Professional Documents
Culture Documents
BED 101 RESEARCH Composite Slides 2-1
BED 101 RESEARCH Composite Slides 2-1
Topic:
The world is facing some serious issues which different countries and communities need to
respond to urgently. Among these issues re: Climate Change; Disaster and Risk
Management; Poverty; Health; and Gender Equality.
Select ONE of the issues stated above. Design a research proposal on how you would go about
researching on the impact of that issue in a local community of your choice (urban or rural).
Use the following structure for your proposal:
Background/Motivation of the study (half a page maximum) (15)
Statement of the problem (a paragraph) (7)
Research questions (at least three) (12)
Significance of the study (half a page maximum) (10)
Assumptions of the study (at least three) (6)
The research design (half a page maximum) (15)
Population, sample and sampling procedures (a page maximum) (20)
Description of the data collection instruments that you would use (half a page max) (15)
You are advised to adhere to the maximum lengths indicated for each aspect of your
proposal.
Presentation: Typed, using Times New Roman; Font Size 12; and 1.5 Line Spacing.
Due Date: …………………….
Research in Education
2. Tradition
Refers to culture, religion etc as sources of
knowledge
E.g. manner of greeting elders, dress code, communication
modes.
Long established customs are sources of knowledge
3. Experience
Our personal experience in matters of problem
solving is a fundamental source of knowledge
E.g. experience in teaching particular topics within a subject
discipline, handling indiscipline at school level
Sources of knowledge (contd)
immediate solutions
Localised in nature; replication of other studies
Types of research (contd)
science
Teachers who are action researchers are constantly observing
E.g.
The main Research question
What are the levels of poverty that influence learners’ learning
and performance in Lupane District secondary schools ?
Sub – research questions
1. What are the levels of poverty in Lupane rural communities?
2. How does poverty influence learners’ learning and
performance in Lupane secondary schools?
3. What strategies can be put in place to minimize the negative
effects of poverty on learners’ learning and performance?
Research questions (contd)
Structured interviews:
are, essentially, verbally administered questionnaires
a list of predetermined questions are asked, with little or no
variation and with no scope for follow-up questions to
responses that warrant further elaboration.
They are relatively quick and easy to administer
Data collected can be quantified for further analysis
Types of Interviews (contd.)
Unstructured interviews:
Are conducted with very few or no pre-set questions
Are conducted in a conversational manner
Questions asked are open-ended and have follow-up questions
Interviewee does most of the talking while interviewer listens
without judging the responses
Types of Interviews (contd.)
Semi-structured interviews:
Consist of several key questions that help to define the areas to
be explored
Have some pre-set and logically organized questions
Allow the interviewer or interviewee to diverge in order to
pursue an idea or response in more detail
Can include some closed questions
Data collected is qualitative and may be difficult to quantify
Observations
Quantitative observations:
are more structured,
standardized,
Complete participant:
the researcher becomes a member of the group being studied and
does not tell the members that they are being studied
Participant-as-observer:
the researcher spends extended time with the group as an insider
and tells the members they are being studied
Observer-as-participant:
the researcher spends a limited amount of time observing group
members and tells members that they are being studied
Complete observer:
the researcher observes as an outsider and does not tell the people they are
being observed
Tests
Used in research to measure personality, aptitude,
achievement, and performance.
A test must measure what it is supposed to measure
Specific knowledge, skills, behavior, or cognitive
activity can be measured using tests
QUALITIES OF A GOOD TEST
Validity: in general a test possesses validity to the extent that it
measures what it claims to measure.
Reliability: a test is reliable to the extent that it measures
a. Variable
- characteristic of data that can take on different
values e.g. age, teacher qualification, test score
b. Parameter
- a numerical measure to describe the characteristics
of a population e.g. popn standard dev (σ) popn mean (
µ)
c. Statistic
- a numerical measure to describe the characteristics
of a sample. e.g. sample standard dev (s)
Key terms - contd
d. Data
raw materials of research.
can be quantitative or qualitative
suppose you are measure height in research (H)
H 1.4 m (i.e. quantitative – thus objective)
or
Very short (i.e. qualitative – thus subjective)
A. Descriptive statistics
1.1 The distribution of data
Bivariate Data:
Switch on calculator
Press ‘Mode’
Select ‘1’ for Statistics mode
Select ‘1’ for Two Variable Statistics
Then start entering your data, in pairs
Enter first number, press ‘STO’, enter second number, press
‘M+’
Repeat for all the pairs until the last pair
To retrieve your summary statistics:
Press ‘RCL’ then the statistic that you want, e.g. ‘r’ for
correlation coefficient (on division sign)
Example
Σ = sum of
M = mean of all scores
n = sample size (number of scores)
Types of distributions
Positive skew
Negative skew
Normal /symmetrical distribution
Skewness is the pattern of the distribution of
scores
It shows whether the majority of scores are high or
low or whether half of the scores appear below or
above the average
Positive skew
Positively skewed distribution
E.g
Negatively skewed distribution
E.g.
