GMS 2014 Module 2

INTRODUCTION
The main purpose of this module is to introduce you to statistical method in research. In
order to effectively handle this task, the module has been divided into eleven units. Unit one is
an introduction to statistics. Unit two looks at the measures of central tendency while unit three
focuses on measures of dispersion. Unit four gives an introduction to probability. Unit five
discusses frequency distributions. Unit six is an introduction to estimates with samples. Unit
seven focuses on hypotheses testing. Unit eight looks at analysis of variance (ANOVA). Unit 10
focuses on regression. Unit eleven presents the format for the PAS 2014 research report and
gives guidance on how to write the research report.
AIM OF THE MODULE
The aim of this module is to introduce you to statistical methods in research.
OBJECTIVES OF THE MODULE
By the end of this module, you should be able to:
Explain the major statistics concepts used in research.

Demonstrate an understanding of descriptive statistics.
Acquire knowledge of inferential statistics.
Analyze data and write a research report
TIMES REQUIRED
It will take you approximately three months to undertake the requirements of this
module. The time should be spent on studying the module, reading recommended material, doing
self-help questions and writing the research report.
1
STUDY SKILLS
As a distance student, your study skills will be different from students that are resident
at the university. In your case, you will choose what to study, when to study and where to study
from on your own. You will also need to fit your study activities around other professional
and/or domestic responsilities. It is therefore important for you to allocate enough time to your
studies and secure a conducive place to study from.
NEED HELP
In cases where you need help, you should contact the Director, Institute of
Distance Education, University of Zambia, P.O Box 32379, Lusaka. Cell 0978 772259 or 0979
772248, Email: [email protected]
ASSESSMENTS
The assessments under this module are in two categories. These are Continuous
Assessment (CA) and the Final Examination as shown below.
ASSESSMENT
Continuous Assessment: 50%
Broken down as follows:
a) Test 1 10%
b) Test 2 10%
c) Proposal 15%
d) Research Report 15%
Final Examination: 50
2
UNIT ONE: INTRODUCTION TO STATISTICS
1.0 Introduction
This unit gives an introduction to statistics. The unit begins by defining statistic and
thereafter outlines the importance of statistics. We will then examine different types of frequency
distributions such as ungrouped frequency distributions, grouped frequency distribution, absolute
frequency distributions, relative frequency distributions, cumulative frequency distributions and
decumulative frequency distributions. The last part of this unit will focus graphical presentation
of data.
2.0 Aim of the Unit
The aim of this unit is to introduce you to statistics.
3.0 Objectives of the Unit
By the end of this unit, you should be able to:
Explain the meaning of statistics.

Outline the importance of statistics.
Distinguish between descriptive and inferential statistics.
Identify different types of frequency distributions.
Construct graphs used for data presentations.
4.0 Required Material
In this unit, you will require the following readings:
Bless C. and Kathuria R. (1993) Fundamentals of Social Statistics: An African

Perspective, Cape Town: Juta and Co. Ltd
Kenny D. (1987) Statistics for the Social and Behavioural Sciences, Canada: Little
Brown and Co. Ltd
3
5.0 Time Required
It will take you approximately two hours thirty minutes to work through this unit.
6.0 Main Contents

6.1 Defining Statistics
Statistics is the plural of the word statistic. A statistic is a figure or a number. Statistics
therefore refers to numbers, figures or numeric information. For example, the number of students
studying PAS 2014 at the University of Zambia, the proportion of people who are unemployed in
Zambia, the population of Zambia, the number of voters in a constituency, the number of people
without access to clean drinking water and others. As a field of study, statistics is the study of the
collection, analyzing and interpretation of data (numbers/information). In research, we are
concerned with the study of methods of statistical analysis which involves descriptive and
inferential statistics. Descriptive statistics is concerned with organizing, summarizing and
describing of data. On the other hand, inferential statistics involves making generalizations for
the entire population based on results obtained from a sample.
6.2 Importance of Statistics
Statistics provides tools that you need in order to react intelligently to information you hear
or read. In this sense, statistics is one of the most important things that you can study. Statistics
help one to understand and evaluate quantitative information with minimum difficulties. For
example, the following statements are statistical in nature and require knowledge of statistics to
fully comprehend:
 63% of people in Africa live in rural areas;

 The poverty level in Zambia is estimated at 68%;
 Almost 80% of youths in Zambia are unemployed;
 The average amount of time PAS 2014 students spend on studying per day is 30 minutes;
 The HIV/AIDS prevalence rate in Zambia is 14.3%;
4
 The voter turnout in the last Presidential and Parliamentary elections in Zambia was
about 61%;
 51% of the national budget in Zambia is spent on salaries for civil servants;
 Approximately 10% of Zambian Members of Parliament in Zambia are women and;
 The literacy level in rural areas is almost 52%.
Knowledge of statistics is important in solving practical problem as it enables one to know

the severity of a problem. For instance, statistics on unemployment, prevalence of a disease,
crime rate and others demonstrate the need for an intervention or practical solutions in solving
the problem. This implies that statistics play a crucial role in policy formulation. Statistics is also
important in development planning and decision making because statistics reinforce or lend
credibility to an argument. Statistics is also important in business, particularly when it comes to
market research. Statistics is also important when undertaking opinion polls or opinion research.
6.3 Descriptive Statistics
Descriptive statistics is a way of organizing, describing and summarizing data in way that
is easy to understand. The organization of data takes several forms such as graphical techniques
and numeric techniques. The simplest way of organizing data consists of listing observations in
form of frequencies. In terms of the graphical techniques, data is presented graphically by way
of, for instance, pie – charts, bar charts, histograms, scatter plots and other. In numerical
presentation of data, we are interested in identifying some general characteristics by way of
measures of central tendency such as the mean, mode and median. We are also interested in
presenting data in terms of measures of dispersion such as the range, variance and standard
deviation.
Descriptive statistics are very important because if we simply presented our raw data it would
be hard to visualize what the data is showing, especially if there is a lot of it. Descriptive
statistics therefore enables us to present the data in a more meaningful way, which allows
simpler interpretation of the data. For example, if we had test results for PAS 2014, we may be
interested in the overall performance of these students. We would also be interested in the
distribution or spread of the marks. Descriptive statistics allow us to do this.
5
6.4 Frequency Distributions and Tables
Frequency distributions help us to organize data by showing us the frequency (regularity)

with which a particular variable or observation occurs. For example, suppose a sample of 20 civil
servants were asked to indicate their level of job satisfaction on a 5-point scale ranging from (i)
Very low; (ii) Low; (iii) Medium; High; and Very High.
Table 1.1: Level of satisfaction as indicated by 20 civil servants
Interviewee No. Level of job satisfaction

1 Low
2 Very high
3 Medium
4 Medium
5 High
6 Medium
7 Low
8 Very low
9 Medium
10 Low
11 Medium
12 High
13 Low
14 High
15 Very low
16 High
17 Medium
18 Medium
19 Low
20 Medium
6
The above responses obtained do not give us a clear view as to whether civil servants are
satisfied with their job or not. Therefore, the answers should be regrouped according to their
content whereby the answer to each category is given and this leads to a more adequate
description of the situation.
Table 1.2: Frequency distribution of the level of job satisfaction by 20 civil servants
Level of job satisfaction Frequency Interviewee No.

very high 1 2
High 4 5;12;14;16
Medium 8 3;4;6;9;11;17;18;20
Low 5 1;7;10;13;19
very low 2 8;15
n = 20
Table 1.2 above shows that civil servants who are neither satisfied nor dissatisfied with their
jobs constitute the largest group (8) whereas there are more dissatisfied civil servants altogether
(7) than those satisfied with their jobs. The total number of civil servants interviewed is equal to
the sum of all the frequencies and is indicated by n = 20, n represents the sample size, that is, the
total number of civil servants interviewed in this case.
1. Ungrouped Frequency Distribution
This is a type of frequency distribution which shows the number of times each observation
appears separately on its own. For example, suppose you collected data on the age of distance
students doing PAS 2014 at the University of Zambia. You sampled 25 (n = 25) students who
gave the following responses about their age.
Table 1.3: Responses to the question on age by 25 distance students
48 36 44 32 35 43 27 40 31 62 50 36
51 41 55 56 46 47 41 28 33 36 22 53 49
7
The raw data on the age of 25 distance students presented on page 7 would be more
confusing if hundreds of students were interviewed. The list would not make sense to most
people. To make table 1.3 more meaningful, one has to arrange the years in their ascending
order, starting from the lowest, which is 22 years to the highest which is 62 years. One can then
find out the number of times each score (age) appears, the frequency of the score is represented
by the letter (f).
Table 1.4: Ungrouped Frequency Distribution Constructed from the raw data in Table 1.3
Age f Age f Age f
22 1 43 1 58 1
26 1 44 1 62 1
27 1 46 1
31 1 47 1
32 1 48 1
33 1 49 1
35 1 50 1
36 3 51 1
40 1 53 1
41 2 55 1
Although the ungrouped frequency distribution in table 1.4 gives a better description of the
data than mere presentation of the raw data as in table 1.3, it is still a long and cumbersome list.
Therefore, there is need to make it more meaningful and manageable by grouping raw data
before counting their frequencies.
8
2. Grouped Frequency Distributions
The purpose of grouping data is to reduce the number of figures or scores presented in a
distribution so as to enable the reader grasp the main features of the data and to present the
information more comprehensively. However, the grouping must be done in accordance with
some rules, without the distortion or loss of too many items of information contained in the data
set.
Table 1.5: Grouped Frequency Distribution Constructed from the Ungrouped Frequency
Distribution in table 1.4.
Age (Class – interval) Frequency
20 – 24 1
25 – 29 2
30 – 34 3
35 – 39 4
40 – 44 5
45 – 49 4
50 – 54 3
55 – 59 2
60 – 64 1
n = 25
This description of the age of 25 distance students in PAS 2014 shows clearly the pattern of
the distribution. Most of the students are concentrated between 30 and 54 years whereas few
students are aged below 30 and above 54 years.
3. Absolute Frequency Distributions
Absolutes frequency distributions are frequency distributions where there are only absolute
values. Absolute frequency distributions can be applied to both grouped and ungrouped data. For
example, the frequency table below is an example of an absolute frequency distribution because
it does not show percentages which were observed at each score value.
9
Table 1.6: Absolute Frequency Distribution
Level of job satisfaction Frequency

very high 1
High 4
Medium 8
Low 5
very low 2
n = 20
4. Absolute Frequency Distributions
Relative frequency distributions are frequency distributions representing the percentage

of cases which were observed at each score value or class – interval. Thus relative frequency
distributions can be applied to both grouped and ungrouped data. The procedure is simply to
express each frequency as its corresponding percentage. For example, table 1.6 when expressed
as a relative frequency distribution will be like table 1.7.
Table 1.7: Relative Frequency Distribution
Level of job satisfaction F F%

very high 1 5
High 4 20
Medium 8 40
Low 5 25
very low 2 10
n = 20 100%
5. Cumulative Frequency Distribution
This is a kind of frequency distribution used when you want to know the percentage of
people/elements falling below a certain point. Cumulative frequency distribution is abbreviated
10
as (CF). For example, the cumulative frequency distribution from the data given in table 1.5
would be shown as follows:
Table 1.6 Cumulative Frequency Distribution
Age (Class – interval) F CF CF %
20 – 24 1 1 4
25 – 29 2 3[1+2] 12
30 – 34 3 6[3+3] 24
35 – 39 4 10[6+4] 40
40 – 44 5 15[10+5] 60
45 – 49 4 19[15+4] 76
50 – 54 3 22[19+3] 88
55 – 59 2 24[22+2] 96
60 – 64 1 25[24+1] 100
n = 25
In interpreting the cumulative frequency, you are supposed to use the upper limit of the class
interval. The upper limit is the number which shows where the class interval ends. For example,
in a class interval of 45 – 49 years, the upper limit is 49 years. On the other hand, the lower limit
is the number which shows where the class interval begins. In the class interval of 45 - 49 years,
45 is the lower limit. Therefore, the cumulative frequency above shows that 76% of distance
students are below 49 years. In absolute terms, 19 distance students are below 49 years.
6. Decumulative Frequency Distribution
This is a kind of frequency distribution which is used when you want to know the percentage
of observations located above a certain point. Decumulative Frequency Distribution is
abbreviated as (DCF). For example, the decumulative frequency distribution for the data given in
table 1.5 would look as follows:
11
Table 1.7: Decumulative Frequency Distribution
Age (Class – interval) F DCF DCF %
20 – 24 1 25[24+1] 100
25 – 29 2 24[22+2] 96
30 – 34 3 22[19+3] 88
35 – 39 4 19[15+4] 76
40 – 44 5 15[10+5] 60
45 – 49 4 10[6+4] 40
50 – 54 3 6[3+3] 24
55 – 59 2 3[1+2] 12
60 – 64 1 1 4
n = 25
In interpreting the decumulative frequency distribution, you use the lower limit of a class
interval. Therefore, the cumulative frequency distribution shows that 96% of distance students
are above 25 years. In absolute terms, 24 distance students are above 25 years.
Note that to find the mid-point of a class interval, you add the lower and the upper limit and
divide by 2. For example the mid-point the class interval 20 – 24 years is arrived at as follows:
(20 +24) / 2 = 22
6.5 Graphical Presentation of Data
Once raw data has been organized into a frequency distribution, it can be visually presented
by various types of graphs, bars charts, pie charts and other forms of pictorial presentations.
Graphical representations have a great advantage of allowing one to grasp immediately the main
characteristic of the information. Although a frequency distribution describes a set of data, it still
remains a list of figures which has to be studied carefully to get information, for example the
variations of the data. However, there are ways of depicting a frequency distribution in order to
allow an immediate grasp of its characteristics, this is based on visual distributions where
12
numerical relationships are expressed in spatial relationships. The following are some of the
ways in which data can be presented:
1. Pie Charts
Pie charts are diagrams that use a circle which is subdivided into sections. Each section is
proportionate to the size of the figure represented. It is a very simple method of expressing
various components and their relationship to the whole. Each figure must be transformed into an
angle in degrees, the total corresponding to 360º.
For example, suppose are the analyzing the number of street kids in 4 major towns in Zambia
and you have the following;
City No. of Street Kids

