Research Designs Used in Quantitative Research: Descriptive Research Design Experimental Research Design

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 55

Research Designs Used

in Quantitative Research

 Descriptive research design


 Experimental research design
Descriptive Research Design
 the type of design that can be
quantitative and qualitative.
 when used in quantitative research it
will enable the researcher to describe
events (as to effects, satisfaction,
awareness, perception, etc.) or
definite correlations, inferences and
causal relationships.
Descriptive Research Design
Methods of Data Collection Statistics

 survey and observational checklist are


the most common tools used in this
study.
 reports of this kind of research design
usually reveals a lot based on measures of
central tendency, percentages,
correlations and variations.
Experimental Research Design

 Creswell (2012) defined experimental


research as intervention study or group
comparison study because of its ability
to test groups using different
interventions and to determine if
significant effects or changes happened
after the intervention.
Experimental Research Design

Methods of Data Collection

 post-test design – you must have a


control and an experimental group.
Treatment is applied directly to the
experimental group. Then a post-test
will be given to both groups to assess
the effect of the manipulation.
Experimental Research Design
Methods of Data Collection

 pre-test post-test design – the only


difference in this design is the pre-test to
both dependent and independent group.
This allows you to see if there is a big gap
of change between the dependent and
independent variable after the intervention
(treatment) was introduced and
implemented to the experimental
(dependent) group.
Writing Your Questionnaire
1. First, specifically identify your variables
in your research questions and
hypothesis.
Example: Level of Awareness of Different
Political Issues Among Grade 9 Students
in the School Year 2016-2017
What is the variable?

Level of Awareness of Different Political


Issues Among Grade 9 Students in the
School Year 2016-2017
Writing Your Questionnaire
Example: Level of Awareness of Different
Political Issues Among Grade 9
Students in the School Year 2016-2017

Instrument?

Researcher-made questionnaire
using 4.3,2,1 scale
Writing Your Questionnaire
Example: Level of Awareness of Different
Political Issues Among Grade 9
Students in the School Year 2016-2017
Interpretation of Scores:

4 Very much aware


3 Aware but does not understand the
issue
2 Not aware but have heard of the issue
1 Not aware of the issue
Examples of Likert Scale
 Agreement (Strongly Agree, Agree,
Undecided, Disagree)
 Importance (Very Important,
Important, Of Little Importance, Not
Important)
 Likelihood (Always True, Usually True,
Not True, No Idea at all)
 Awareness (Always Aware,
Moderately Aware, Aware, Not Aware)
Second, identify the instrument that
you will use
Am I measuring individual’s ability using a / an
____quiz
____achievement test
____subject mastery exam
____aptitude test
____interest inventory
____personality assessment

Am I after attitudinal measures?


______awareness _____perception
______satisfaction _____acceptance
Second, identify the instrument that
you will use

What type of questionnaire should I utilize?


____researcher made
____standardized questionnaire
____published questionnaire
____published questionnaire
____modified questionnaire
Second, identify the instrument that
you will use
Will I gather personal information among my
respondents?
____gender/sex
____age range or exact age
____socio-economic status
____highest educational attainment
____religion

How will I conduct my data gathering?


______actual/face-to-face _____mail surveys
______phone surveys _____online surveys
 Researcher-made questionnaire is a
questionnaire made and constructed by
the researcher himself. Its construction
must be guided by his research purpose
and research questions.

 Standardized questionnaire is an
instrument made by an established
assessment firm and is known for its
reliability and validity.
 Modified questionnaire can come from
published research materials and modified
or changed, upon permission of the
author, to best fit the needs of the
researcher. Creswell (2012) noted that
typically “authors of the original
instrument will ask for a copy of the
modified version and the result of the
study after their instrument was
borrowed and used in another study”.
Pre-test and Post-test Research
Design
Classic Experimental Design
 involves the selection and assignment of
two comparable groups – the
experimental and the control group.

 the experimental group is subjected to


the independent variable while the
control group is not.
Classic Experimental Design

Pre-test Intervention Post-test


Group A

Experimental
group

Group B
Control
group
Pre-test and Post-test Research
Design
Post-test Only Control Group Design
 primarily aims for randomization.
 the researcher can randomize his/her
samples into groups by sampling.
 since there is no pre-test stage, there is no
interaction effect to be measured.
 the advantage of this design, is that the
researcher can test more than two groups.
Post-test Only Control Group Design

Pre-test Intervention Post-test


Group A

Experimental
group

Group B
Control
group
Pre-test and Post-test Research
Design

Solomon Four-Group Design


 as the name suggests, involves the
selection of four groups
Pre-test Intervention Post-test
Group A
Experimental v
group
Group B
Control group v v