Normal distribution
Measure of association
Degree/strength of the linear relationship between
variables
1. Scatter graph
simplest way of representing bivariate data
it has two axes: x –axis showing the independent
variable and the y-axis showing the dependent
variable.
Measure of association-contd
Student A B C D E F G H
RES 16 18 13 26 21 15 35 28
METH
STATS 18 15 15 14 16 19 10 13
Scatter graph
20
18
16
14
12
10
10 15 20 25 30 35 40
Graphical Analysis of Associations
Student A B C D E F G H
RES. 16 18 13 26 21 15 35 28
MET
STATS 18 15 15 14 16 19 10 13
Solution
Rank x Rank y d d2
6 2 4 16
5 4.5 0.5 0.25
8 4.5 3.5 12.25
3 6 -3 9
4 3 1 1
7 1 6 36
1 8 -7 49
2 7 -5 25
148.5
Cont...
Student A B C D E F G H I J
Stats 57 77 64 62 45 66 58 60 40 80
Test
Types of correlation coefficients (contd)
Cont-
Student A B C D E F G H
RES. 16 18 13 26 21 15 35 28
MET
STATS 18 15 15 14 16 19 10 13
In-class task1
Student 1 2 3 4 5 6 7 8
Entry 90 48 51 60 41 63 55 81
test
Maths 22 6 21 13 14 22 9 33
score
Inferential statistics
Key terms:
hypothesis: null & alternative,
types of errors,
level of significance,
degrees of freedom
Inferential Statistics -contd
Suppose
A certain man was arrested for stealing from the
bank and was to appear in court.
The judge has two options; either to convict or acquit
the man of the charges based on evidence provided
by the prosecutor.
1. Hypotheses
Ho : The man is innocent
HI : The man is guilty of the charges
Inferential Statistics -contd
t=
Steps in hypothesis testing
, , n1 = 12 , , n2 = 10
t=
Solution – example 1
Step 2
α = 0.025
Step 3
t-calc = 10.07 (2d.p)
t-crit = t(α;df) = t(0.025;n1+n2 -2)= t(0.025; 20)
t-crit = 2.09
So t-calc > t-crit
Solution - contd
Step 4
Since t-calc > t-crit we reject Ho in favour of H1.
Step 5
Therefore there is sufficient evidence that students
who sit for the calculator version perform better than
those who sit for the non-calculator version tested at
2.5% level of significance.
Non-parametric test
The table below shows the numbers of pupils who passed and failed Mathematics mock
examination at Gray school. The school practices streaming, in which case Form 3A is the
best class and Form 3C is the weakest class.
Ability
[Formula: ]
Solution
Step 1
H0 : Ability and outcome of the mock test are
independent
H1 : Ability and outcome of the mock test are
dependent
Step 2
α = 0.05
Solution - contd
Step 3
Compute E = (R.T X C.T)/ G.T for each observed
freq. (O)
Step 3 – contd
O E
23 17 36 2.1176
13 14 1 0.0714
7 12 25 2.0833
11 17 36 2.1176
15 14 1 0.0714
17 12 25 2.0833
8.5446
Solution - contd
Solution -contd
Step 5
Therefore ability and outcome of the mock test are
dependent tested at the 5% level of significance. That
is knowledge about one factor can help predict the
other factor.
Example 1 for practice
Exam grade
Commercials 22 16 8 5
Arts 3 10 12 24
Example 2
1) A researcher wanted to establish if there is a significant difference in pass rate between mission
and government primary schools at Grade 7 level. she randomly selected samples of 12 schools
each and obtained the following summary statistics;
= 87 % = 74 %
=5% =8%
Carryout a test at 5% level of significance to determine if mission schools perform better than
government schools at Grade 7 level. [30]
[t-cal = ]