Lusaka 450
Kitwe 300
Ndola 150
Livingstone 100
Total 1000
Since 1000 represents 360º,
The number of street kids in Lusaka is equivalent to 450/1000 x 360 = 162º
The number of street kids in Kitwe is equivalent to 300/1000 x 360º = 108º
The number of street kids in Ndola is equivalent to 150/1000 x 360 = 54º
The number of street kinds in Livingstone is equivalent to 100/1000 x 360 = 36º
13
Figure 1.1: Number of Street Kids in Four Selected Towns in Zambia
Livingstone,
100
Ndola, 150
Lusaka, 450
Kitwe, 300
2. Bar charts
Simple bar charts represent data by a series of bars where the height of each bar indicates the
size of the figure or the value represented. The width is the same for all the bars. To construct a
bar chart, label the frequencies vertically, locate the categories of a variable on a horizontal scale
and label the frequencies on the vertical axis by using whatever scale you consider appropriate.
For example, suppose 400 students at the University of Zambia were asked to rate the quality
of internet connectivity at the institution and gave the following responses.
Response Number of students

Very Bad 20
Bad 35
Average 210
Good 90
Very Good 45
14
Figure 1.2: Students perception of the quality of internet connectivity at UNZA
250
200 210
150
100
90
50
45
20 35
0
Very Bad Bad Average Good Very Good
3. Multiple Bar Charts
A multiple bar chart allows the representation of many variables simultaneously. This is used
when more than one relationship has to be depicted. For example, one might be interested in
studying the students’ population as it is spread over the various schools over a given period of
time. This offers the possibility for comparing the students’ population in different schools over
a given period of time. For example, hypothetical data on the enrollment rate in three schools at
the University of Zambia from 2010 to 2013 is presented in the table below.
Academic Year
School 2010 2011 2012 2013
Education 700 790 1000 1300
Humanities 600 880 900 1200
Natural Sciences 500 800 850 1000
15
Figure 1.3: Students Enrollment in three Schools at UNZA from 2010 - 2013
1400
1200
Number of Students
1000
School of Educaion
800
600 School of Humanities and

Social Sciences
400 School of Natural
Sciences
200
0
2010 2011 2012 2013
Academic year
4. Graphs
A graph is a representation of data by a continuous line within the framework of two axes,
the horizontal axis and the vertical axis. Social scientists tend to use a specific kind of graph
where (y) denotes the frequency of the variable or variables and (x) denotes the periods under
consideration. If the graph is about one variables or item, then there will be one continuous line
in the graph. However, if two variables or items are considered, then there will be two
continuous lines. For example, data on Maize and Cotton production in Gwembe district from
2000 to 2004 is shown below.
Farming Season
Crop 2000 2001 2002 2003 2004 2005
Maize 500 450 750 575 700 900
Cotton 430 350 760 630 650 700
16
Figure 1.4: Maize and Cotton Production in Gwembe, 2000 – 2004
Maize and Cotton Production

1000
900
800
700
Metric Tonnes
600
500 Maize
400 Cotton
300
200
100
0
2000 2001 2002 2003 2004 2004
7.0 Summary
You have been informed in that statistics refers to numbers, figures or numeric
information. As a field of study, statistics is concerned with the analysis of data and this involves
descriptive statistics and inferential statistics. Descriptive statistics is concerned with the
organising, summarizing and describing of data while inferential focuses on making
generalizations on results obtained from a sample for the whole population. Various techniques
are used under descriptive statistics such as frequency distributions and graphs.
8.0 Self Assessment Questions
After reading through unit one, you should answer the questions below. It is
important for you to attempt the questions in order to gauge how much you have comprehended
the material presented in this unit. If you are having challenges in answering the question, read
unit one again.
17
Given data below:
Class – interval Frequency
12 – 18 2
19 – 25 1
26 – 32 5
33 – 39 1
40 – 46 4
47 – 53 4
54 – 60 2
61 – 67 2
68 – 74 3
75 – 81 2
82 – 88 2
89 – 95 2
Complete the frequency table by including the following;

a. The midpoint of each class interval
b. The relative frequency of each class interval
c. The cumulative frequency for the above data
d. The de-cumulative frequency for the above data
Distinguish between the following;

a. Ungrouped frequency distribution and grouped frequency distribution.
b. Cumulative frequency distribution and decumulative frequency distribution.
c. Pie chart and bar chart.
d. Inferential statistics and descriptive statistics.
18
UNIT TWO: MEASURES OF CENTRAL TENDENCY
1.0 Introduction
This unit discusses the measures of central tendency. The unit begins with explaining the
mode and thereafter discusses the median. We will then focus on the mean for both ungrouped
and grouped data.
2.0 Aim of the Unit
The aim of this unit is to introduce you to the measures of central tendency.
Interpret the measures of central tendency

Compute the measures of central tendency

Brown and Co. Ltd
5.0 Time Required
It will take you approximately forty minutes to work through this unit.
19
6.0 Main Contents
Measures of central tendency are ways in which a set of figures can be summarized by using
only one number which expresses the central feature of the whole set. These measures are used
in situations where graphical methods may be inappropriate such as when making statistical
inference. These methods are also used for reasons of expediency because it is sometimes easier
to verbally communicate than to use graphs. The measures of central tendency used in research
are the mode, median and the mean.
6.1 The Mode
The mode of a set of measures is defined as the measurement or observation that occurs most
frequently in that distribution or has the highest frequency in that distribution. For example,
n=25 (n stands for sample size or number of observations), number of hours PAS 2014 students
spend studying per week.
7, 10,8, 11,9, 9, 9, 8, 9, 8, 9, 9, 9, 8, 9, 8, 8, 9, 10, 11, 10, 7, 10, 9, 8
Arrange the data from the smallest to the highest according to the number of times
(frequency) they occur.
7, 7, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 11, 11
The number whose observation occurs most frequently is 9. Therefore, the mode is 9,
which means that most PAS 2014 students spend 9 hours per week studying. Sometimes it is
possible to have more than one mode, as is the case in bimodal or trimodal distributions.
The mode is an important measure of central tendency insofar as it enables one to easily
determine what is typical or common among a group. However, the shortcomings of the mode
are that it does not take into account all the observations or values in its computation.
20
6.2 The Median
The median of a set of measurements or observations is the middle value when these
measurements or observations are arranged in order of magnitude. Where there is an even
number of observations, the median is simply the average of the two midpoint values. In
situations where we have an old number of observations, the midpoint value becomes the
median. Suppose you had the following number of observations where n=10;
95, 86, 78, 90, 62, 73, 89, 92, 84, 76
Rearrange the numbers in order of magnitude
62, 73, 76, 78, 84, 86, 89, 90, 92, 95
Median = (84 + 86) /2 = 170/2 = 85
If we include 60 to make n=11, we will have;
60, 62, 73, 76, 78, 84, 86, 89, 90, 92, 95
Median = 84
6.3 The Mean
The mean is often referred to as the average. It is the sum of measurements divided by the
total number of measurements. The population mean is denoted by µ while the sample mean is
denoted by ̅.
a. The Mean for Ungrouped Data
For example in class of 10 students, the marks obtained by the students in a test were as
follows;
50, 59, 65, 63, 85, 76, 39, 41, 51, 51
21
Mean ( ̅ ) =
= 58
The interpretation of the mean mark we have computed is that on average, each student is
expected to have scored 58 marks in the test.
b. The Mean for Grouped Data
When calculating the mean for grouped data, you need to have the midpoints of the class
intervals. The midpoint of a class interval is denoted by xi and is computed by adding the upper
and lower limit of the class interval and dividing the answer by 2. For example, for the class
interval 20 – 24, the midpoint will be (20 + 24) / 2 = 22. The formula for grouped data mean is:
̅=∑
Where ̅ = mean
Σ = is a summation symbol
= is the mid - point of a given class interval
= is the frequency for a given class interval
= is the product of the frequency of a class interval and its midpoint
Suppose you have the following grouped data.
22
Table 2.1: Grouped Data on Age
Age group
20 – 24 22 3 66
25 – 29 27 4 108
30 – 34 32 5 160
35 – 39 37 6 222
40 – 44 42 5 210
45 – 49 47 4 188
50 – 54 52 3 156
n = 30 Σ =1110
Therefore, the mean age for the grouped data above is:
̅=∑
= 37 years
7.0 Summary
You have been informed in this unit that measures of central tendency are ways in
which a set of figures can be summarized by using only one number which expresses the central
feature of the whole set. These methods are also used for reasons of expediency because it is
sometimes easier to verbally communicate than to use graphs. The measures of central tendency
used in research are the mode, median and the mean.
23
After reading through unit one, you should answer the questions below. It is
unit two again.
The following are the marks scored by students in a test:

40 28 38 78 82 56 74 20 62 46
80 90 52 30 58 82 72 24 48 70
Using the data above compute the following and give an interpretation of each: (65 Marks)
a. Mean
b. Median
c. Mode
The data below shows the number of 50 kg bags of maize harvested by 1, 000 small scale
farmers. The number of bags is grouped in intervals of 5.
Table 2.1: Grouped Data on Bags of Maize Harvested
Class-interval Frequency
10 – 14 8
15 – 19 30
20 – 24 40
25 – 29 94
30 – 34 126
35 – 39 156
40 – 44 176
45 – 49 138
50 – 54 118
55 – 59 70
60 – 64 20
65 – 69 16
70 – 74 8
a. Compute the mean and interpret the answer.
24
UNIT THREE: MEASURES OF DISPERSION
1.0 Introduction
This unit discusses measures of dispersion. The unit begins by explaining the ranges as a
measure of dispersion. Thereafter, we shall look at the average deviation from the mean as a
measure of central tendency. We will then examine the variance as a measure of central tendency
and the last part in this unit will discuss standard deviation.
2.0 Aim of the Unit
The aim of this unit is to explain measures of dispersion.
Identify the measures of dispersions.

Compute the measures of standard deviation.

