1. Raising a new Generation of Leaders
www.covenantuniversity.edu.ng
College: Business & Social Sciences Dept: Mass Communication
Programme: Mass Communication
Course Title: Analysis in Communication Research
Code: MAC411 (2 Units)
SESSION: 2017/2018 SEMESTER: ALPHA
2. 2
Course Compact
COVENANT UNIVERSITY
College: Business & Social Sciences Dept: Mass Communication
Course Title: Data Analysis in Communication Research Code: MAC411 2 Units
Semester: Alpha Time: 5-7pm (Mondays). Location: C401
Course Lecturers
2
3. 3
This course builds on the knowledge gained in MAC230, MAC220 (Introduction to
Statistics for Social Sciences 1 & 11) and MAC223 (Foundation of Communication
Research).
The course provides a broad overview of statistical methods for advanced
undergraduates. The course provides applied experience with conducting, interpreting,
and reporting results of data collected.
The course adopts research studies and examples that connect the statistical concepts
to data analysis problems.
It also exposes students to statistical techniques, data processing techniques
Brief Overview of Course
3
4. 4
OBJECTIVES:
To enable students have in-depth understanding of the application of basic methods of statistics in
mass communication.
The course provides applied experience with conducting, interpreting, and reporting
results of data collected
C). METHOD OF LECTURE DELIVERY/TEACHING AIDS: The teaching methods include
class lectures, individual and group assignments, tests and examinations. Students will be
divided into groups to discuss and examine the relevant issues of interest concerning
statistics and the broadcast industry
4
5. 5
(E). Tutorials
Two hours will be devoted to tutorial sessions per week. Such periods will be used for practical
demonstration as well as practice the use of computer in data entry, processing, analysis,
interpretation and presentation.
(F). Method of Grading
(A) Mid-Semester Exams, Enjoyments and Other Class Tests = 30%
(B) Alpha Semester Exams = 70% TOTAL = 100%
(G). Ground Rules and Regulations
The University guidelines on attendance requirements will be strictly observed in this course.
No latecomer will be allowed into the lecture hall 10 minutes LATE –that is 10 minutes into the
lecture without an acceptable excuse.
Assignments (Enjoyments) submitted after the due and terminal dates will not be accepted.
All assignments (Enjoyments) undertaken in this course are graded.
Two hours will be devoted to tutorials every week.
Improper and Indecent dressing will not be allowed into the lecture hall
5
6. 6
(H). Topics for Term Papers/ Assignments/Student Activities
Each topic with a calculation will be followed with related problems. Some will be solved in the class
while others will be taken home and returned at agreed dates.
Students will be expected to come to class with their laptops for practical exercises and demonstration.
The mid-semester exam will be conducted electronically.
(I). Alignment with Covenant University Vision/Goals
This course will assist students become leaders and experts in communication research and other
scholarly endeavors that are research based.
(J). Contemporary Issues/Industry Relevance
The course is intended to prepare students to solve problems encountered in research projects, to
make decisions based on data in general settings both within and beyond the university setting, and
finally to become critical readers of statistical analyses in research papers and in news reports.
6
7. 7
WEEK 1: Collecting Data
Introduction and Abstract of a Research Study
Data Collection Tools: surveys (Questionnaire and Interview),
Questionnaire: Face-to-face interviews, self-administered
questionnaires,
Interview: In-depth Interview, Focus Group Discussion
Observational Studies
Week 2 & 3: Sampling Designs for Surveys
Probability sampling: Simple Random, Stratified Sampling, Cluster
Sampling, Systematic Sampling and Multistage Sampling.
Non-probability sampling: Quota sampling, Convenience sampling,
Purposive sampling, Self-selection sampling and Snowball sampling.
Course Content
7
8. 8
Week 4, 5 & 6: Methods of Analysis
Quantitative data analysis
Univariate (single variable): mean, median, standard deviation, and so on
Bivariate analysis (two variables):
Correlation, coefficient of determination, Chi-Square, contingency coefficient,
Graphs and Simple Regression analysis.
Multivariate analyses (multiple variables):
Regression techniques
Analysis of variance (ANOVA)
multivariate analysis of variance (MANOVA)
Course Content (contd.)
8
9. 9
Week 4, 5 & 6: Methods of Analysis (contd)
Qualitative data analysis:
Content Analysis,
Grounded Analysis (theory),
Social Network Analysis,
Discourse Analysis,
Narrative Analysis and,
Conversation Analysis.
Research Study: Exit Polls versus Election Results
Course Content (contd.)
