warmth.

c) Father’s monthly income and paternal warmth

The analysis in this study employed Pearson's correlation test to investigate the

relationship between the father's monthly income and paternal warmth. The primary objective

of this test was to determine if there exists a significant correlation between these two

continuous variables. Table 5 provides an overview of the results of this assessment.

The null hypothesis posited that there is no significant relationship between a father's

monthly income and paternal warmth. Pearson's correlation test was the chosen method for

quantifying the relationship between these variables. The findings revealed that there was no

significant correlation between the father's monthly income and paternal warmth, with an

r-value of 0.03 and a p-value of 0.63 (p > 0.05). As a result, the original hypothesis failed to

be rejected, signifying that no significant correlation exists between the father's monthly

income and paternal warmth. The data does not support the idea that changes in the father's

monthly income are associated with changes in paternal warmth. The two variables are not

significantly linked in this study.

Table 5
Correlation of adolescent’s demographic profile and paternal warmth.
 30

Variable Paternal Warmth

r p

Adolescent’s Age -0.12* 0.02

Father’s Age -0.10 0.06
Father’s Monthly Income 0.03 0.63
Note. *p<0.05, **p<0.01, ***p<0.001
d) Family size and paternal warmth

The statistical analysis in this study utilized Pearson's correlation test to examine the

relationship between family size and paternal warmth. The main aim of this test was to

determine whether a significant correlation exists between these two continuous variables.

The results of this analysis are presented in Table 6, which illustrates the association between

family size and paternal warmth.

The null hypothesis states that there is no significant relationship between family size

and paternal warmth. Pearson's correlation test was applied to measure the relationship

between family size and paternal warmth. The results showed that there was no significant

correlation between family size and paternal warmth, as indicated by r = -0.04, p = 0.50 (p >

0.05). Consequently, the original hypothesis was accepted, suggesting that there is no

significant correlation between family size and paternal warmth. In a nutshell, the statistical

analysis did not find evidence to support a significant relationship or correlation between the

size of a family and the level of paternal warmth.

Table 6
Correlation of family size and paternal warmth.

Variable Paternal Warmth

r p

Family Size -0.04 0.50

 31

e) Adolescent’s age and adolescent self-esteem

The statistical analysis in this study employed Pearson's correlation test to examine

the relationship between adolescent age and adolescent self-esteem. The primary objective of

this test was to ascertain whether a significant correlation exists between these two

continuous variables. Table 7 presents the results of the assessment of the relationship

between adolescent age and adolescent self-esteem.

The null hypothesis posited that there was no significant relationship between

adolescent age and self-esteem. Pearson's correlation test was applied to measure this

relationship. The results revealed a significant correlation between adolescent age and

self-esteem, with r = -0.14**, p = 0.01 (p < 0.05). Consequently, the original hypothesis was

rejected, indicating a significant negative correlation between adolescent age and self-esteem.

In practical terms, it suggests that as adolescents get older, their self-esteem tends to decrease.

In other words, there is an inverse relationship between age and self-esteem, with self-esteem

declining as age increases.

f) Father’s age and self-esteem

The statistical analysis in this study employed Pearson's correlation test to investigate

the relationship between a father's age and adolescent self-esteem. The primary objective of

this test was to determine whether a significant correlation exists between these two

continuous variables. The results of this analysis are presented in Table 7, which outlines the

assessment of the relationship between a father's age and adolescent self-esteem.

The null hypothesis posited that there was no significant relationship between father's

age and adolescent self-esteem. Pearson's correlation test was utilized to examine this

relationship. The results, however, indicated that there was no significant correlation between
 32

the father's age and adolescent self-esteem, as evidenced by r = -0.01, p = 0.93 (p > 0.05).