Brown and Co. Ltd
Nachmias, C. and Nachmias, D. (1996) Research Methods in the Social Sciences. 5th ed,
London: St. Martins Press
5.0 Time Required
It will take you approximately one hour to work through this unit.
25
6.0 Main Contents
When dealing with quantitative data, the mode, median and the mean can be computed
but variability is equally important. Variability is concerned knowing the extent to which values
in the distribution deviate from the central value. Measures of variability are also called
measures of dispersion; they constitute another type of statistic which summarizes a distribution.
Just like measures of central tendency, there are different types of measures of dispersion.
6.1 The range
The range is the difference between the highest and the lowest score in a distribution. Its
computation however, depends on whether the data is grouped or ungrouped. The range for
ungrouped data is given by the formula; Range = highest score –lowest score. For example, if a
distribution has the following scores, 4, 8, 9, 7, 5, 3. Then range is 9-3 = 6
For grouped data in class intervals, the range is defined as the difference between the real upper
limit of the highest class interval and the real lower limit of the lowest class interval. Therefore,
Range = real upper limit of the highest class interval – the real lower limit of the lowest class
interval.
Table 3.1: Grouped data on Age
Age (Class – interval) F CF DCF
20 – 24 1 1 25
25 – 29 2 3 24
30 – 34 3 6 22
35 – 39 4 10 19
40 – 44 5 15 15
45 – 49 4 19 10
50 – 54 3 22 6
55 – 59 2 24 3
60 – 64 1 25 1
n = 25
26
Referring to the data above, the real upper limit of the highest class interval, 60 – 64 years is
64.5 and the real lower limit of the lowest class interval 20 – 24 years is 19.5 years. Hence the
range of this distribution is 64.5 – 19.5 = 45. The frequency of the class interval does not affect
the range. Note that the upper and lower limits of the class interval you have in the table above
are stated limits. To find a true upper limit, you add 0.5 to the stated upper limit. Similarly, to
find the true lower limit of a class interval, you subtract 0.5 to the stated lower limit
6.2 Average deviation from the mean
The range is not very suitable for measuring the extent to which values differ from the mean.
We need to have a measure which takes into account the variations from the mean. Deviation
from the mean is a more robust measure as it gives the relationship between each value and the
dispersion from the mean. For example, if 5 students got the following results in a test; n = 5
Student Marks (%) / x
A 68
B 67
C 66
D 63
E 61
Σx = 325
̅=
= 65%
27
The margins by which these students deviated from the mean can be as shown in the table below.
Student Marks (x) d = (x – ̅) |x - ̅|
A 68 +3 3
B 67 +2 2
C 66 +1 1
D 63 -2 2
E 61 -4 4
Σ | x – x | = 12
The third column indicates the difference between each score and the mean. Since some
of the values have a negative sign, it would make it impossible for us to find the average because
when we add the deviations, the sum will be zero. To overcome this, the fourth column has been
included with a heading of | x – ̅ | these signs transform all the values obtained in the third
column into positive value. Therefore, the general formula for average deviation is
∑ –̅
AD =
= 2.4
28
The formula for average deviation for grouped data is:
∑ –̅
AD =
Age group Xi F - ̅ f| – ̅|
20 – 24 22 3 -15 45
25 – 29 27 4 -10 40
30 – 34 32 5 -5 25
35 – 39 37 6 0 0
40 – 44 42 5 5 25
45 – 49 47 4 10 40
50 – 54 52 3 15 45
n = 30 Σ f | - ̅ | = 220
∑ –̅
AD =
= 7.3
6.3 Variance
While discussing the average deviation, it became clear that a deviation expresses the
distance of a particular score from the mean, symbolically written as d = x – ̅. However, the
average of all the deviations from the mean cannot be taken as a measure of variability. This is
so because as we discussed with average deviations, the mean deviation is sometimes zero. Since
Σ (x – ̅) = 0. This is avoided by using the square of the deviations (instead of the absolute values
as done for the average deviations). The average of the squared deviations of a distribution from
the mean of a distribution is called the variance and its formula is:
–̅
σ² =
29
This formula is only considered appropriate only for large population distributions. In
this case, the Greek letter sigma (σ) is used. However, when dealing with samples, a more
precise formula of the variance is used.
–̅
S² =
In this formula, the sum of the squared deviations from the mean is divided by n – 1 instead
of n and the S serves as a symbol for variance. The variance S² is 0 only if all the scores in the
distribution do not differ from the mean. Otherwise, since the variance depends on the distance
of each score from the mean, the more the scores differ, the larger the spread of the variance and
the more heterogeneous the group of data. The variance is expressed in squared units of
measurements. If scores refer to people or years, the variance will be in square people or years
which might not make sense.
6.4 Standard Deviation
Standard deviation is an improvement of the variance. In simple terms, the standard deviation
of a distribution is the square root of the variance. As in the case of the variance, the Greek letter
σ indicates that the formula refers to the population whereas the letter S refers to a sample. There
are two ways of calculating standard deviation valid for grouped and ungrouped data. The first
one is based on the actual definition of standard deviation as a square root of the variance.
( –̅ )
σ= √
The formula above is called the definitional formula for standard deviation and is used
for large populations. When using samples, n as the denominator is replaced with n – 1.
–̅
S= √
30
1. Standard Deviation of Ungrouped Data
The standard deviation for ungrouped data on the marks obtained by 5 students in a test
presented on page 27 can be computed as follows;
Student Marks (x) Deviations Square of Deviations
(x – ̅) (x – ̅)²
A 68 +3 9
B 67 +2 4
C 66 +1 1
D 63 -2 4
E 61 -4 16
Σ (x – ̅ )² = 34
The standard deviation formula for ungrouped data is;
–̅
S= √
This formula requires you to first calculate the mean, find the deviation from the mean
for each, square the deviations and find the sum of the squared deviations as shown in the table
above. Therefore;
–̅
S= √
= √
31
=√
= √
= 2.9
This means each student is expected to have deviated from the mean by 2.9 marks.
2. Standard Deviation for Grouped Data
Age group xi F x– ̅ (x – ̅ )² f(x – ̅ )²
20 – 24 22 3 -15 225 675

25 – 29 27 4 -10 100 400
30 – 34 32 5 -5 25 125
35 – 39 37 6 0 0 0
40 – 44 42 5 5 25 125
45 – 49 47 4 10 100 400
50 – 54 52 3 15 22 675
n = 30 Σ f(x – ̅ )² =2400
–
S= √
= √
= √
= √
= 9.1
32
This means that on average each individual is expected to have deviated from the mean by
9.1 marks.
7.0 Summary
You have been informed in this unit that measures of dispersion help us to know the
extent to which values in the distribution deviate from the central value. At this point you should
know that the most commonly used measures of dispersion are the range, average deviation from
the mean, variance and standard deviation. Standard deviation is the most refined measure of
dispersion and there are two ways of calculating standard deviation valid for grouped and
ungrouped data.
The following frequency table of rent cost for a bed space was constructed after collecting
data on rentals from boarding houses near the University of Zambia.
Table3.1: Rental Charges for Boarding Houses
Rent ( in Kwacha) Frequency
226 – 250 2
251 – 275 5
276 – 300 9
301 – 325 7
326 – 350 3
351 – 375 2
376 – 400 4
401 – 425 1
426 – 450 3
451 – 475 2
(i) Compute the mean and interpret the answer.

(ii) Compute the Standard deviation and interpret the answer.
33
UNIT FOUR: INTRODUCTION TO PROBABILITY
1.0 Introduction
This unit gives an introduction to probability. The unit begins by discussing the common
uses of probability. We will then examine properties of probability by focusing on additional,
conditional and multiplication rules of probability. Thereafter, we will demonstrate the
application of these properties of probability through the use of an example, which will be the
last part of this unit.
2.0 Aim of the Unit
The aim of this unit is to introduce you to probability.
Explain the uses of probability.

Identify the properties of probability.

Brown and Co. Ltd
5.0 Time Required
It will take you approximately one hour thirty minutes to work through this unit.
34
6.0 Main Contents
Whereas descriptive statistics are simply procedures for condensing information about a
set of measurements or observations, inferential statistics refer to techniques for making
generalizations on the basis of partial information. The study of probability introduces inferential
statistics. When testing hypotheses or making generalizations, we use probability to determine
the likelihood of an event occurring. For example, the level of significances, which is the
probability of rejecting a null hypothesis when it is in fact true, can be easily understood if one
has knowledge of probability.
4.1 Common uses of Probability
Probability statements are made by everyone when attempting to explain the present or
predict the future. For example, UNZA students are demonstrating, there is a probability that
they have not been paid their meal allowance. The young doctor is very dedicated, she will go
far in life. Probability is used when evaluating the likelihood of an event happening such as
students demonstrations at UNZA. Since there have been several instances when students have
demonstrated due to late payment of meal allowance, there is a likelihood that when students
demonstrate, delayed payment of meal allowances could be the issue. Probability reasoning is
most used to predict the future, the young doctor satisfies pre-conditions for success in her job,
so her chances for a bright future are good.
In everyday life, probability evaluations are expressed in words such as chance, likelihood,
possibility, expectations or prospects. However, the probability assessments in day to day
situations are usually inaccurate and often subjective or intuitive. Some common sense laws of
probability are recognized such as that of the probability of an event occurring increases as more
preconditions are satisfied. Some events have no probability of occurring at all, for example, the
probability of a newly born baby committing suicide. On the other hand some events are certain
to occur such as the a normal healthy baby learning how to walk
The basics of probability presented here aim at identifying and making more precise probability
assessment based on a few probability laws and then applying them to specific fields of statistics.
35
4.3 Some Properties of Probability
In a questionnaire, a respondent has the probability of ticking one of the answers given to
each of the multiple choice questions. Assuming that the respondent answers all the questions,
the probability that any respondent ticks the answer (Yes) in a Yes or No question is . This can
be expressed in the following ways.
If E denotes the event under discussion (E = to tick the answer yes) the probability of this
event occurring is symbolized by p (E). Therefore p (E) = = 0.5. If the response category has 5
alternatives (a 5 point scale) For instance, ‘Strongly agree; Agree; No opinion; Disagree and
strongly disagree’ and one wants to calculate the possibility of the answer ‘No opinion’ being
given. One probability out of 5 possibilities and it is p (E) = = 0.2. In the same way, the
possibility of any answer other than ‘No opinion’ is 4 out of 5 possibilities and it is p (E) = =
0.8 and the possibility of a positive answer (Strongly agree or agree) is p (E) = = 0.4. In other
words, the possibility of an event occurring is:
1. p (E ) =
If an event is impossible and can never occur, it means that the number of favourable outcomes
is zero, the p (E) = 0.
If an event is certain that it will occur, it means the number of favourable outcomes is equal to
the number of possible outcomes, then p (E) = 1
2. 0 ≤ p (E) ≤ 1
The probability of any event occurring varies from 0 to 1. Therefore, it is evident that to
each event E there exists there exists its complementary, non – E. If one does not pick the answer
E = ‘No opinion’ one has to pick one of the remaining answers or possibilities which altogether
constitute the complementary event of E. In simpler terms, a new born baby is either a girl (E) or
a boy (non - E). Therefore, the possibility of a newly born being either a girls or a boy becomes a
36
certainty, since there is no alternative. The probability that an event will occur and its
complementary adds up to certainty.
3. p (E) + p ( non – E) = 1
The events E and non – E are called mutually exclusive events since they cannot occur
simultaneously. Only complementary events are mutually exclusive. For example, you cannot
toss a coin and have a head and tail up at the same time.
Addition Rule of Probability
If two events and are mutually exclusive, the probability that one or the other will
occur is simply the sum of their respective probabilities. This rule is expressed by;
4. p( U )= p( )+p( )
The symbol U is borrowed from set theory and is used to denote either of the two events occurs.
If two events and are not mutually exclusive, that is they can happen at the same
time, then the probability of either of them happening is the sum of the probability that each
occurs without the other plus the probability that each occurs with the other. Using the symbol ∩
borrowed from set theory to denote that two events occur simultaneously; this probability can be
written as p ( ∩ ). Therefore, the addition rule of probability for two events which are not
mutually exclusive is;
5. p ( U )= p( )+p( )-p( ∩ )
When two events are mutually exclusive, ∩ cannot exist, in set theory it is an empty set
and p ( ∩ ) = 0.
Conditional Probability
37
A frequent situation in reality is that the likelihood of an event occurring depends on
whether another event has already occurred. For example, if a child is born into a wealthy
family, the chance of that child enjoying a good education is higher than if he or she comes from
a poor economic background. The probability of an event occurring given that some other
event has already taken place is called conditional probability and is denoted by p ( ).
Based on the foregoing, we can distinguish between two types of events, dependent
events and independent events. If the probability of one event occurring is dependent on another
taking place, these two are called dependent events. Such events have been illustrated above.
On the other hand, if the probability of one event occurring does not affect the probability
of the other occurring, and vice versa, these events are called independent events. In this case,
the probability p ( is not affected by having taken place and it is equal to the conditional
probability. Symbolically;
6. p( )=p(
The above events are independent but not mutually exclusive. They can occur simultaneously.
Multiplication Rule of Probability
Given two events, mutually exclusive or not, the probability of at least one of them taking
place is p ( U ). What is the probability of both of them occurring? In other words, how
is p ( ∩ ) found? Here a distinction has to be made between independent events and
dependent events.
The probability of two independent events and both occurring is simply a product of their
respective probabilities.
7. p ( ∩ )=p( )xp( )
However, if the events are dependent, one has to take into account that once one event
has taken place, the probability of the second event taking place changes accordingly. Thus one
has to use the conditional probability instead of the probability of the event alone.
38
The probability that two dependent events and both occur is the product of the
probability of one by the conditional probability of another.
8. p ( ∩ )=p( )xp( ) Or
p( ∩ )=p( )xp( )
Clearly, if events are independent, using formula 6 in the above expression, formula 7 is
obtained. Moreover, formula 8 allows for the conditional probability to be expressed by;
9. p( )= or
p( )=
Example
Let us illustrate the various properties of probability by an example related to probability

sampling. A certain population of 150 women is divided into categories as shown below.
Table 4.1: Marital Status of Women and Household Status
Non Heads of Households Heads of Households Total
Housewives 45 0 45
Divorcees 5 10 15
Widows 5 25 30
Unmarried 35 25 60
Total 90 60 150
a. What is the probability of each category to be drawn from the population?

i. The probability of selecting a Housewife is p ( )= = 0.3
39
ii. The probability of selecting a Divorcee is p ( )= = 0.1
iii. The probability of selecting a Widow is p ( )= = 0.2
iv. The probability of selecting an Unmarried woman is p ( )= = 0.4
b. What is the probability of drawing out of the population a woman who has been
previously married?
The probability of selecting out of the population a woman who has been previously
married is: p ( U )=p( )+p( )
= 0.1 + 0.2
= 0.3
c. What is the probability of selecting a woman who is not presently married?