9
10. 10
Week 7 -9: Probability and Probability Distributions
Introduction and Abstract of Research Study
Finding the Probability of an Event
Basic Event Relations and Probability Laws
Conditional probability
Dependent events
Probability Distributions for Discrete Random Variables Two Discrete
Random Variables
Probability Distributions for Continuous Random Variables
Random Sampling
Sampling Distributions
Research Study: Inferences about Performance-Enhancing Drugs among
Athletes
Summary and Key Formulas
Course Content (contd.)
10
11. 11
Week 10 & 11: Inferences about Population Central Values
Statistical estimation
Hypotheses testing
Point and interval estimation
Level of confidence
Exercises
Week 12: Revision
Course Content (contd.)
11
12. 12
K) Recommended Books:
1.Spiegel Murray R. & Stephens Larry J. (2008). Theory and Problems of Statistics. Schaum’s
Outlines. Fourth Edition. TATA McGRAW-HILL Edition
2.Agbadudu A.B.(1994). Statistics for Business and Social Sciences. Revised Edition. URI
3. Ott, L &Longnecker, M.(2010). An Introduction to Statistical methods and Data Analysis. Belmont:
Centage Learning
4.Wilmer and Dominick (2003). Mass Media Research. An Introduction. Belmont, CA:
Thomson/Wadsworth
5.Obikeze, D.O. (1990). Methods of Data Analysis in the Social and Behavioural Sciences. Enugu:
Auto-Century Publishing Company Ltd.
Other references:
1.https://www.cfsecure.com/stats.cyberk/instructor.cfm?ft-s5
2.http://icp.giss.nasa.gov/education/statistics/index.html?
3.http://www.shodor.org/interactivate/lessons/sm1.html
4.http://www.mste.uiuc.edu/hill/dstat/dstat.html
Course Content
12
13. 13
Dr. E.O Amoo 13
Distinction between Statistic and Data
Statistics as information
It entails processed data, as a group of numbers that represent facts or describing a situation.
A statistic is a single figure that summarizes a situation or event.
Statistics are obtained from data.
What is data?
Data are information or facts used in making calculations.
They are figures to be further processed or such that computer can use.
In mass communication, data are the raw materials out of which social and economic statistics are
produced.
They are information collected and stored at the level at which the unit of analysis could be observed.
Have you gathered your data?
Review
Module I
Introduction
14. 14
Statistics as the art and science of collecting, organizing, analyzing,
summarizing and presenting and interpreting data (COASPI) so that valid
conclusions and reasonable decisions could be made about an event or
phenomenon.
Social Statistics: described as the use of statistical measurement systems to
study human behavior in a social environment through the information obtained
from the sub-set of the population or the whole population
Students to differentiate between STATISTICS & STATISTIC
Do you still remember the following?
14
Statistics, Types and Data Descriptions
15. 15
Statistics is divided into two main branches, namely: Deductive and Inductive statistics or
simply put “Descriptive and Inferential Statistics”.
Its scope therefore covers two major areas, namely:
Descriptive (or Deductive) statistics and
Inductive (or Inferential) statistics.
SCOPE of STATISTICS
What are the scope/branches of Statistics?
16. 16
Descriptive statistics is concerned with collection and description of a set of data for meaningful
understanding.
Descriptive statistics can be described as a straightforward presentation of facts, in which modeling
decisions made by a data analyst have had minimal influence.
It draws no conclusion concerning the entire set.
Specific tools of descriptive statistics are tables, graphs, charts, rates, etc.
Examples include: company balance sheet, tables of football leagues, etc
Students to give more examples
Descriptive vs. Inferential
17. 17
Inferential statistics deals with analysis of a part of data leading to predictions or inferences about
the entire set.
Simply put, it is used to make predictions or comparisons about a larger group (called population)
on the basis of information gathered from the part of the group (i.e. the sample).
Inferential statistics deals with drawing conclusions, generalizations, making some predictions or
estimations on the basis of data obtained from samples.
Examples are: Paired T-test, Independent T-test, ANOVA, Correlational, Pearson correlation, Spearman correlation, Chi-square,
Simple and multiple regression
Descriptive vs. Inferential (contd.)
18. 18
Types of Data
Numerical data
Categorical data (e.g. labels i.e. words. Examples include a list of products bought at a grocery store {milk, eggs,
toilet paper, bread, and so on)
Types of Numerical Data
Discrete Data
Continuous data
(a) Examples of Discrete data
Discrete data are whole numbers. They are data from counting. They are whole integers and do not
take fraction. A count of number of animals in a house gives discrete data.
(b) Examples of Continuous Data
A continuous data implies a data whose set of assumed values is uncountable, arising from
measurement, e.g. 2.5, 3.123, 0.92, etc. They are measured data and may take on any real value.