Consequently, the original hypothesis was accepted, signifying that there is no significant

correlation between the father's age and adolescent self-esteem. In summary, the statistical

analysis did not reveal evidence of a meaningful relationship or correlation between the age

of fathers and the self-esteem of adolescents.

g) Father’s monthly income and self-esteem

The statistical analysis in this study utilized Pearson's correlation test to explore the

relationship between a father's monthly income and adolescent self-esteem. The primary

objective of this test was to ascertain whether a significant correlation exists between these

two continuous variables. The results of this analysis are presented in Table 7, illustrating the

assessment of the relationship between a father's monthly income and adolescent self-esteem.

The null hypothesis posited that there was no significant relationship between the

father's monthly income and self-esteem. Pearson's correlation test was employed to

investigate this relationship. The results, however, revealed a significant positive relationship

between father's monthly income and adolescent self-esteem, as evidenced by r = 0.140**, p

= 0.01 (p < 0.05). Therefore, the original hypothesis was rejected, indicating a significant

positive relationship between the father's monthly income and adolescent self-esteem. In

summary, the statistical analysis suggests that as the father's monthly income increases, the

adolescent's self-esteem also tends to rise.

Table 7
Correlation of adolescent’s demographic profile and adolescent’s self-esteem.

Variable Adolescent’s Self-Esteem

r p

Adolescent’s Age -0.14** 0.01

 33

Father’s Age -0.01 0.93

Father’s Monthly Income 0.14** 0.01
Note. *p<0.05, **p<0.01, ***p<0.001
h) Family size and self-esteem

The statistical analysis in this study employed Pearson's correlation test to investigate

the relationship between family size and adolescent self-esteem. The primary objective of this

test was to determine whether a significant correlation exists between these two continuous

variables. The findings regarding the correlation between family size and adolescent

self-esteem are presented in Table 8.

The null hypothesis posited that there is no significant relationship between family

size and self-esteem. Pearson's correlation test was utilized to examine this relationship

among young adults. The results, however, indicated that there was no significant relationship

between family size and youth self-esteem, as denoted by r = -0.10, p = 0.06 (p > 0.05).

Consequently, the original hypothesis was accepted, signifying that there is no significant

correlation between family size and adolescent self-esteem. In summary, the statistical

analysis did not find evidence to support a significant relationship or correlation between

family size and the self-esteem of adolescents.

Table 8
Correlation of family size and adolescent’s self-esteem.

Variable Adolescent’s Self-Esteem

r p

Family Size -0.10 0.06

i) Paternal warmth and self-esteem

 34

The statistical analysis in this study utilized Pearson's correlation test to explore the

relationship between paternal warmth and adolescent self-esteem. The main objective of this

test was to determine whether a significant correlation exists between these two continuous

variables. The results of this examination are presented in Table 9, depicting the correlation

between paternal warmth and adolescent self-esteem.

The null hypothesis postulates that there is no significant relationship between

paternal warmth and self-esteem. Pearson's correlation test was employed to quantify the

association between paternal warmth and self-esteem. The results revealed a significant and

positive relationship between paternal warmth and adolescent self-esteem, with a correlation

coefficient (r) of 0.24**, and a p-value of 0.00 (p < 0.05). Consequently, the original

hypothesis was rejected, signifying a substantial positive correlation between paternal

warmth and adolescent self-esteem. In practical terms, this suggests that increased levels of

paternal warmth are linked to higher self-esteem among adolescents.

Table 9
Correlation of paternal warmth and adolescent’s self-esteem.

Variable Paternal Warmth

r p

Adolescent’s Self-Esteem 0.24** 0.000

Note. *p<0.05, **p<0.01, ***p<0.001

D2.

Perform t-test to determine sex group differences

The assessment was performed by t-test to determine sex group differences with

respect to paternal warmth and adolescent’s self-esteem. The t-test analyses the group

statistics between the gender groups, paternal warmth, and self-esteem.

 35

a) Paternal warmth

An independent-sample t-test was conducted to compare the paternal warmth scores

for male adolescents and female adolescents. There was no significant difference in paternal

warmth score for males (M = 25.42, SD = 8.18) and females (M = 23.20, SD = 8.45; t (371) =

2.58, p = 0.01, two-tailed). The magnitude of the differences in the means (mean difference =

2.22, 95% CI: 0.53 to 3.92) was very small (eta squared ≈ 0.02).