The probability of selecting a woman who is not presently married is:
p( U =p( )+p( )+p( )
= 0.1 + 0.2 + 0.4
= 0.7
d. What is the probability of selecting a woman who is neither a housewife nor a head of
household?
The probability of choosing a woman who is neither a housewife nor a head of household
is: p (non - U =1–p U )
= 1 – 0.17
= 0.83
e. What is the probability of selecting a woman who is either unmarried or is head of
household?
The probability of selecting a woman who is either unmarried or is head of household is:
p U )=p +p – p (( ∩ )
= 0.4 + 0.4 – 0.17
= 0.63
40
f. What is the probability of selecting a woman who is either divorced or is head of
household.
The probability of selecting a woman who is either divorced or is head of household is;
p( U )=p )+p( )–p( ∩ )
= 0.1 + 0.4 – 0.07
= 0.43
g. What is the probability of selecting a selecting a head of household knowing that this
woman is divorced? This refers to the conditional probability p ( ). Since there are
15 divorcees forming and out of them 10 are heads of households, this conditional
probability is; p( )=
= 0.67
h. What is the probability of selecting a divorcee knowing that this woman belongs to the
group of heads of households? This refers to the conditional probability p ( ).
Since there are 60 heads of household forming and out of the 10 are heads of
households, the conditional probability is; p( )=
= 0.17
i. What is the joint probability of being a widow and being a head of household?
Here the multiplication rule of probability applies. Thus the probability is;
p( ∩ )=p( )xp( )
= 0.2 x 0.4
= 0.08
8.0 Summary
You have been informed in that probability is the likelihood of an event occurring.
The study of probability is important in inferential statistics as it helps in making generalisations.
The probability of any event ranges between 0 and 1. The unit has also explained that dependent
events are events whose occurrence is dependent on another event occurring. On the other hand,
independent events are those events whose occurrence does not dependent on another event
taking place.
41
The table below shows the preferred sports of 2500 students at Kalomo College of Health
Sciences:
Sport Sex
Male Female
Football 380 570
Golf 440 660
Cricket 180 270
Using the data above, compute the following;

a. The probability of selecting a student who like golf
b. The probability of selecting a student who either likes football or is male
c. The probability of selecting a female student knowing that she likes cricket
d. The probability of liking football and being male
42
UNIT FIVE: FREQUENCY DISTRIBUTIONS
1.0 Introduction
This unit discusses frequency distributions. The unit begins with explaining the types of
frequency distributions. Thereafter we shall discuss the standard normal distribution and its
significance in the field of research. We will then focus on Z – score and give examples to
illustrate the use of the standard normal curve.
2.0 Aim of the Unit
The aim of this unit is to discuss frequency distributions.
Identify different types of frequency distributions.

Distinguish between the standard normal distribution and an ordinary normal distribution.
Explain the characteristics of the standard normal curve.
Calculate Z - scores

Brown and Co. Ltd
43
5.0 Time Required
6.0 Main Contents
One of the first things examined in a frequency table are the peaks in the distribution. A
peak in a distribution is a frequency or sets of adjacent frequencies that are larger than most of
the other frequencies. For example, suppose you collected the following data on the height of
students in a class n = 30
Table 5.1: Height of students in a class of 30
Height (m) No. Of Students
1.2 4
1.5 6
1.6 10
1.7 6
1.8 4
When the above data is presented in form of a distribution, it will take the shape of a
curve as shown in figure 5.1 below. The distribution of the data helps in identifying its general
form.
44
Figure 5.1: Graphical distribution of students’ height
Frequency
Height
Distributions may have distinctive forms with few low scores and many high scores; with
many low scores and few high scores; with many score concentrated on the left or right side of
the distribution and; with many scores concentrated on the in the middle of the distribution and
few scores on either side of the distribution. The simplest way to describe a distribution is by
visual representation. Examples are presented below.
Figure 5.2: Symmetrical Distribution
0 ̅̅̅
Mean (𝒙
Mode
Median
45
Figure 5.3: Positively Skewed Distribution
Mode
Median
̅̅̅
Mean (𝒙
Figure 5.4: Negatively Skewed Distribution

Mode
Median
̅̅̅
Mean (𝒙
The values of the variables are presented along the baseline and the area under the curve
represents the frequencies. The distribution in figure 5.2 is symmetrical, that is the frequencies at
the right and left tails of the distribution are identical, so if the distribution is divided into two
halves, each will be the mirror image of the other. This means that most of the observations are
46
concentrated at the middle of the distribution and there are few observations with very high or
very few scores. Many variables tend to be distributed symmetrically and this form of
distribution plays an important role in the field of statistics.
When there are more extremely high scores, the distribution is positively skewed as
shown in figure 5.3 and when there are more extremely low scores, the distribution is negatively
skewed as shown in figure 5.4. Skewed distributions are also referred to as non symmetrical;
they have more extreme cases in one direction of the distribution than in the other.
Skewness can also be identified according to the measures of central tendency. In a

symmetrical distribution, the mean will coincide with the median and the mode. In skewed
distributions, there will be discrepancies between these measures. In a negatively skewed
distribution, the mean will be pulled in the direction of lower scores. In a positively skewed
distribution, the mean will be pulled closer to higher scores. The property of skewed distributions
makes the choice of a measure of central tendency a critical issue. Since the mean is pulled in the
direction of the extreme scores, it loses its usefulness as a representative measure. In such
instances, it might be useful to employ the median or the mode instead.
6.1 The Standard Normal Distribution
One type of symmetrical distribution called the standard normal distribution has great
significance in the field of statistics. The standard normal distribution has the following
properties.
 It is symmetrical and bell shaped. Each side of the curve constitutes 50% of the
distribution.
 It is unimodal, this means that the mean, the median and the mode coincide at the centre
of the distribution.
 It is asymptomatic, this means the area between the curve and baseline on both sides
extends to infinity.
 The area under the normal curve represents a probability of 1. Each side of the curve
represents a probability of 0.5.
47
 In a normal curve, a fixed proportion of observations lie between the mean and fixed
standard deviation units. The proportions can be seen from the figure below.
Figure 5.5: The Normal Distribution
The mean of the distribution divides it exactly in half, 0.3413 (34.13%) of observations
lie between the mean and 1 σ (one standard deviation) to the right of the mean and the same
proportion falls between the mean and one standard deviation to the left. The plus sign indicates
one standard deviation above the mean, the minus sign indicates standard deviation below the
mean. Thus 68.28 % of all observation fall between x , 95.46% of all observations fall
between x 2 and 99.73% of all the observations fall between x
In any univariate analysis that is normally distributed, the proportion of observations

included within fixed distances of the mean can be determined. For example, in a distribution of
intelligence tests with a mean score of 110 and a standard deviation of 10, 68.28% of subjects
will have an IQ of 110 ; that is an IQ of between 100 and 120 and 95.46% of subjects will
have an IQ of 110 ; that is an IQ of between 90 and 130.
48
6.2 Standard Normal Distribution and Standard Scores
On the basis of the mean and the standard deviation of the randomly distributed data it is
possible to construct a standard normal distribution. The standard normal distribution resembles
an ordinary normal curve. There is only one standard normal curve distribution unlike ordinary
normal curve which can be several. In order to go round these problems it becomes necessary to
standardize the ordinary normal curve. This standard normal curve has a mean of 0 and a
standard deviation of 1 whereas ordinary normal curves will have different means and standard
deviations. For example, examination results for PAM 2010 for the year 2012, 2013 and 2014
may conform to a normal distribution but these three normal distributions may have different
means and standard deviations. Therefore, these normal distributions have to be standardized so
that they have the same mean and standard deviation. Standardization uses a formula referred to
as standard score or standard normal deviation or z-score. The formula is:
̅
Z=
Where Z = Number of standard deviation units
X = Any observation or score of the distribution;
̅ = Mean of the distribution and;
S = Standard deviation of the distribution.
Z also referred to as a standard score, expresses the distance between a specific

observation (x) and the mean in terms of standard deviation units. Special tables have been
constructed for the standard normal curve. The tables enable you to determine the proportion of
observations that lie between the mean and any observation in the distribution. A table of Areas
Under the Normal Curve is attached at the end of the module. The first two digits of Z are
listed in the left hand column; the third digit is shown across the row. For example, a Z – score of
1 is .3413 (Z – scores in the table are expressed to two decimal places. In this case, the value of a
49
Z – score of 1 was found by first locating 1.0 in the first column and then moving across to
horizontally to .00). The table only shows half of the curve’s proportions because the curve is
symmetrical. Thus the distance between the mean and a Z of -1.0 is identical to the distance
between the mean and a Z of 1.0.
Researchers use the standard normal curve to evaluate the proportion of observations
included within a desired interval and to determine the probability of various events. To do this,
the raw scores must be converted to standard deviation units in order to use the tables which
report the area under the standard normal curve. When raw data has been converted into standard
scores, a single table can be used to evaluate distributions regardless of the scale on which the
data were measured. Therefore, distributions measured on different scales can be compared.
Observations are converted into standard deviational units using the formula stated on page 36.
To illustrate the use of the normal curve, suppose the distribution of income in a particular
community is normal, its mean is K15, 000 and the standard deviation is K 2,000.
1. Calculating proportions
a. What is the proportion of people in the community who have an income between
K11,000 and K 15,000?
Figure 5.6 Proportion of people earning between K11, 000 and K15, 000
.4773
̅
𝒙
11, 000 15,000
50
First we convert the figure into standard deviation units.
̅
Z=
–
=
= -2
= .4773
When you go to the areas under the normal curve, the value of a Z – score of 2 is .4773,
therefore, the proportion of people who have an income between K11, 000 and K15, 000
is .4773.
b. What is the proportion of people in the community who get between K16, 000 and K20,
000?
Figure 5.7: Proportion of people earning between K16, 000 and K 20,000
.3023
̅
𝒙 16, 000 20, 000
K15, 000
First we need to convert the two figures into standard deviation units.
̅ ̅
= =
– –
= =
= =
51
= 0.5 = 2.5
= .1915 = .4938
The Z – score value of 0.5 is .1915 and the Z- score value of 2.5 is .4938
Therefore, the proportion of people in the community who get between K16, 000 and K 20,000
is .4938 - .1915 = .3023
c. What is the proportion of people who get between K11, 000 and K 17,500?
Figure 5.8: Proportion of people earning between K 11,000 and K 17, 500
.8717
̅
𝒙
11, 000 15,000 17, 000
̅ ̅
= =
– –
= =
= =
= -2 = 1.25
= .4773 = .3944
52
Therefore, the proportion of people who get between 11,000 and 17,500 is .4773 + .3944 = .8717
2. Calculating percentages
a. What is the percentage of people whose income is between K11, 000 and K15, 000?
̅
Z=
–
=
= -2
= .4773
= .4773 X 100 (Note that to find percentages, you calculate the Z – score value by 100)
= 47.73%
This means that 47.73% of people in this community have an income of between
K11, 000 and K 15, 000.
b. What is the percentage of people in the community who get between K16, 000 and
K20, 000?
First we need to convert the two figures into standard deviation units.
̅ ̅
= =
– –
= =
= =
= 0.5 = 2.5
= .1915 = .4938
The Z – score value of 0.5 is .1915 and the Z- score value of 2.5 is .4938
The proportion of people in the community who get between K16, 000 and K 20,000 is .4938 -
.1915 = .3023, hence the percentage of people who get between K16, 000 and K20, 00 is .3023 x
100 = 30.23%
53
c. What is the proportion of people who get between K11, 000 and K 17,500?
̅ ̅
= =
– –
= =
= =
= -2 = 1.25
= .4773 = .3944
The proportion of people who get between 11,000 and 17,500 is .4773 + .3944 = .8717 , hence
the percentage of people who get between 11,000 and 17,500 is .8717 x 100 = 87.17%
3. Calculating probabilities
a. What is the probability of a member of this community getting between K11,000 and
K15,000?
̅
Z=
–
=
= -2
= .4773
The probability of a member of the community getting between K11,000 and 15, 000 is
0.4773. Note that the proportion of the area under the normal curve is also indicate of the
probability of a given event.
4. Computing the total number of people in a given interval

a. If the community has a population of 1200, how many people get between K11, 000 and
K17, 500?
̅ ̅
= =
– –
= =
= =
54
= -2 = 1.25
= .4773 = .3944
The proportion of people who get between 11,000 and 17,500 is .4773 + .3944 = .8717, hence
the percentage of people who get between 11,000 and 17,500 is .8717 x 100 = 87.17%. To find
the actual number of people who get between K11, 000 and K17, 000, we simply find 87.17% of
1200.
= X1200
= 87.17 X 12
= 1046.04
= 1046 people
The number of people cannot be expressed in decimals, so when dealing with people, the answer
should be rounded off to a whole number.
The basic rule in using z –scores is that if the z – scores are located on the opposite sides of the
centre, you add the corresponding proportions of the area under the curve. On the other hand, if
z-scores are on the same side of the centre, you subtract the corresponding proportions.
7.0 Summary
You have been informed in this unit that one type of symmetrical distribution called
the standard normal distribution has great significance in the field of statistics. It is symmetrical,
unimodal, asymptomatic, the areas under the curve represents a probability of 1 and it has a fixed
proportion of observations which lie between the mean and fixed standard deviation units. The
unit has also shown that z – scores or standard scores express the distance between a specific
observation and the mean in standard deviation units.
55
After reading through unit five, you should answer the questions below. It is
unit five again.
Using the table of the areas under the normal curve attached in the appendices, find the
following;
a. The areas associated with a Z- score of 2.33

b. The area associated with a Z- score of -1.08
c. The area between a Z – score of -0.52 and 2.01
d. The area between a Z –score of – 3.06 and -0.08
In a normally distributed sample of 1200 children, the mean age is 14.4 years and the standard
deviation is 2.5 years.
a. How many children are between 12 and 16 years old?

b. How many children are older than 18 years?
c. What is the percentage of children who are younger than 8 years?
d. What is the probability of a child in this sample being above 15 years?
56
UNIT SIX: INTRODUCTION TO ESTIMATES WITH SAMPLES
1.0 Introduction
This unit gives an introduction to estimates with samples. The unit begins with discussing the
sampling distribution of means. Thereafter, we shall discuss the standard error of the mean and
the standard error of a proportion. We will then focus on confidence levels, confidence limits and
confidence intervals. The last part of this unit will dwell on the estimation of parameters based
on samples.
2.0 Aim of the Unit
The aim of this unit is to introduce you to estimation of parameters based on sample
statistic.
Estimate parameters based on samples.