E.g. Amount of time spent watching TV (2hr, 30 mins) ; Distance between point A & B (2½ km).
Introduction to Data Description and Data Collection
18
19. 19
Primary Sources
This refers to data collected either by or under the direct supervision and instruction of the investigator. Such
data usually implies considerable knowledge of the conditions under which the data are collected and so of the
limitations which must be placed on their use
The investigator is familiar with the background of data
Secondary Sources
This refers to all other data where a lesser degree of control by the investigator exists.
There is possibility of misinterpretation as the degree of control decreases.
List of Methods
Censuses
Sample surveys,
Personal interviews,
Self-administered questionnaires,
Postal and telephone enquiries.
Social Media (Students to list examples)
List of Sources of Data
19
20. 20
Highlights on Methods of Data Collection
20
Census
A census is a study that obtains data from every member/element of a population. In most studies, a
census is not practical, because of the cost and/or time required.
Sample survey
A sample survey is a study that obtains data from a subset of a population, in order to estimate
population attributes.
Experiment
An experiment is a controlled study in which the researcher attempts to understand cause-and-effect
relationships. The study is "controlled" in the sense that the researcher controls (1) how subjects are
assigned to groups and (2) which treatments each group receives.
Observational study
Observational studies attempt to understand cause-and-effect relationships but the researcher is not
able to control the event(s).
21. 21
Types of sampling techniques
Probability sampling methods
Non-probability sampling methods
Probability sampling methods
In this method, each element of the population has a known chance of being selected or included in the sample.
It allows the calculation of sampling errors
It permits the evaluation of goodness of fit of the estimates
Goodness of fit
• Goodness of fit describes how well the result obtained fits a set of observations.
• It is measures by evaluating the discrepancy between observed values & the values expected under a model or theory.
This features prominently in statistical hypothesis testing. Examples include:
• Test for normality of residuals,
• Testing whether two samples are drawn from identical distributions
• Whether outcome frequencies follow a specified distribution (e.g. Pearson's chi-squared test)
• Likely test available:
(A) Kolmogorov–Smirnov test, Anderson–Darling test, Shapiro-Wilk test, Chi-squared test, Hosmer–Lemeshow test, etc
(B) Coefficient of determination (R-squared), Reduced chi-squared.
Sampling Designs for Surveys (contd.)
21
22. 22
Types of probability sampling
Simple random Sample
All elements has equal chance of being selected
Steps
Assign numbers to all the elements in the population.
Record each number on a slip of paper
Draw ‘n’ slips (or specific desired number of) slips of paper
The selected numbers forms your sample
Caution:
Difficult to use when population is larger
When encounter a very large population, a table of random numbers can be used.
22
Probability sampling methods (contd.)
23. 23
Systematic Sampling
It is a modification of simple random sample
Less time consuming
An estimated number of elements in a larger population is divided into desired sample size resulting into
Sampling Interval
Sample is then drawn by listing the population elements in a specific order
Then, every nth case is selecting starting with a randomly selected number between 1 and n
Exercise
Given the total number of health centres in a city is 1995, show the procedure to select a sample size of 285 health
centres from the population
Probability sampling methods (contd.)
23
24. 24
Systematic sampling
24
Systematic sampling selection approach
Total Population 500
Desired/Required sample size 25
Sampling Interval 500/25 = 20
You may choose a random number of where to start (e.g. 5).
So, you start from 5th house 5th
Next house = 5+20 = 25th
Next health center = 25th + 20 = 45th
--- ---
Continue until quota completed ( at No 485)
Example
Demonstrate how you can select a representative sample of 25 houses from a community
with 3500 houses using a Systematic Sampling technique.
25. 25
Systematic sampling
25
Exercise
Conduct a systematic sampling procedure to select 285 students from a total population of
1995 student in Mass Communication Department.
Systematic Sampling
Total Population 1995
Desired/Required sample size 285
Sampling Interval 1995/285 = 7
You may choose a random number of where
to start (e.g. 3). So, you start from 3rd student 3rd
Next Student = 3+7= 10th
Next Student = 10+7 = 17th
--- ---
Continue until quota completed
26. 26
Systematic sampling
26
Exercise
Conduct a systematic sampling procedure to select 285 students from a total population of
1995 student in Mass Communication Department.
