Professional Documents
Culture Documents
DEFINITION OF
CONCEPTS, VARIABLES AND
LEVELS OF MEASUREMENT
Statistics
Number of Activities
Demographic
No 1-3 4-6 7 or More
Characteristics
Activities Activities Activities Activities
Age Group
12 or 13 5.8 26.4 32.2 35.7
14 or 15 7.6 27.0 32.0 33.4
16 or 17 9.5 27.9 30.0 32.6
Gender
Male 8.5 30.4 32.8 28.3
Female 6.7 23.7 30.0 39.6
Family Income
Less Than Php20,000 13.0 33.7 30.5 22.9
Php20,000 to Php49,999 9.4 30.7 31.5 28.4
Php50,000 to Php74,999 6.1 25.1 31.4 37.4
Php75,000 or More 3.6 20.7 31.8 43.9
Descriptive Research
Participation in Youth Activities and Substance Use
Sample Table of Percentile Norms
Raw Score f cf PR
50 3 40 99.9
45 7 37 92.5
40 5 30 75
35 5 25 62.5
30 6 20 50
25 4 14 35
20 5 10 25
15 4 5 12.5
10 1 1 2.5
Correlation and
prediction
Correlation and prediction
Example:
• A researcher wish to determine the extent to which
personal, socio-economic and psychological factors
influence the academic performance of BS Psychology
students.
Personal
Attributes
Academic
SES
performance
factors
Psychological
factors
Basic Considerations in Data Analysis
Typology (developed
by American
psychologist Stanley Smith Stevens in
1946):
Nominal
Ordinal
Interval
Ratio
Levels of Measurement
Nominal – differentiates between items or subjects
based on categories
gender: male, female
age
scholastic aptitude test scores
height
time of day
length
Cont: Measurement Scale
Ratio – has fixed intervals between scores, and has
a fixed zero point, which means that the values can
be compared with each other with zero as a
reference point.
amount of savings;
crime incidence;
number of convictions
bacteria in a specimen
number of passers in the board exams
number of children
years of work experience
sales figures
Measurement Properties of Measurement
Scale
Categorical Hierarchical Equal- True Zero-
(Classification) (Rank-order) distance point
Nominal √ x x x
Ordinal √ √ x x
Interval √ √ √ x
Ratio √ √ √ √
Number of Sample
various programs
Quality of education of PUPSPC as assessed by
Dichotomous Dichotomized
Variable – a Variable – a
categorical variable that continuous variable
has been divided into that has been divided
two categories (e.g. into two categories
pass and fail). (e.g. poor and not
poor).
INDEPENDENT VS. DEPENDENT
Independent Variable –
the presumed cause in a Independent
study; a variable that can Variable
be used to predict or
explain the values of
another variable.
Dependent Variable –
the presumed effect in a
study; the variable whose Dependent
values are predicted by Variable
the independent variable.
Effect of Computer-Aided
Instruction on Achievement Effect of Music on Mood
in Mathematics among
Grade Four Pupils
Music
(classical, pop, rock)
Computer-Aided
Instruction
Achievement in
Mood
Mathematics
MEDIATING VS. MODERATING
X Y
Interactional justice as a mediator Pain, physical activity and the
of the relationship between pay mediating effect of self-
for performance and job efficacy among athletes
satisfaction
Pay
Self-efficacy
Interactional
Justice
(participation
in pay Physical
systems) Pain
activity
Job-Satisfaction
MEDIATING VS. MODERATING
X Y
Moderating A moderator may
Variable – a variable be qualitative (gender,
that influences or education, etc.) or
“moderates” the relation quantitative (IQ level,
between two other age, income, etc.)
variables and thus variable that affects the
produces an interaction direction and/or
effect; the presence of strength of relationship
this third variable (whether causal or
modifies the original correlational) between
relationship between the an independent (or
independent variable and predictor) and a
the dependent variable. dependent (or
criterion) variables.
Intelligence
Stress D
Computer- Other E
Gender School
Aided P
Instruction Factors
R
Social E
Support
S
S
I
Achievement in Stress + O
Mathematics Social N
support
Alternate path diagram
representations of the
moderation model.
Why sample?
Time: Census may be financially feasible but may take too long
to complete, which may seriously reduce the value of the results
Systematic Sampling
• Interval sampling
Purposive Sampling
Snowball Sampling
Hypothesis Testing
Two distinct statistical functions
Accept H0 Reject H0
H0 is No error Type I
State of true Correct Decision Wrong Decision
population
H0 is Type II No error
false Wrong Decision Correct Decision
Significance Level
Non-directional hypothesis
Ho: π = πo
Ha : π ≠ πo
For example, if
you are testing at the
5% significance level
in two-tail, then the
critical region is ±
1.96 standard
deviations above and
below the mean.