Distinguish between point estimates and confidence interval estimates

5.0 Time Required
57
It will take you approximately two hours to work through this unit.
6.0 Main Contents
The purpose of sampling is to enable researchers estimate the population parameters

based on sample statistics. For example, when you want to estimate the outcome of a presidential
election, you may sample a few people and use the results obtained from the sample to estimate
what the results of the actual election will be. Similarly when you want to know the average
income of people in Zambia you would select a sample and use the value obtained from the
sample to estimate the population mean, although the sample mean is unlikely to be exactly the
same like the population mean. There will inevitably be some degree of error in the sample.
If you took a very large number of samples of the population, you would expect all the
sample means to be close to the true population means, but spread around the mean with some
sample means higher and some sample means lower than the true population mean. These means
will not all be the same and they can be plotted as a frequency distribution. This distribution is
called a sampling distribution of the means. For example, if you draw 100 samples of the same
size to find the mean income of households in Zambia, you will end up with 100 different sample
means. The sampling distribution of means will be normal and it will bear all the characteristics
of the normal curve.
The sampling distribution of sample means is important in statistical inferences because it

is used in significance tests and hypothesis testing. Although the sampling distribution of means
is similar to the ordinary normal curve, the differences are that whereas the ordinary normal
curve deals with ordinary observations, the sampling distribution of means deals with the
distribution of sample means.
58
6.1 The concept of Standard Error of the Mean
The standard deviation of a sampling distribution is known as the standard error of the
mean. It is given by the formula;
̅ =
√
̅ = Standard error
S = Standard deviation of the sample
√ = Square root of the sample size
The importance of the standard error is that it is used to estimate how accurately the
sample mean estimates the population mean. The smaller the standard error, the more accurately
the sample mean can estimate the population mean. When you increase the sample size, the
Standard Error (SE) reduces.
6.2 Standard Error of a Proportion
Many surveys and samples attempt to estimate a proportion, rather than an arithmetic
mean. This is true of surveys of attitudes or opinions about an issue or the percentage of times an
issue occurs. Suppose you want to know what proportion of people would vote for a name
political party in elections, several sample may be obtained and the proportion of votes in each
sample could be arranged into a sampling distribution which would be; normally distributed; It
would have a mean equal to the proportion of voter in the population who would vote for a
named political party and; it would have a standard error equal to the standard deviation of a
proportion.
The standard error of a proportion is given by the formula;
̅ √ where; p = proportion; q = 1- p and; n = sample size
59
6.3 Confidence Levels, Confidence Limits and Confidence Intervals
The properties of a normal distribution, together with the rule that sample means are
normally distributed around the true population mean with a standard deviation of equal to the
standard error, we can predict the following (using the normal distribution table);
 68% of all the sample means will lie within 1 standard error of the mean. In other words,
there is a 68% probability that the population mean lies within the range: sample mean ±
1standard deviation.
 95% of all sample means will be within 1.96 standard errors of the population mean. In
other words, there is a 95% probability that the population mean lies within the range:
sample mean ± 1.96 standard deviations.
 99% of sample means will be within 2.58standard errors of the population mean. In other
words, there is a 99% probability that the population mean lies within the range: sample
mean ± 2.58 standard deviations.
These degrees of certainty are known as confidence levels and the ends of the ranges around
the sample mean are called confidence limits. The ranges are called confidence intervals. The
word confidence is used because you are trying to find out how confident you can be that your
sample is representative of the population as a whole.
6.4 Estimation of Parameters
There are two types of parameter estimates that can be used; these are point estimates and
confidence interval estimates.
a. Point estimates are the simplest way of estimating population parameter; you merely take
the sample statistic to represent the population parameter.
Sample statistic = Population parameter
When using point estimates, an error in the estimation process is likely to occur as the
sample statistic might be below or above the population parameter. This error is referred
to as sampling error, which is the difference between the population parameter and the
sample statistic.
60
b. Confidence Interval Estimates. A confidence interval estimate is a particular kind of
estimate of a population parameter. It denotes the range within which a population
parameter might be found it also denotes the level of confidence or probability that the
population parameter lies within the stated confidence limits.
C.I.E = Sample statistics ± Critical value x Standard error
A critical value is the Z – score value which shows the ranges in which a given
confidence level falls under the normal curve. It is like a confidence limit for a given
level of confidence. For example, a confidence level of 99% will be represented by a Z –
score of ± 2.58.
Example 1: Suppose you are trying to estimate the mean age of UNZA students, you
sample 100 students; the mean age for this sample is 24.7 years and the standard deviation is
1.02 years. Use 95% confidence level to establish the mean age of UNZA students.
C.I.E = Sample statistic ± Critical x standard error
= 24.7 ± 1.96 x 0.1

= 24.7 ± 0.196
= 24.5 – 24.9
Interpretation: You can be 95% certain that the mean age of all UNZA students lies between 24.5
years and 24.9 years.
Note that the standard error was computed using the formula;
̅ =
√
=
√
= 0.1
61
The critical value for was found by following the following steps;
95% confidence level means we should identify the z – scores in which 95% of the area
under the normal curve falls. You know that the whole area under the normal curve adds up to
100% with 50% on either side of the mean. Similarly, a confidence interval of 95% will have
= 47.5% on either side of the mean. 47.5% represents a proportion of = 0.4750 or .4750
of the area under the normal curve. Having done this, you know you go to the table of the areas
under the normal curve in the appendices of this module and look for the z – score which will
represent an area of .4750. In this case you will note that a z – score of 1.96 represent .4750 of
the area under the normal curve.
Example 2: A researcher randomly selects 100 students at UNZA to determine their

utilization of the University library. Out of the sample, 55% said they prefer studying from their
rooms. Estimate the proportion of students who prefer studying from the library at 99% level of
confidence.
Standard error of a proportion is = √
=√
= 0.04975
In this case p = = 0.55
q = (1 –p)
1 – 0.55
0.45
C.I.E = Estimated proportion ± critical value x standard error

= 0.55 ± 2.58 x 0.4975
= 0.421645 – 0.678355
= 42% - 68%
Interpretation: You can be 99% confident that 42% - 68% of UNZA students prefer studying
from their rooms.
62
7.0Summary
You have been informed in this unit that the purpose of sampling is to enable
researchers estimate population parameters based on sample statistics. You have further been
informed that the importance of the standard error is that it is used to estimate how accurately the
sample mean estimates the population mean. The smaller the standard error, the more accurately
the sample mean can estimate the population mean. When you increase the sample size, the
standard error reduces.
After reading through unit six, you should answer the questions below. It is important
for you to attempt the questions in order to gauge how much you have comprehended the
material presented in this unit. If you are having challenges in answering the question, read unit
six again.
Given a random sample of 100 malnourished adults with a mean weight of 26kg and a standard
deviation of 5.2kg;
a. Find the point estimate of the malnourished adults mean weight.

b. Compute the standard error and interpret the answer.
c. Establish the 90% confidence interval estimate for the population mean.
d. Establish the 98% confidence interval for the population mean.
63
UNIT SEVEN: HYPOTHESIS TESTING
1.0 Introduction
This unit discusses hypothesis testing. The unit begins with highlighting the steps in
hypothesis testing and defining the major concepts in hypothesis testing. Thereafter, the unit will
discuss hypothesis testing between means of large samples. We will then look at hypothesis
testing between proportions of large samples as the last part of this unit.
2.0 Aim of the Unit
The aim of this unit is to discuss the procedure for hypothesis testing.
Outline the steps in hypothesis testing.

Define the major concepts in hypothesis testing.
Conduct hypothesis testing for difference between means of large samples.
Conduct hypothesis testing for differences between proportions of large samples.

McNabb D. (2009) Research Methods for Political Science: Quantitative and Qualitative
Methods, New Delhi: PHI Learning Private Ltd
64
5.0 Time Required
6.0 Main Contents
We use samples to make estimates of the population mean or proportion. In the previous
section, it was explained how we can estimate a confidence interval for a population mean or
proportion, at a given level of confidence. We can also use sample statistics to test a theory or
assumption that we have made about the population. This is referred to as hypothesis testing. The
purpose of hypothesis testing is to test an assumption that we can make about the population.
6.1 Procedure in Hypothesis Testing
The following are the steps which are followed in hypothesis testing:
1. Formulation of the research and null hypothesis

a. Research hypothesis. A research hypothesis is a hypothesis which is derived from
theory, observations and literature review. A research hypothesis is a hypothesis which a
researcher is interested in testing in order to prove or disapprove it. Hypotheses are
expressed in a directional way. Research hypothesis is denoted by . For example, a
research hypothesis may read as; Male students participate more in demonstrations than
female students. In this case a sign will be used to denote male students and will be
used to denote female students. Hence the research hypothesis in question would be
presented as:
:
b. Null hypothesis. A null hypothesis is an antithesis or opposite of the research
hypothesis. It is the hypothesis that is directly tested and contradicts the true hypothesis.
The null hypothesis is denoted by The null hypothesis of the research hypothesis
above would read as; Male students are less involved in demonstrations than female
students. Hence the null hypothesis in question would be presented as:
65
:
In hypothesis testing rejection of the null hypothesis increases the probability that the
research hypothesis would be correct. Whereas acceptance of the null hypothesis increases the
likelihood that the research hypothesis could be wrong.
2. Choosing appropriate statistical tests and making assumptions
Once the hypothesis has been formulated, the next thing is to decide on the appropriate
statistical test that must be done. This choice has certain assumption or conditions which must be
met. These assumptions are used to justify the choice of the appropriate statistical test. The most
important assumptions are that; (i) assumptions relating to the distribution of parameters; (ii)
assumptions relating to the scale of measurements of variables and; (iii) assumptions relating to
the sampling designs, this is always assumed to be random sampling. Based on these
assumptions, there are two major types of statistical tests used in hypothesis testing, these are
parametric tests and non parametric tests. A parametric test is a statistical test based on several
assumptions about the parameters of the population from which it is drawn. Among the most
important ones are that the observations must be drawn from a normally distributed population
and that the variables are measured on either the interval or ration scale. On the other hand, a non
parametric test does not require the normality of the distribution or the variables to be interval or
ratio.
3. Obtaining the appropriate sampling distribution
There are two types of sampling distribution which are used in hypothesis testing. These
are determined by the sample size. The standard normal distribution or Z – distribution will be
used if the sample size is more than 30 (n 30). If the sample size is less than 30 (n 30), then
the t – distribution may be used.
Once the choice of an appropriate sampling distribution has been made, the sampling
distribution selected will be used to choose the critical values.
66
4. Choosing the significance level and critical regions
a. Critical Regions
Once the sampling distribution has been obtained, determination of the critical values
follows. The critical value in the case of a Z – distribution is the Z – score obtained from the
standard normal curve. The critical value in hypothesis testing determines the boundary or
basis on which you accept or reject the null hypothesis. In hypothesis testing, a test statistic
must be computed. The test statistic is also known as the observed value.
Figure 7.1: Areas of rejection and acceptance in hypotheses testing
Area of Acceptance
Area of Rejection
b. Level of significance
The level of significance is used to compute the critical values to use in demarcating the
areas of rejection or levels of acceptance. The level of significance is the probability of
rejecting the null hypothesis ( ) when it is in fact true. The probability of rejecting a null
hypothesis when it is in fact true is referred to as type I error. The level of significance is
denoted by (Alpha) symbol. The opposite of type one error is type II error. Type two error
is the probability of accepting the null hypothesis when it is false. Symbolically, it is denoted
by beta (β).
67
c. Setting the critical regions using the level of significance.
i. Critical regions for a two tailed hypothesis test. A two tailed hypothesis test refers to a
situation where the research hypothesis is non – directional. For example, the average income of
civil servants is not equal to K2,500.00
: ≠ K2, 500.00
At 5% level of significance, the total proportion of the area of rejection is .05. Since this
proportion is evenly split between the two sides of the standard normal distribution it is going to
be divided by two in order to find the area of rejection on both sides of the curve. Therefore, the
critical region will be =0.025. The Z-score which demarcates the region with a proportion of
0.025 on the left of the curve is -1.96 and on the right of the curve is +1.96 as shown in the
diagram below.
Figure 7.2: Areas of acceptance and rejection for a two tailed hypothesis at 5% level
of significance Area of Acceptance
0.4750 0.4750
0.025 0.025
-1.96 +1.96
Area of Rejection
68
At 1% significance level, the proportion of the area of rejection is .01. Since this
proportion is evenly split between the two sides of the standard normal distribution, it is going to
be divided by two in order to find the region of rejection on either side of the curve. Therefore,
the critical region will be = 0.05. The Z –score which demarcates the region of 0.05 on the
left of the curve