Group C
Experimental v v
group
Group D
Control group v

The result of the post-test of Group C and Group D


allows the researcher to see if pretesting affects post
testing. If the result of the post-test of Group C and D is
different from that of Group A and B, then the researcher
can assume that the pre-test influences the post-test.
Pre-test Intervention Post-test
Group A
Experimental v
group
Group B
Control group v v

Group C
Experimental v v
group
Group D
Control group v

The comparison between the post-test of Group B and


Group D can help the researcher determine if external
factors influenced the effect.
Pre-test Intervention Post-test
Group A
Experimental v
group
Group B
Control group v v

Group C
Experimental v v
group
Group D
Control group v

Upon comparison of the post-tests of Groups A and C, if


there is a difference in the result, the researcher can
assume that the pre-test is affected the post-test and the
research is flawed.
Pre-test Intervention Post-test
Group A
Experimental v
group
Group B
Control group v v

Group C
Experimental v v
group
Group D
Control group v

The post-test of Group B as compared to Group D can


reveal if the pre-test is affecting the behavior, which is
independent of the treatment. If the results lead to a
significant difference then the pre-test is affecting of
influencing the over-all results and is in need of
refinement.
Pre-test Intervention Post-test Pre-test Post-test
results results

Group A
Exp. v 6 12
Group B
Control 6 8
Group C
Exp. 7
Group D
Control 3

A student who took the pre-test and was


subjected to the intervention had a higher score
in the post-test.
Pre-test Intervention Post-test Pre-test Post-test
results results

Group A
Exp. v 6 12
Group B
Control 6 8
Group C
Exp. 7
Group D
Control 3

A student who took the pre-test but never


underwent intervention also had a higher score in
the post-test.
Pre-test Intervention Post-test Pre-test Post-test
results results

Group A
Exp. v 6 12
Group B
Control 6 8
Group C
Exp. 7
Group D
Control 3

A student who had no pre-test but underwent


intervention got a good score.
Pre-test Intervention Post-test Pre-test Post-test
results results

Group A
Exp. v 6 12
Group B
Control 6 8
Group C
Exp. 7
Group D
Control 3

A student who was not subjected to the pre-test


and received no intervention had a very low
score.
Common Statistics Used in a
Descriptive Study

 Mean or Average
 Median
 Mode
 T-test
 ANOVA
 Correlation
Common Statistics Used in a
Descriptive Study
 Mean or Average is the total of
all scores divided by the total
number of scores. To find the
mean, add up the values in the
set of scores and then divide the
sum by the number of values
that you added.
Common Statistics Used in a
Descriptive Study

Example: On a quiz of 10 items,


the scores are 4,5,8,7,7,9. What
is the mean score?
Common Statistics Used in a
Descriptive Study

 Median divides the total number


of scores into two groups. To find
the median, arrange the scores in
numerical order from least to
greatest. Find the score that
appears in the middle of the list.
Common Statistics Used in a
Descriptive Study
 Note: If the number of scores or
data is an even number, add the
two middle terms and divide the
sum by 2.
 If the total number of score
resulted to an odd number, simply
get the middle term, or the middle
score, or middle data.
Common Statistics Used in a
Descriptive Study
 Mode or modal score is the
score that appears most often in
a group of scores. An array of
scores or data can have one
modal score or unimodal. Some
can have two modals or bimodal,
others have multiple modals.
Common Statistics Used in a
Descriptive Study

Suppose the array scores are:


3,3,4,4,4,4,3,5,5,5,7,7,7,7,8,9
Common Statistics Used in a
Descriptive Study
Lights Challenges

Measures of central The mean can be


tendency provide us easily affected by
with the most typical outliers or can
or the usual score change abruptly
that is found in an when there are
array of scores. extreme scores.
Lights Challenges

Example: If your study will Example: The weights of 5


delve on the level of representatives from a
achievement, awareness, class were recorded as 43
or satisfaction through kg, 43 kg, 47 kg, 35 kg, 55
achievement tests or kg with a mean of 44.6 kg.
surveys using the MEAN
can be very powerful in Suppose that the same
showing the most respondents have their
common score or weights recorded after 2
response of the weeks as 43 kg, 43 kg, 59
respondents. kg, 35 kg, 53 kg and a
mean = 44.6 kg.
Lights Challenges

In the study “Satisfaction of Notice that when there is a


the Students in the Food score that is too low or too
Services Provided by the high the mean changes
Canteen,” the category abruptly. The effect of
“Food Variety” has a mean outliers is such that they
of 3.4 which falls under the can simply pull the average
rating Quite Satisfied. From up or down.
this result, the researcher
can have an idea that the
students are “quite
satisfied” with the food that
the canteen is serving them.
Lights Challenges