The decision of significant value p (p < 0.05 = reject Hypothesis , p > 0.05 = failed to

reject Hypothesis). In conclusion, as p = 0.01 (p < 0.05), male adolescents' paternal warmth

was significantly higher than that of female adolescents due to the fact that the mean value of

male adolescents' paternal warmth was higher than the mean value of female adolescents'

paternal warmth. As a result, there was a significant difference in paternal warmth between

male and female adolescents.

Table 10
Paternal warmth between male and female adolescents.

Variable Paternal Warmth

Male Female t p

Mean Mean

Sex 25.42 23.20 2.58** 0.01

 36

Note. *p<0.05, **p<0.01, ***p<0.001

b) Self-esteem

An independent-sample t-test was conducted to compare the self-esteem scores for

males and females. There is no significant difference in self-esteem score for males

(M=21.51, SD=3.86) and females (M=22.13, SD=4.36; t (371) = -1.436, p = 0.152,

two-tailed). The magnitude of the differences in the means (mean difference = -0.613, 95%

CI: -1.45 to 0.226) was very small (eta squared ≈ 0.01).

The decision of significant value p (p < 0.05 = reject Hypothesis, p > 0.05 = failed to

reject Hypothesis). In conclusion, as p = 0.15 (p > 0.05), there is no significant difference

between gender group and self-esteem. The hypothesis failed to be rejected among gender

groups and self-esteem.

Table 11
Adolescent’s self-esteem between male and female adolescents.

Variable Adolescent’s Self-Esteem

Male Female t p

Mean Mean

Sex 21.51 22.13 -1.44 0.15

Note. *p<0.05, **p<0.01, ***p<0.001

 37

D3.

Perform ANOVA test to determine race group differences

A one-way analysis of variance (ANOVA) was conducted to assess the influence of

race group differences on the levels of paternal warmth and self-esteem experienced and

developed by adolescents. The analysis included group statistics for paternal warmth among

different race groups, allowing for an examination of variations in these factors.

a) Paternal warmth

Table 12 shows a one-way within-group ANOVA to discuss paternal warmth by race.

The null hypothesis was that there was no significant difference between the means of

paternal warmth across races. Respondents were divided into three groups according to their

race (Malay, Chinese, and Indian). The result states that there was a significant difference in

paternal warmth between the three races at the p < 0.05 level: F(2, 370) = 12.24, p = 0.00.

Thus, the null hypothesis was rejected and there was a significant difference between the

means of paternal warmth for different races.

Table 12
One-way analysis of variance of paternal warmth by adolescent's race.

Source dƒ SS MS F 𝘱

Between Groups 2 1867.12 933.56 12.24* 0.000

Within Groups 370 24265.61 65.58
Total 372 26132.73

Note. *p<0.05
 38

In Table 13, post hoc analysis using the Tukey Honestly Significant Difference (HSD)

test was conducted to examine mean differences among different racial groups, specifically

Malays, Chinese, and Indians, in relation to the variable of interest. The results revealed

notable distinctions in the mean scores across these ethnic groups.

Table 13 shows the post hoc responses using the Turkey HSD test for Malays (mean:

24.51, standard deviation: 8.13), Chinese (mean: 20.32, standard deviation: 7.73) and Indians

(mean: 29.28, standard deviation: 8.43).

The Tukey HSD test indicated that these differences in means among Malays,

Chinese, and Indians were statistically significant. In other words, there was a significant

variance in the levels of paternal warmth experienced by individuals from these different

racial backgrounds.

Furthermore, the effect size was calculated using eta squared (η²), resulting in an

effect size of .07. This effect size suggests that ethnicity, as a factor, accounts for a notable

proportion of the variability in paternal warmth among the study participants. In practical

terms, the racial background of individuals in this study appears to have a meaningful impact

on the perceived levels of paternal warmth.