e is – 2.58 and on the right of the curve is + 2.58.
ii. Critical regions for a one tailed hypothesis test. A one tailed research hypothesis refers
to situation where the research hypothesis is directional. For example, the average income of
civil servants is more than K2, 500.00 ( : K 2,500.00) or there are more men involved
more in decision women ( : ). In this case, you place the critical value in one tail only,
in the right side of the curve. The area of acceptance is the area to the left of the critical value;
the extreme right is the area of rejection.
At 5% level of significance, the proportion of the area of rejection under the standard
normal curve is 0.05. Since this area is only on one side of the curve, it will not be divided by
two as was the case in a non – directional hypothesis. Therefore, the z – score which represents
the critical region of 0.05 is 1.65. At 1% level of significance the proportion of the area of
rejection under the normal curve is 0.01, the Z-score which represents this critical region is 2.58.
The z – score values which correspond to the level of significance determine where the critical
regions are located, that is the cut off points or values for the rejection or acceptance of the null
hypothesis.
5. Computing the test statistic
At this stage you run the data in your computer and come up with a test statistic also
known as the observed value. This will come in form of a z – score or t- score depending on the
sample size.
6. Decision making
Depending on the area of acceptance or rejection of the null hypothesis, make a decision
whether to accept or reject it based on the test statistic or observed value.
69
7. Conclusion
Make a conclusion based on your decision, go beyond the mathematics and express
yourself in your own words.
Example of hypothesis testing
In a dispute between Chimuka Mining Company Limited and the workers’ union over
employee remuneration, management claims that the workers earn more than K 7,500 but the
union dispute the claim. As a labour consultant, you are hired to ascertain this claim which the
union disputes. You select a random sample of workers at the mine, n = 100. The mean income
of workers based on the sample is K7, 900 and a standard deviation of K 1,500. You decide to
establish if the claim by management is true by using a 5% level of significance.
Step 1. Formulate the research and null hypothesis
: = K 7,500
: K 7,500
Step 2. Assumptions
i. The subjects are randomly and independently selected

ii. The population distribution is normal
iii. The scale of measurement is interval
Step 3. Decision Rules
Given 5% level of significance and using a standard normal distribution, the critical value
for a one tailed directional test is 1.65.
If +1.65 Accept
If +1.65 Reject
70
Figure 7.3: Areas of Acceptance and rejection for a one tailed hypothesis at
Area of Acceptanc
Area of Rejection
0.05
1.65
Step 4. Computation
=
̅
̅ = Standard error
̅ =
√
=
√
= 150
=
̅
–
=
= 2.67
71
Step 5. Decision
Compare the observed value with a critical value of 1.65. Because is greater than 1.65, we
reject the null hypothesis ( ).
Step 6. Conclusion
It is highly likely that Chimuka Mining Limited is right in its claim that it pays its workers more
than K 7,500.
6.2 Testing Hypothesis about a Proportion
A nutritionist at the Ministry of Health in Lusaka claims that more than 50% of children
in Lusaka are malnourished. A researcher selects a sample of 400 children and finds that 54% of
children have a protein deficiency. Test this claim by the researcher at 5% level of significance.
Step 1. Formulation of research and null hypothesis
: P = .50
:P .50
Step 2. Assumptions
i. The children are randomly and independently selected

ii. The population distribution is normal
iii. The level of measurement is interval
Step 3. Decision rules
Given α .05, for a directional hypothesis we require a one tailed test, with the critical
value of 1.65.
If +1.65 Accept
If +1.65 Reject
Step 4. Computation
72
=
P = Hypothesized proportion
SE = Standard error of a proportion
SE = √
q=1–p
= 1 - .50
= .50
SE = √
=√
= 0.025
=
–
=
= 1.60
Step 5. Decision
Compare the observed value with a critical value of 1.65. Because is less than 1.65,
we accept the null hypothesis ( ).
Step 6. Conclusion
The claim by the nutritionist at the Ministry of Health that more than 50% of children in Lusaka
are malnourished is not true.
73
6.3 Testing a Hypothesis concerning differences between Means for Large Sample
Test of difference between two means is used when you want to compare two means
reflecting two samples or you may compare two means of two populations and then establish if
there is a significant difference between in the mean of female student and male student.
For example, there is a debate going on at the University of Zambia concerning the
performance of concerning self sponsored students and those on bursary. Management argues
that the performance of self sponsored students and those on bursary is the same while others are
saying it is not. As an independent researcher, you are asked to establish management’s claim
that there is no significant difference in the performance of the two categories of students at 1%
level of significance. You sample both sets of students and come up with the following
information.
Self Sponsored Students Sponsored Students
= 50 = 50
= 88.7 = 83.1
= 7.8 = 11.4
Step 1: Formulation of the research and null hypothesis
Step 2: Assumptions
i. Subjects are randomly and independently selected

ii. The groups are independent from one another
iii. The population distribution is normal
iv. The scale of measurement is interval
74
Step 3: Decision Rules
Given α .01, for a non – directional hypothesis, if -2.58 +2.58, accept
If -2.58 or ≥ +2.58, Reject
Step 4: Computation
=
√
=
√
= 2.87
Step 5: Decision
Compare the observed value with the critical value formulated under decision rules. Because
is greater than 2.58, you reject the null hypothesis ( ).
Step 6: Conclusion
There is no significant difference between the performance of self sponsored and students on
bursary. The claim by management is correct.
6.4 Testing Hypothesis concerning differences between Proportions
Test of difference between two proportions is used when you want to compare two two
populations and then establish if there is a significant difference between the proportions of the
two populations.
75
For example; A research firm is investing citizen’s satisfaction with service delivery by
government. The comparison is being made between a rural and urban area.
Rural Urban
= 1000 = 1, 091
= =
= 0.2 = 0.22
=1- =1-
= 1 – 0.2 = 1 – 0.22
= 0.8 = 0.78
Given this information would you agree that people in urban areas are more satisfied with
service delivery by government than those in rural areas? Test the hypothesis using 5% level of
significance.
Step 1: Statement of hypothesis
: =
:
Step 2: Assumptions
i. are randomly and independently selected

ii. The groups are independent from one another
iii. The population distribution is normal
iv. The scale of measurement is interval
Give α .05 for a directional hypothesis, if > - 1.65 Accept ,
76
if < -1.65 Reject
Step 4: Computation
=
√
=
√
= - 1.12
Step 5: Decision
Compare the observed value with the critical value formulated under decision rules.
Because is greater than - 1.65, we accept the null hypothesis ( ).
Step 6: Conclusion
There is no difference in satisfaction with service delivery by the government between

people in rural and urban areas.
7.0Summary
You have been informed in this unit that the purpose of hypothesis testing is
to test an assumption that we can make about the population. The procure for hypothesis testing
involves a series of steps, theses are; formulation of the research and null hypothesis; choosing
the appropriate statistical test and making assumptions; obtaining the appropriate sampling
distribution; choosing the significance level and critical regions, computing the test statistic,
decision making and; conclusion.
77
After reading through unit seven, you should answer the questions below. It is
unit seven again.
a. The management of Spar Supermarket in Choma is suspicious that the 500 grams of
beans
supplied to the store are below the required weight. To test this suspicion, 100 packages of
beans are selected at random, and the average weight of the packages in the sample is 497
grams with a standard deviation of 5 grams. Is the suspicion of management justified? Test
at 5% level of significance.
b. A sample of 400 voters in Lubasenshi constituency and 100 voters in Katombora

constituency showed that only 33% and 10% of were interested in casting their ballots in
the next Presidential, Parliamentary and Local Government elections. At 10% level
significance, test the hypothesis that there is no significant difference in interest in voting
between the two constituencies.
78
UNIT EIGHT: ANALYSIS OF VARIANCE (ANOVA)
1.0 Introduction
This unit discusses the analysis of variance (ANOVA). The unit begins with explaining how
the ANOVA is used in hypothesis testing. Thereafter, will describes the steps used when
undertaking an analysis of variance. A practical example will be used to explain how the analysis
of variance is computed.
2.0 Aim of the Unit
The aim of this unit is to explain the use of ANOVA in hypothesis testing.
Distinguish between ANOVA and other methods of hypothesis testing.

Outline the steps in the computation of the F test statistic
Test hypotheses using ANOVA.

5.0 Time Required
79
6.0 Main Contents
Analysis of variance tests are tests of hypothesis that compare the variation or dispersion
between different samples or groups to establish whether there is a significant difference
between the samples. This hypothesis test makes use of the F test which is based on the F
distributions. To carry out a test, we compute an F test statistic from the data in the samples and
compare this with a critical value in an F distribution table. There is a different F distribution
table for each level of significance. If the F test exceeds the critical value, we reject the null
hypothesis and assume that there is a significant difference between the samples. This F
distribution table is attached as an appendix at the end of this module.
6.1 The F test Statistic
The F test statistic compares two or more different samples and the variances of the
values in those samples.
F=
VBG is the abbreviation for variance between groups and VWG is an abbreviation for variance
within groups.
For example, suppose you are studying three tutorial groups to establish whether there is
a significant difference between in the performance of students in three different tutorial groups.
Four (4) random samples are collected from each of the three tutorial groups to establish if there
is a significant difference in the performance of students among the three groups using 5% level
of significance.
Tutorial Group 1 Tutorial Group 2 Tutorial Group 3 Total

80 70 63 213
92 81 76 249
87 78 70 235
83 74 58 215
Total 242 303 267 912
80
Step 1: Statement of Hypothesis
: = =
: ≠ ≠
Step 2: Assumptions
i. Subjects are randomly and independently selected

ii. Groups are independent from one another (mutually exlusive)
iii. Population variances are equal
iv. Population distribution is normal with respect to the variable of interest which is
performance
v. The scale of measurement is interval
Step 3: Decision rules
ANOVA uses a distribution related to the normal distribution called the F – distribution.
In testing hypothesis concerning three means, the computed F – ratio has to be compared to the
critical value given in the table. You need the degree of freedom (df) for the two variance
estimates of between means and variance within means and also the level of significance.
Degree of freedom between groups (dfb) = j – 1
j stands for the number of groups; therefore dfb = 3 – 1 = 2 df
Degree of freedom within groups (dfw) = n – j
n = total sample size for all the groups; therefore dfw = 12 – 3 = 9 df
To get the critical value of F at 5% level of significance, look for the point of intersection
the degree of freedom between means (Degree in freedom in the numerator) and the degree of
freedom within means (Degree of freedom in the denominator). Using the F distribution, the F
critical value for the degrees of freedom calculated for the data given is 4.26
81
If 4.26, Accept
If 4.26, Reject
Step 4: Computation
i. Computation of the grand mean ( ̿ )
This involves summing all the observations in all the columns and in all the roles and then
dividing by the total number of observations.
∑
̿=
= 76
ii. Computation of the Total Sum of Square (TSS)
Computation of the TSS involves finding the sum of squared deviations of each observation from
the grand mean.
SST = ∑ (xij – ̿ )²
̿ = 76
T.G 1 ( - ̿ )² T.G 2 ( - ̿ )² T.G 3 ( - ̿ )²

80 16 70 36 63 169
92 256 81 25 76 0
87 121 78 4 70 36
83 49 74 4 58 324
442 69 529
SST = 442 + 69 + 529
= 1040
iii. Computation of the sum of squares between groups
82
This involves first computing means for each group or category and thereafter finding the
squared deviation of each category to the grand mean.
T.G 1 T.G 2 T.G 3