On the other hand, the The median can be defined


study “Student as the 50th percentile of
Understanding of Historical scores. However, the
Concepts” revealed a mean median can be insensitive to
of 13 from a 30-item test. It the distribution of scores.
is quite alarming since the Thus, a distribution group
mean score of 13 doesn’t with a different set of
meet the 50% passing score scores might have the same
of 15. It means that most of mean regardless of how
the students had scores high or low the scores in
below the passing mark. the set.
Lights Challenges

Remember: The mean or The mode, through the


average can be a good easiest to find, oftentimes
indicator of change (that provide an insignificant
can be attributed to a description of the sample
shift above or decline). scores. It is insensitive to
This can be attributed to values of scores in the
factors causing its ups and distribution and gives a
downs. narrow scope as to its
data.
GETTING THE MEANOF A
LIKERT SCALE

By computing for the weighted


average – this allows the responses
to be computed based on their
respective weights or value.
GETTING THE MEANOF A
LIKERT SCALE
For example, in the study “Satisfaction of the
Students in the Food Services Provided by the
Canteen” the criteria for Food Variety reveals
the following data:
Food Variety 4 3 2 1
Very Satisfying Non Inapplicable
Satisfying Satisfying
The school 13 17 2 5
canteen is
constantly
reviewing the
food to be
served for
every day.
GETTING THE MEANOF A
LIKERT SCALE
Verbal Interpretation Legend

1.00 – 1.75 INAAPLICABLE


1.76 – 2.50 NOT SATISFYING
2.51 – 3.25 SATISFYING
3.26 – 4.00 VERY SATISFYING
TESTING THE HYPOTHESIS
Most research work would require you to prove
if the results of your study agree with your
hypothesis or not.

It was noted that researches conducted by


students in the Philippines almost always end
with the disapproval or rejection of the null
hypothesis. Most research start with a null
hypothesis and nothing follows.
TESTING THE HYPOTHESIS
But what must you do? Go back to your goal.
Do you want to prove that your intervention
will strengthen the behavior of the
respondents? Or do you want to see their
satisfaction over the services provided by an
agency? In this case, you must write an
alternative hypothesis. On the other hand, if
your goal is to disprove a certain practice, you
need to write a null hypothesis.
TESTING THE HYPOTHESIS

A null hypothesis is a hypothesis that


states there is “no statistical significance”,
mean? Should you use the term
“significant difference” or “significant
relationship”?
TESTING THE HYPOTHESIS

Using the term “statistical significance”


means that the researcher is very sure
that the statistic in his/her study is very
reliable. It doesn’t mean that the findings
or results of the research are important
or useful. Significance is a statistical term
that says how sure you are that a
difference or relationship exists.
TESTING THE HYPOTHESIS
The term “significant difference” is used
when you want to test whether an
intervention has different effects toward
different groups. Say for example, the
learning achievement of students who
attend school regularly versus those who
are enrolled in an online school. Common
statistical treatment used if your hypothesis
is asking for "significant difference” is the t-
Test and Analysis of Variance or ANOVA.
TESTING THE HYPOTHESIS

The term “significant relationship” implies


a correlation between the two variables
in a study. The most commonly used
statistical test for this is the Pearson
Product-Moment Correlation or
Spearman’s Rho Test.
t-TEST

Comparing the weekly study time (in


hours) of male and female students. The
purpose is to determine whether the
weekly study time of female students is
higher than the weekly study time of
male students. The test is to use 0.05
level of significance.
t-TEST
 If the t-stat value (p-value) is less than
the value of the t-critical one-tail and t-
critical two-tail. Accept the null
hypothesis (no significant difference)
 If the p-value of one-tail and two-tail are
greater than the alpha level (0.05).
Accept the null hypothesis (no significant
difference)
t-TEST

 After making a new classroom


intervention, Teacher Gaddi gathered the
results shown in the table.
ANOVA
Analysis of Variance

It is the study of the effects of the


independent variable.
ANOVA
 If the computed ANOVA (F) is higher
than the critical value. Fail to accept the
null hypothesis (there is significant
difference)
 If the p-value is less than the alpha level
(0.05). Fail to accept the null hypothesis
(there is significant difference)
REGRESSION ANALYSIS
Regression analysis can provide us the
relationship of the dependent and
independent variables, estimate their
relationship, and determine if the
relationship is significant.

 It can validate our claim or theory.

You might also like