Table 13
Multiple comparisons and mean differences in paternal warmth by adolescent's race.

Comparisons Mean Difference Std. Error 95% CI

Malay vs. Chinese 4.19* 1.15 1.49, 6.88

 39

Malay vs. Indian -4.76* 1.44 -8.14, -1.39

Chinese vs. Indian -8.95* 1.70 -12.96, -4.94
Note. *p < 0.05, where p-values are adjusted using the Tukey method.
b) Self-esteem

Table 14 presents a one-way between-groups ANOVA to discuss the relationship

between different races and adolescent self-esteem. The null hypothesis was that there was no

significant difference in self-esteem between the different races. In accordance with their race

makeup, the interviews were split into three groups: Malay, Chinese, and Indian. There was

no significant difference in self-esteem among the three races. This is because F(2, 370) =

0.32, p = 0.73 (p>0.5). Thus, the original hypothesis was accepted and there was no

significant correlation between different races and self-esteem.

Table 14
One-way analysis of variance of adolescent’s self-esteem by adolescent's race.

Source dƒ SS MS F 𝘱

Between Groups 2 10.91 5.46 0.32 0.73

Within Groups 370 6320.76 17.08
Total 372 6331.67

In Table 15, post hoc analysis was conducted using the Tukey Honestly Significant

Difference (HSD) test to examine mean differences among different racial groups,

specifically Malays, Chinese, and Indians, concerning a specific variable. The results

revealed that there were no significant distinctions in the mean scores across these ethnic

groups.

Table 15 shows the post hoc responses using the Turkey HSD test for Malays (mean:

21.89, standard deviation: 4.03), Chinese (mean: 21.44, standard deviation: 4.19) and Indians
 40

(mean: 21.97, standard deviation: 4.75). Therefore, there is also no significant difference in

the mean scores between Malay, Chinese and Indian.

The findings from the Tukey HSD test indicated that these mean differences among

Malays, Chinese, and Indians were not statistically significant. In other words, there was no

substantial variation in the levels of self-esteem experienced by individuals from these

different racial backgrounds.

Additionally, the effect size was calculated using eta squared (η²), resulting in an

effect size of .002. This effect size suggests that ethnicity, in this context, explains only a very

small proportion of the variability in self-esteem among the study participants. Therefore, the

racial background of individuals in this study does not appear to have a significant impact on

the levels of self-esteem they report.

Table 15
Multiple comparisons and mean differences in adolescent’s self-esteem by adolescent's race.

Comparisons Mean Difference Std. Error 95% CI

Malay vs. Chinese 0.45 0.58 -0.93, 1.82

Malay vs. Indian -0.08 0.73 -1.80, 1.64
Chinese vs. Indian -0.53 0.87 -2.57, 1.51
 41

D4.

Perform a Chi-square test to determine the association between sex and race.

The assessment was performed by Chi-square to determine the association between

gender and race. Table 16 shows the relationship between gender and race using a chi-square

test. The null hypothesis states that there is no significant association between gender and

race. The findings of the SPSS analysis showed that there was not a significant relationship

between gender and race, with 𝒳2 (2, n =373) = 5.34, 𝘱 = 0.07. There was no significant link

between gender and race, as the null hypothesis was accepted and the significant value was

greater than 0.05. Symmetric measures, including Phi and Cramer's V, were also examined.

The Phi coefficient was 0.120 with an approximate significance level of 0.07, while Cramer's

V also showed a value of .120 with a significance level of 0.07. These symmetric measures

confirmed the absence of a strong association between gender and smoking status, as the

coefficients were close to zero and the significance levels were greater than .05.

Table 16
Chi-square analysis between and race.

Variable Sex
n (%) 𝒳2 𝘱

Male Female

Race 5.34 0.07

Malay 128 (69.9) 148 (77.9)
Chinese 31 (16.9) 30 (15.8)
Indian 24 (13.1) 12 (6.3)
Total 183 (100.0) 190 (100.0)
 42

MULTIVARIATE ANALYSIS

E1.