80 70 63
92 81 76
87 78 70
83 74 58
Total 342 303 267
Mean( ̅ ) = 85.5 75.75 66.75
The means as calculated above are T.G 1 = 85.5, T.G 2 = 75.75 and T.G 3 = 66.75
Squared deviations of the category mean to the grand mean ( ̅ - ̿ ) for each tutorial will be
calculated as follows:
N ( ̅ - ̿) ( ̅ - ̿ )² n( ∑ ̅ - ̿ )²
T.G 1 4 9.5 90.25 361
T.G 2 4 -0.25 0.0625 0.25
T.G 3 4 -9.25 85.5625 342.25
703.5
The calculations above show that 703 is the sum of squares between groups (SSB).
iv. Computation of the variance between groups (VB)
This involves dividing the sum of squares between groups by the degree of freedom between
groups,
VB =
Note that the degree of freedom between groups (dfb) was calculated earlier in this section.
83
VB =
= 351.75
v. Computation of the sum of squares with groups
This involves subtracting the total sum of squares between groups (SSB) from the total sum of
squares (TSS)
SSW = TSS – SSB
= 1040 – 703.5
= 336.5
vi. Computation of the variance within groups (VW)
This involves dividing the sum of squares within groups (SSW) by the degree of freedom within
groups.
VW =
Note that the degree of freedom within groups was calculated earlier in this section..
VW =
= 37.39
84
vii. Compute the F statistic
= 9.41
Step 5: Decision
We will reject the null hypothesis because the F statistic (ratio) is greater than the critical value
of 4.26. If 4.26, Reject
Step 6: Conclusion
There is a significant difference in the avaerage performance of students is the three tutorial
groups.
7.0 Summary
You have been informed in this unit that Analysis of variance tests are tests of
hypothesis that compare the variation or dispersion between different samples or groups to
establish whether there is a significant difference between the samples. This hypothesis test
makes use of the F test which is based on the F distributions. The F test statistic compares two or
more different samples and the variances of the values in those samples. F = , where VBG is
the abbreviation for variance between groups and VWG is an abbreviation for variance within
groups.
85
8.0 Self Assessment Question
After reading through unit eight, you should answer the question below. It is
important for you to attempt the question in order to gauge how much you have comprehended
unit eight again.
Mr. Kakoma planted four different varieties of beans. Each variety was planted on three separate
plots of land and when he harvested, he obtained the following 25 kg bags of beans.
Variety A 60 61 56
Variety B 59 52 51
Variety C 55 55 52
Variety D 58 58 55
Perform an analysis of variance to test at 5% level of significance to establish if there is a

significant difference between the samples. (The F distribution table is provided in the
appendices).
86
UNIT NINE: CHI – SQUARE TEST
1.0 Introduction
This unit discusses the chi-square as a test for significance. The unit begins by explaining the
applications of chi-square. Thereafter, we will describe the steps to be followed when
undertaking a chi–square test. A practical example will be used to explain how to test for
significance using chi – square. The last part of this unit will look at the chi-square test of
goodness fit.
2.0 Aim of the Unit
The aim of this unit is to explain chi-square in significance testing.
Distinguish between chi-square and other methods of hypothesis testing.

Test hypotheses using chi-square.

Brown and Co. Ltd
87
5.0 Time Required
6.0 Main Contents
The chi – squared distribution is another distribution used in significance testing. It is

often referred to by means of the Greek letter χ². The chi – squared distribution can be used to
test a sample of items in order to decide whether items in a sample are distributed according to a
preconceived or expected distribution pattern. Because chi – square tests whether something
conforms to a given pattern or not, chi – square tests are often referred to as goodness of fit tests.
The main application of χ² test is testing for association or correspondence between

attributes. Significance tests on attributes usually take the form of a contingence table. A number
of research problems in the social sciences often involve the use of more than one variable. If
you have observations taken on more than two or more variables we have what is referred to as a
bivariate table and where you have more than two variables, it is referred to as a multivariate
table. Bivariate data is often drawn in a two way table with one of the variables along the column
and another variable down the rows. Two way tables are called contingency tables. This is
because the research hypothesis normally states that two variables are related and this is the same
as saying the two tables are contingent.
88
Example:
A research wants to determine whether there is a relationship between religious affiliation

and attitudes towards abortion.
Attitudes towards Religious Affiliation

abortion Protestant Catholic Total
For 126 (64%) 99 (38%) 225
Against 71 (36%) 162 (62%) 233
Total 197 261 458
: There is no relationship between religious affiliation and attitude towards abortion
: There is a relationship between religious affiliation and attitude towards abortion
Step 2: Assumptions
i. The subjects for each group are randomly and independently selected
ii. The groups are independent
iii. Each observation qualifies for only one category
iv. The scale of measurement for chi – square can either be nominal or ordinal. In this
case, the data is nominal.
Step 3: Decision rules
Given α = 0.05 and df = (r – 1) ( c – 1)
r = number of row
c = number of columns
df = (2 -1) (2-1) = 1
89
To find the critical value, go to the Critical Values of Chi–Square, the table attached at the
module, find the intersection between the df of 1 (in the first column) and the probability of 0.05
or .05 across the row. The intersection between these two gives you the critical value of 3.84.
If ˂ 3.84 accept
If ≥ 3.84 reject
Step 4: Computation
=∑
O = Observed value
E = Expected value
- ( – )² –
126 96.78 29.22 853.84 8.82

99 128.22 -29.22 853.84 6.66
71 100.22 -29.22 853.84 8.52
162 132.78 29.22 853.84 6.43
= 30.43
Step 5: Decision
We will reject the null hypothesis because 3.84.
Step 6: Conclusion
There is a relationship between religious affiliation and attitude towards abortion.
90
6.1 Chi–Square Goodness of Fit Test
The chi–square goodness of fit test is used to determine the extent to which our
expectations match with reality.
For example
The Managing Director of Michelo Stores believes that the number of customers who
come to the shop from Monday to Saturday is 1,632. However, the cashier disputes this claim.
You are hired as a consultant to verify this claim.
: Number of customers is evenly spread
: Number of customers is not evenly spread
Step 2: Assumptions
i. Nominal scale
ii. Sample is large
Given α + 0.05 the df = K – 1= 6-1 = 5df
The critical value is 11.07
If ˂ 3.84 accept
If ≥ 3.84 reject
91
Step 4: Computation
–
χ² = ∑
O E O–E (O – E)² –
Monday 1525 1632 -107 11499 7.02

Tuesday 1711 1632 79 6241 3.82
Wednesday 1655 1632 23 529 0.32
Thursday 1497 1632 -135 18225 11.17
Friday 1603 1632 -29 841 0.52
Saturday 1801 1632 169 28561 17.5
= 40.35
Step 5: Decision
We will reject the null hypothesis because 3.84.
Step 6: Conclusion
The number of customers who visit the shop is not evenly spread.
7.0 Summary
You have been informed in this unit that the chi – squared distribution can be
used to test a sample of items in order to ascertain whether items in a sample are distributed
according to a preconceived or expected distribution pattern. Because chi – square tests whether
something conforms to a given pattern or not, chi – square tests are often referred to as goodness
of fit tests. The main application of χ² test is testing for association or correspondence between
attributes.
92
After reading through unit nine, you should answer the questions below. It is
unit nine again.
Suppose a bottling company expects to sale 480 pallets of soft drinks over the next four months,
with an equal volume of sales in each month. In the past four months, sales by the company were
150, 110, 100 and 120 pallets. Does the historical data suggest that the demand each month is
uniform? Test at 5% level of significance.
93
UNIT TEN: REGRESSION
1.0 Introduction
This unit discusses regression and correlation. The unit begins with giving an overview of
regression and discussing linear regression. Thereafter, we will discuss the formulation of the
regression equation. We will then discuss correlation and will later focus on the coefficient of
determinant as the last part of this unit.
2.0 Aim of the Unit
The aim of this unit is to discuss regression.
Distinguish between regression and correlation.

Compute the regression equation.
Calculate the correlation coefficient.
Explain the coefficient of determinant.

Brown and Co. Ltd
94
5.0 Time Required
It will take you approximately three hours to work through this unit.
6.0 Main Contents

Regression is a type of analysis which focuses on understanding the relationship or
association between two or more variables. It deals with variables that are measured on an
interval scale. The relationship or association is shown by two different samples of numbers that
are linked together. For example, questions of association might be the relationship between
population growth and crime, the association between the time that a car travels and the
relationship between years of work experience and income.
If two variables are associated, then as the numbers in one sample vary, their partner
numbers in the second sample vary in a related fashion. In statistical analysis we say that the
numbers co-vary. The simplest way in which numbers can co-vary is in a linear fashion. A
relationship is said to be linear if a change in one variable results in a change in the second
variable. The term linear is used when you plot a line on the graph which depicts that the change
in X results in a change in Y. Therefore, regression helps to predict the future situation based on
the established relationship or association.
6.1 Linear Regression
Linear regression is a way of expressing a relationship between two interval variables

using the linear function. Scientists use regression to find some algebraic expression by which to
represent formal relationships between variables. The equation Y = a + bx is a linear regression
equation meaning that the function describing the relationship between x and y is a straight line.
The letter a in the equation is a constant, it shows the value of y when x is 0. The letter b in the
equation is the regression coefficient, it shows the changes in y which will be occasioned by a
change in x. Ordinarily, researchers display the observations of x and y and the regression line
connecting them in form of a graph. The variables x and y are represented by two intersecting
axes. X which is the independent variable is also referred to as the prediction or predicator. Y
95
which is the dependent variable is also referred to as the criterion. The formula for finding the
∑ – ∑ ∑
value of a in the regression equation is a = ̅ - b ̅ while the formula for b = ∑ ∑
An example of the regression of x and y is depicted using the data set on time and distance which
a car travels as shown in figure 10.1.
Figure 10.1: Time and Distance Travelled by a Car
Time and Distance

16
14 y=x
R² = 1
12
Distance (Km)
10
8
6
4
2
0
0 2 4 6 8 10 12 14 16
Time (minutes)
The regression line does not always pass through the intersection of x and y axes. When a
straight line intersects the y axis, a constant needs to be introduced to the regression equation.
The constant is represented by the letter a is called the y intercept. The intercept reflects the
value of y when x is 0. For example, three different regression line which are considered in
figure 10.2, 10.3 and 10.4 have different values for a and b. The three different values of a (6, 1,
2) are reflected in the three different intersections of the graphs being considered (figure 10.2,
10.3 and 10.4). The different values of b (-3, 0.5, 3) reflect the steepness of the slope. The higher
the value of b, the steeper the slope. The sign of b expresses the relationship between x and y.
When b is positive, an increase in x is accompanied by an increase in y. On the other hand, when
b is negative, y decreases as x increase
96
Figure 10.2: Negative Regression
8
6
4
2
0
Y - Axis
-2 0 1 2 3 4 5 6 7
-4
-6
-8
-10
Y = 6 -3x
-12
-14
X - Axis
Figure 10.3: Positive Regression

5
4.5 y = 1 + 0.5x
4
3.5
3
Y - Axis
2.5
2
1.5
1
0.5
0
0 1 2 3 4 5 6 7 8
X - Axis
97
Figure 10.3: Steep Positive Regression
16
14 y = 2 + 3x
12
10
Y - Axis
8
6
4
2
0
0 2 4 6 8 10 12 14 16
X - Axis
98
Given the data below on percentage of urban population and robbery rates in the USA, compute
the following:
Table 10.1: Percentage of urban population and robbery rates in 10 states in the USA.
(Adopted from Nachmias and Nachimias, p. 378)
State Percentage of Robberies per XY X² Y²
urban 100,000
population population
(x) (y)
Massachusetts 84.3 150 12,645 7,106.49 22,500