Z Critical Values
98% 2% 2.33
99% 1% 2.575
One-tailed
Directional hypothesis
Right-tailed Left-tailed
Ho: π = πo Ho : π = πo
Ha : π > πo Ha : π < πo
Level of Significance
and the Rejection Region
H0: m 1 m2 a Critical
H1: m 1 < m2 Value(s)
Rejection 0
Regions a
H0: m 1 m2
H1: m 1 > m2
0
a/2
H0: m 1 = m2
H1: m 1 m2
0
Hypothesis Testing Procedure
1. State the null hypothesis and alternate
hypothesis.
2. Choose test statistic.
3. Specify level of significance α.
4. Decide on a sample size.
5. Define the critical region in terms of test
statistic.
6. Compare the observed/calculated value of the
test statistic with the critical value and decide to
accept or reject the null hypothesis.
MEASURE OF CENTRAL TENDENCY
DECISION MATRIX
SAMPLE STATISTICAL TOOLS DECISION MATRIX
FOR TESTING SIGNIFICANCE OF DIFFERENCE
Sample Scale
Number Type Nominal Ordinal Interval/Ratio
Chi-square test of Kolmogorov-Smirnov t test
One goodness-of-fit test of goodness-of- z test
sample Binomial test fit
Runs test
Chi-square test of Mann-Whitney U test t test for independent
Inde- independence (for Kolmogorov-Smirnov sample
Two- pendent large sample) for two samples
sample Fisher’s Exact test (for Wald-Wolfowitz for
small sample) two samples
Depen- McNemar Test Wilcoxon Signed- t test for related sample
dent (for dichotomous data) Rank Test Sandler’s A test
SPSS Applications
Parameter vs. Statistics
Statistics – characteristics of a
sample (numerical descriptive
measures computed from a sample).
What is SPSS?
A computer application that supports statistical analysis of
data
1968: the Statistical Package for the Social Sciences (SPSS)
1975: incorporated as SPSS, Inc.
Statistical Products and Service Solutions
2009: changed product name from SPSS to Predictive
Analytics Software (PASW).
2010: acquired by IBM Corporation for US$1.2 billion, and
changed the name of the software to IBM SPSS Statistics.
2014: IBM SPSS Statistics version 22.
What is SPSS?
widely used program in the social sciences; used also by health
researchers, survey companies, government, education
researchers, marketing organizations, data miners, and others.
Perhaps, the most user-friendly statistical software
Statistics included in the base software:
Descriptive statistics: Cross tabulations, Frequencies, Percentages,
MCT, Mvar
Correlations: bivariate, partial
Comparing means: t-tests, ANOVA
Generalized Linear Model: Two-way ANOVA
Non-parametric tests: Chi-square, U test, H test, Friedman ANOVA
Reliability analysis: Cronbach‟s alpha
Prediction for numerical outcome: Linear regression
Prediction for identifying groups: Factor analysis, Cluster analysis,
Linear discriminant analysis, Logistic regression
When using SPSS:
Consider the research design you employed, the
number of variables you manipulated and/or
measured, and the type of data you have collected
codes, etc.
Click
Assigning/Naming Variables
Gender 1 = Male
2 = Female
Example 2:
Variable Name Variable Label
Performance 1 = Poor
2 = Fair
3 = Good
4 = Very Good
5 = Excellent
Statement of Problems:
Determining
1. What are the demographic the
characteristics of the respondent respondents‟
police personnel in terms of: profile
1.1 gender
1.2 age
Measuring
2. What is the level of job performance the two major
of the respondent police personnel? variables:
job
3. What is the level of job commitment performance
of the respondent police personnel? and job
commitment
Gender f %
Age f %
40 to 49 years old
216 32.0
50 to 54 years old
127 18.8
Total 676 100.0
Table 3. Level of Performance of Police Personnel
r p Decision
Mean
Variable Male Female Mean df t p Decision
Police Police diff.
n = 569 n = 107
Job 85.545 86.201 -.661 674 -2.030 .043 Reject H0
performance
Job 85.589 85.280 .308 674 0.914 .361 Do not Reject
H0
commitment
Table 7. Summary of ANOVA on Job Performance
of Police Personnel across Police Districts