Perform a multiple regression analysis on the following variables:

a) Paternal warmth

A multiple regression analysis was conducted to explore the influence of several

predictor variables on paternal warmth. The predictor variables included adolescent age,

adolescent sex, adolescent race, father's age, father's monthly income, and family size as

shown in Table 17. The results indicated that this was a statistically significant model (F(2,

370) = 2.37, p < 0.05). The adjusted R2 value suggested that approximately 2.2% of the

variance in paternal warmth can be explained by the variance in the predictor variables. The

analysis further revealed that among the predictor variables, adolescent sex (β = -0.12*)

emerged as the most influential predictor in the model. In contrast, the father's monthly

income (β = 0.00) appeared to be the least influential predictor in the model. Notably,

adolescent sex (t = -2.39, p < 0.05) and adolescent age (t = -1.98, p < 0.05) were found to be

statistically significant predictors of paternal warmth. Conversely, adolescent age, adolescent

race, father's age, father's monthly income, and family size did not exhibit statistically

significant predictive power in relation to paternal warmth.

Table 17
Multiple regression analysis for paternal warmth.

Variable B SEB 𝛃

Adolescent’s Age -0.88 0.45 -0.10*

Adolescent’s Sex -2.07 0.87 -0.12*
Adolescent’s Race 0.57 0.71 0.04
Father’s Age -0.10 0.08 -0.07
Father’s Monthly Income 1.08E-5 0.00 0.00
Family Size 0.11 0.27 0.02
 43

constant 41.18 6.79

Note. R2 = 0.04; Adj. R2 = 0.02; F = 2.37; p < 0.05
*p<0.05, **p<0.01, ***p<0.001

b) Self-esteem

In the multiple regression analysis for adolescent's self-esteem (Table 18), the effects

of several predictors were assessed, including adolescent age, adolescent sex, adolescent race,

father's age, father's monthly income, and family size. The model proved to be statistically

significant (F(2, 370) = 3.40, p < 0.01), and the adjusted R2 indicated that 3.7% of the

variance in self-esteem could be explained by the predictor variables. Notably, father’s

monthly income (β = 0.13*) and adolescent’s age (β = 0.13*) emerged as the most influential

predictors in the model, while adolescent’s race (β = -0.03) was the least influential predictor.

Furthermore, father's monthly income (t = 2.53, p < 0.05) and adolescent's age (t = -2.54, p <

0.05) were deemed significant predictors of self-esteem. In contrast, adolescent’s sex,

adolescent’s race, father's age, and family size did not exhibit statistically significant

associations with self-esteem. The findings suggest that higher father's monthly income and

older adolescents were associated with higher levels of self-esteem.

Table 18
Multiple regression analysis for adolescent’s self-esteem.

Variable B SEB 𝛃

Adolescent’s Age -0.55 0.22 -0.13*

Adolescent’s Sex 0.72 0.42 0.09
Adolescent’s Race -0.18 0.35 -0.03
Father’s Age 0.04 0.04 0.07
Father’s Monthly Income 0.00 0.00 0.13*
Family Size -0.23 0.13 -0.10
constant 27.65 3.32
Note. R2 = 0.05; Adj. R2 = 0.04; F = 3.40; p < 0.01
*p<0.05, **p<0.01, ***p<0.001
 44

Conclusion

The findings from this study shed light on the intricate relationships between various

factors, including adolescent age, parental warmth, parental income, family size, and

self-esteem. Through a comprehensive analysis, the study aimed to understand the

multifaceted dynamics influencing adolescents' self-esteem and their perceptions of parental

warmth.

In the initial stages of our analysis, we examined bivariate relationships among these

variables. We observed significant correlations and differences in certain pairs. Notably, there

was a negative correlation between adolescent age and paternal warmth, indicating that older

adolescents perceived lower levels of paternal warmth. Furthermore, father's monthly income

was positively correlated with adolescent self-esteem, suggesting that as income increased, so

did self-esteem. These initial findings laid the foundation for more in-depth analyses.