Wisconsin 65.7 105 6,898.5 4, 316.49 11,025
South Dakota 50 26 1,300 2,500 676
Virginia 69.4 132 9,160.8 4, 816.36 17,424
Sacramental 54.6 176 9, 609.6 2, 981.16 30,976
Texas 80.3 180 14, 454 6, 448.09 32,400
Arizona 87.5 174 15, 225 7,656.25 30,276
California 92.6 331 30, 650.6 8,574.76 109,561
Arkansas 53.5 126 6, 741 2,862.25 15,876
Illinois 89 131 11,659 7,921 17,161
Totals 726.9 1,531 118,343.5 55,182.85 287, 875
a. The regression equation
The regression equation is given byY = a + bx. To find the regression equation, we first have to
calculate the value of b and a.
∑ – ∑ ∑
b= ∑ ∑
–
=
99
–
=
= 3.01
a= ̅–b ̅
∑
Therefore, to find a, we should first compute the mean of the y variable. ̅ = = = 153.1,
∑
similary, we should also calculate mean of x. ̅ = = = 72.69.
a = 153.1 – (3.01) (72.69)
= 153.1 – 218.76
= -65.7
Therefore, the regression equation is:
̂ = -67.5 + 3.01 (x)
In the above equation, the intercept is -67.5 and the regression coefficient is 3.01.
6.2 Correlation
Correlational analysis aims at measuring the degree of the relationship between two variables
and expresses it through a correlational coefficient. The correlational coefficient is given by the
formula:
∑ – ∑ ∑
r=
√ ∑ ∑ ∑ ∑
Correlation is an extension of regression analysis insofar as regression expresses

mathematically the law underlying the relationship between variables whereas correlation aims
at measuring the type and strength of the relationship which exists between variables. The other
difference between correlation and regression is that the regression coefficient as a measure of
association is asymmetrical in the sense that its values depends on which variable is considered
100
as the criterion (dependent variable) and which one is the predicator (independent variable). It
measures the amount of change in the criterion as a function of one unit change in the predicator.
However, it is sometimes not possible to specify which variable is the predicator and which
variable is the criterion. For example, when measuring the association between a students’
performance in chemistry and in biology one variable is not clearly the predicator and the other
the criterion. It is therefore desirable to come up with a measure of association that is
asymmetrical and expresses the degree of relationship between variables. The correlation
coefficient meets both these requirements.
The correlation coefficient is represented by ‘r’ and it ranges between -1 and 1. That it -1
is equal to or less than r and r is less than or equal to 1 ( -1 ≤ r ≤ 1). On this basis, a researcher
can conclude whether the observed relationship is real (significant). If the correlation coefficient
(r) is close to -1, then the relationship between variables is a strong negative correlation. If the
correlation coefficient r is close to 1, then the observed relationship is a strong positive
correlation. If r is equal to 1, then it is a perfect positive correlation. If r is equal to -1, then it is a
perfect negative correlation. If r is equal to 0, then there is absolutely no correlation.
The correlation coefficient for the data on population and crime in the USA considered under
regression would be calculated as follows;
∑ – ∑ ∑
r=
√ ∑ ∑ ∑ ∑
–
r=
√
–
r=
√[ – ]
r=
√
r=√
r=
r = 0.63
101
The correlation coefficient we have computed show that there is a strong positive
correlation between the percentage of urban population and crime per 100 000 population in the
USA.
6.3 Coefficient of Determinant
Having computed the correlation coefficient, we can also calculate the coefficient of
determinant. The coefficient of determinant allows one to determine the proportion of variability
of y that can be attributed to x. The coefficient of determinant is given by the formula:
r² x 100 and is expressed as a percentage. In the case of the example we have considered, the
coefficient of determinant will be:
r² x 100
(0.63)² x100
0.3969 x 100
39.69%
This means that the percentage of urban population accounts for 39.69% of the factors which
cause crime in the USA. Other factors, beside percentage of urban population account for
60.31%
7.0 Summary
You have been informed in this unit that regression is a type of analysis which
focuses on understanding the relationship between two or more variables which are measured on
the ratio or interval scale. Linear regression is a way of expressing the relation between variables
using the linear function. You have further been informed that correlation analysis aims at
measuring the degree of the relationship between two variables and expresses it through a
correlation coefficient. The coefficient of determinant allows one to determine the proportion of
variability of Y that can be attributed to X.
102
After reading through unit ten, you should answer the questions below. It is important
for you to attempt the questions in order to gauge how much you have comprehended the
material presented in this unit. If you are having challenges in answering the question, read unit
ten again.
The table below shows the relationship between copper prices and Government revenue
in Zambia during a seven year period.
Government Revenue and Copper Price, 1967 - 1973

Year Copper Price (K/ton) Government
X revenue (K million)
Y
1967 810 276
1968 887 306
1969 1, 048 401
1970 1, 011 432
1971 767 309
1972 746 315
1973 1, 156 385
a. Compute the regression equation to show the dependence of government revenue on

copper price.
b. Interpret the meaning of the observed regression coefficient within the context of the
equation.
c. If government revenue in 1979 was 210 (K million), what was the price of copper?
d. Compute the correlation coefficient and interpret the answer.
e. Compute the coefficient of determinant and interpret the answer
103
UNIT ELEVEN: THE RESEARCH REPORT
1.0 Introduction
This unit explains the format for PAS 2014 research report. The unit begins by giving an
outline of the research report. We will then discuss in detail the each component of the research
report.
2.0 Aim of the Unit
The aim of this unit is to introduce you to research report writing.
Outline the components of a research report.

Write your research report for PAS 2014.
5.0 Time Required
It will take you approximately three to work through this unit.
104
6.0 Main Content
Writing a research report is the last step in the research process. The report informs your
readers what you have done, what you have discovered and what conclusions you have drawn
from your findings. The research report will have all the sections which were in your proposal;
this is because the research report is an extension of the research proposals. You cannot write a
valid research report if you did not write the research proposal. Besides having all the sections
contained in the proposal, the report will have findings, discussions and conclusions.
6.1 Format of PAS 2014 Research Report
The PAS 2014 report should have the following:
1. Cover Page
The cover page should have the title of the research and all the other relevant details pertaining
to the institution, school, course, personal details and other relevant information.
2. Preliminary pages
The preliminary pages will contain acknowledgements, table of contents, list of abbreviations,
list of tables and list of figures.
3. CHAPTER ONE: INTRODUCTION
Note that you already have chapter one in your research proposals. All the contents for this
chapter will be as they are in your approved (marked) proposals. The only thing you have to do
is to change the tense. When writing the contents for this chapter, you should use past tense. The
following will be part of the introduction.
Introduction / Background
Statement of the Problem
Study objectives
Overall objective of the study
Specific of the study
Research questions
Hypothesis to be tested (if applicable)
105
Definition of concepts / conceptual framework or theoretical framework
Literature review
4. CHAPTER TWO: METHODOLOGY
Just like chapter one, you have much of what constitutes chapter two albeit some part should
not be included in the report. The report should have the sections listed below but should also be
written in past tense. Ensure that all the corrections which you were advised to make in your
proposal are fully addressed as you work on your report. The following will be part of the
methodology.
Study design
Study site and study population
Sampling size and sampling design
Data collection instruments
Data processing procedure
Problems and Limitations
Ethical issues involved (if any) and how you propose to deal with the
5. CHAPTER THREE: FINDINGS AND DISCUSSIONS
Findings, discussions and conclusions are what distinguish a proposal from a report. The
chapter on findings and discussions should have the following. You write your findings after
having analyzed your data. Data analysis entails turning raw data you have collected at the
preceding stage in to meaningful information. If your study is purely qualitative, you can write
your report based on your field notes from interviews, observations or reviewing of documents
by manually analyzing the contents of your notes (contents analysis) or by using computer
software NVIVO or Ethnograph. If you are using quantitative data analysis, it is necessary to
decide upon the type of analysis required such as frequency tables, cross tabulation and other
statistical analysis.
When presenting findings, you can firstly present finding on the characteristics of your
respondents such as age, sex and other relevant characteristics. Thereafter, start presenting
106
findings in line with your research objectives or research questions. This means that you will
have sub – headings under this chapter which will conform to your objectives. Having presented
findings on your research objectives, you should then moves on to present findings on your
research hypotheses.
If you study is quantitative, you should use frequency tables, bar charts, pie charts and
cross tabulations to present findings on your objectives. However, this does not mean that you
should present a table for every question you asked, you be selective as to what tables to include
in the report. The other information can be presented in narrative form. When testing you
hypotheses, the scale of measurement for the variables will determine the kind hypothesis test. In
a majority of cases, the variables will be nominal and ordinal, this mean chi – square should be
used. If you are using SPSS, the p – value which you get after running a chi – square tests
determines whether you accepted or reject the null hypothesis. A p – values of ≤ 0.05 means
rejection of the null hypothesis and accept the research hypothesis.
For every set of findings which you present, you should come up with a discussion.
Discussion of findings entails you as researcher being able to contextualize your findings.
Explaining the meaning of your findings in light of the prevailing situation and where possible
you can also compare your findings to those of other people who have conducted similar studies.
You can also link your findings to the literature which you reviewed and the try to see how your
findings fit into your conceptual or theoretical framework.
6. CHAPTER FOUR: CONCLUSIONS
When writing the conclusions, highlight the salient findings. This chapter does not require
you to write a lot of things. All you should do is you again follow this sequence of your
objectives and at each stage explain what you have concluded. You may cite the page in your
report where evidence for arriving at that conclusion can be found. This will mostly be linked to
the chapter on findings and discussions.
Bibliography
Appendices
Appendix A: Proposed time frame for the project.
Appendix B: Questionnaire (s)
107
Appendix C: Interview guide (s)
7.0 Summary
In this unit you have been informed that a writing of a research report is the last step
in the research process. The report informs your readers what you have done, what you have
discovered and what conclusions you have drawn from your findings. You cannot write a valid
research report if you did not write the research proposal. Besides having all the sections
contained in the proposal, the report will have findings, discussions and conclusions. The
research report for PAS 2014 should have four chapters. Chapter one is the introduction, chapter
two is methodology, chapter three is findings and discussion while chapter four is the last chapter
and it consists of conclusions and recommendations
8.0 Self Assessment Question
After reading through unit eleven, you should write your PAS 2014 research report in
line with the guidance provided in this unit.
108
REFERENCES
Bless C. and Kathuria R. (1993) Fundamentals of Social Statistics: An African Perspective,

Cape Town: Juta and Co. Ltd
Bowen, B. and Herbert, F. (1980) An Introduction to Data analysis, San Francisco: W. H.

Freeman and Co. Ltd
Kenny D. (1987) Statistics for the Social and Behavioural Sciences, Canada: Little Brown and
Co. Ltd
Kotari C. (2004) Research Methodology: Methods and Techniques, 2nd Ed. New Delhi: New
Age International Publishers
Kumar, R. (2005), Research Methodology: A Step – by – Step for Beginners, 2nd Ed. London:
Sage Publications
Methods, New Delhi: PHI Learning Private Limited
Nachmias, C. and Nachmias, D. (1996) Research Methods in the Social Sciences, 5th Ed.
O’Sullivan, E. and Rassel, G. (1995) Research Methods for Public Administration, 2nd Ed.
New York: Longman
Upgrade, V. and Shende, A. (2000) Research Methodology, New Delhi: S. Chand and Co.
109
110
111
112
113

GMS 2014 Module 2

Uploaded by

Copyright:

Available Formats

GMS 2014 Module 2

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GMS 2014 Module 2

Uploaded by

Copyright:

Available Formats

INTRODUCTION

AIM OF THE MODULE

The aim of this module is to introduce you to statistical methods in research.

OBJECTIVES OF THE MODULE

By the end of this module, you should be able to:

Explain the major statistics concepts used in research.

2.0 Aim of the Unit

The aim of this unit is to introduce you to statistics.

3.0 Objectives of the Unit

By the end of this unit, you should be able to:

Explain the meaning of statistics.

4.0 Required Material

In this unit, you will require the following readings:

Bless C. and Kathuria R. (1993) Fundamentals of Social Statistics: An African

6.0 Main Contents

6.2 Importance of Statistics

 63% of people in Africa live in rural areas;

Knowledge of statistics is important in solving practical problem as it enables one to know

6.3 Descriptive Statistics

Frequency distributions help us to organize data by showing us the frequency (regularity)

Table 1.1: Level of satisfaction as indicated by 20 civil servants

Interviewee No. Level of job satisfaction

Level of job satisfaction Frequency Interviewee No.

1. Ungrouped Frequency Distribution

Table 1.3: Responses to the question on age by 25 distance students

Age f Age f Age f

Age (Class – interval) Frequency

3. Absolute Frequency Distributions

Level of job satisfaction Frequency

4. Absolute Frequency Distributions

Relative frequency distributions are frequency distributions representing the percentage

Table 1.7: Relative Frequency Distribution

Level of job satisfaction F F%

5. Cumulative Frequency Distribution

Table 1.6 Cumulative Frequency Distribution

Age (Class – interval) F CF CF %

6. Decumulative Frequency Distribution

Age (Class – interval) F DCF DCF %

6.5 Graphical Presentation of Data

City No. of Street Kids

Since 1000 represents 360º,

The number of street kids in Lusaka is equivalent to 450/1000 x 360 = 162º

The number of street kids in Kitwe is equivalent to 300/1000 x 360º = 108º

The number of street kids in Ndola is equivalent to 150/1000 x 360 = 54º

The number of street kinds in Livingstone is equivalent to 100/1000 x 360 = 36º

Response Number of students

3. Multiple Bar Charts

600 School of Humanities and

Maize and Cotton Production

8.0 Self Assessment Questions

Complete the frequency table by including the following;

Distinguish between the following;

2.0 Aim of the Unit

3.0 Objectives of the Unit

By the end of this unit, you should be able to:

Interpret the measures of central tendency

4.0 Required Material

In this unit, you will require the following readings:

Bless C. and Kathuria R. (1993) Fundamentals of Social Statistics: An African