Subsequently, we explored group differences using t-tests and ANOVA. Gender

differences emerged, revealing that male adolescents experienced higher levels of paternal

warmth compared to their female counterparts. However, no significant differences were

observed between genders concerning self-esteem. When assessing race group differences,

we found that race did not have a significant impact on self-esteem, while paternal warmth

significantly varied among different races, with Indian adolescents perceiving higher levels of

warmth.

Our examination of the association between gender and race revealed that these two

variables were not significantly related, implying that gender distribution was similar among

different racial groups.

 45

To provide a comprehensive understanding of the interplay between these factors, we

conducted multiple regression analyses. For paternal warmth, adolescent sex and age were

identified as the most influential predictors. Specifically, female adolescents and older

adolescents tended to experience lower levels of paternal warmth. In contrast, for self-esteem,

father's monthly income and adolescent age were the prominent predictors. A higher monthly

income and older age were positively associated with self-esteem.

In conclusion, this study has contributed valuable insights into the factors that shape

adolescents' self-esteem and perceptions of parental warmth. While adolescent age, gender,

and income are significant factors in influencing these outcomes, it is vital to consider the

complex interplay among multiple factors when assessing the emotional lives of adolescents.
 46

References

Backes, E. P., & Bonnie, R. J. (2019). Adolescent Development. Nih.gov; National

Academies Press (US). https://www.ncbi.nlm.nih.gov/books/NBK545476/

DeVellis, R. F. (2012). Scale development: Theory and Applications. SAGE Publications.

Gravetter, F. J., & Wallnau, L. B. (2016). Statistics for The Behavioral Sciences. Cengage

Learning.

Leung Ling, M. T. W., Chen, H. F., & Chiu, K. C. N. (2019). Parental Warmth and

Involvement and the Self-Esteem of Young People in Hong Kong. Child Indicators

Research. https://doi.org/10.1007/s12187-019-09645-3
 47

FACULTY OF SOCIAL SCIENCES AND LIBERAL ARTS

SP110 STATISTICS FOR PSYCHOLOGY
SEPTEMBER - OCTOBER 2023
ASSIGNMENT SCORING RUBRIC
Subject : SP110 Statistics for Psychology
Lecturer : Ms. Syarifah Maisarah Syed Alwi
Task : Data Analysis Report (20%)
Student : Name: Eddie Yap ID: 1002267012
Name: Chew Hooi Yin ID: 1002267005
Name: Guo Jijie ID: 1002161231
Name: Joserine Chong Zoe Ling ID: 1002370728

Category 4-5 2-3 0-1 Points

Statistical Statistical methods are fully and correctly applied. Some statistical methods are applied but with Statistical methods are completely misapplied.
Method (10%) errors.

Hypotheses The hypotheses are clearly stated and Hypotheses are clearly stated but the relevance is Hypotheses are not/ somewhat stated and the
(10%) relevant. somewhat unclear relevance is somewhat unclear

Data (20%) Accurate representation of the data in tables and Accurate representation of the data in written Data are not shown or are inaccurate
written form. Tables are labelled and titled. form but no table are presented.

Results & Results are discussed, together with decision, Results are discussed but no decision, Results are not discussed.
Discussion (30%) conclusion and interpretation. conclusion and interpretation.

Organization All information is exceptionally organized, neat Most information is organized, neat and easy Information is unorganized and hard to
(10%) and very easy to read. to read. understands.

Mechanics (10%) Written work has 1-2 errors in word selection and Written work has 3-4 errors in word selection Written work has serious and persistent errors
use, sentence structure, spelling, punctuation, and and use, sentence structure, spelling, inword selection and use, sentence
capitalization. punctuation, and capitalization. structure,spelling, punctuation, and capitalization.

Reference (10%) Cites all data obtained from other sources. APA Cites some data obtained from other sources. The report has very limited citation.
citation style is accurate. Citation style is either inconsistent or incorrect.

Total /100
Percentage (%) /20
48