Systematic Sampling
Total Population 1995
Desired/Required sample size 285
Sampling Interval 1995/285 = 7
You may choose a random number of where
to start (e.g. 3). So, you start from 3rd student 3rd
Next Student = 3+7= 10th
Next Student = 10+7 = 17th
--- ---
Continue until quota completed
27. 27
Stratified Sampling
It is a modification of simple random sample
Population is divided into (seemingly ) homogeneous groups) called strata
Each strata is then sampled independently
It is used to endure that all relevant strata are represented in the sample.
It may used in conjunction with simple random or systematic random sampling
In other words,
Population is divided into strata and a random sample is taken from the elements in each stratum
NB: Use where population differ by certain characteristics
Sample can be selected via the following ways:
Proportional Allocation
Cost approach allocation
Optimum Allocation
Probability sampling methods (contd.)
27
28. 28
Proportional Allocation
Sample size in each stratum could be made proportionate or disproportionate to the number
of elements in the stratum.
E.g. if age distribution of sample is the same with age distribution of the population, you can
draw proportionately to the population.
Where ,
N = Population size ,
n = Sample size and,
Ni = Size of the stratum
Stratified Sampling (contd.)
28
29. 29
Cost Approach Allocation
The sample size in each stratum is inversely proportion to the cost of sampling the stratum
N = n(Ni)Ci divided by (∑Ni/Ci)
Simply presented as:
Where,
Ci = unit cost of sampling from ith stratum
n = total sample size
N = population size
Stratified sampling (contd.)
29
30. 30
Optimum Allocation
The sample size in each stratum is proportional to the product of the number of elements
(in the population) in the stratum and the standard deviation of the characteristics being
measured in the stratum.
Stratified sampling (contd.)
30
32. 32
Most commonly used
Samples are selected in two or more stages
Used when there is no adequate sampling frame
Used also when sample random sampling might be too costly or difficult to reach (when too
dispersed)
It increases sample errors
Select a random sample of household
Divide population into groups
Select a random sample of elements from within the cluster/group
Cluster Sampling
32
33. 33
Used when population is extremely complex
Select the primary sampling unit (PSU)
Draw a random selection of household to visit
Randomly select an individual from the household or interview all eligible individuals
Possible to apply a sampling fraction to each group e.g. one-in-five households or one-in-
ten eligible women
33
Multistage Sampling
34. 34
Definition
The process of selecting a sample from a population without using (statistical)
probability theory.
Characteristics of non-probability sampling
Each element of the population DOES NOT have equal chance of being included in the sample
The researcher CANNOT estimate the error caused by not collecting data from all
elements/members of the population.
Non-Probability Sampling
36. 36
Definition
Convenient sampling is the selection of easily accessible participants with no
randomization.
For example, asking people who live in your room or floor to fill your questionnaire.
Interviewing just anyone you see on the road
Asking just anyone you see on the road to comment on agenda of government, or economic situation
Class discussion
Students to cite examples of where convenient sampling could be appropriate
Convenient (Non-Probability) Sampling
37. 37
Definition
Selecting participant in numbers proportionate to their numbers in the larger
population, no randomization.
For example, you may arbitrarily include only 5 males and 15 female as sample
in a particular study of students response to 75% attendance as qualification for
writing exam in Nigerian universities.
Class discussion
Students to cite examples of where only or where Quota system is practicable
Quota Sampling
38. 38
Definition
Selecting participants because they have certain predetermined
characteristics, no randomization.
Decision to include a particular tribe in a sample.
Attempt to select only first class students as reps for CBSS
Judgmental Sampling
39. 39
Definition
In sociology and statistics, snowball sampling is a non-probability sampling technique
also known as ‘chain sampling’, ‘chain-referral sampling’, ‘referral sampling’, where
existing study subjects recruit future subjects from among their acquaintances.
Selecting participants by finding one or two participants and then asking them to refer
you to others.
Thus, the sample group grows like a rolling snowball.
E.g.,
Selection of students who are keeping phones in their rooms.
Selection of individuals with a particular brand of shoe/car, etc
Snowball Sampling
41. 41
Israel Gleen Approach
Where
‘Z’ is z-score for the confidence interval selected.
‘d’ is the upper limit of the range of occurrence of the event
‘c’ the lower limit of the range of occurrence of the event and,
‘n is the sample size
41
Determination of Sample Size
2
42. 42
In a binary/dichotomous variable, the d should imply the “present” or “absence of the event, mostly captured as
1 (one) or 0 (zero).
d = 1 (implying that the challenge is present)
c = 0 (representing that the challenge is not there). This serves as the range for the estimation of the sample size
for this study.
Maximum margin of error is 0.05 indicating a 95% level of significance.
Thus ,the sample size for this study is:
= [1.96 (1-0)/(2*0.05)]2
= [1.96 / 0.1] 2 = 384.
42
Determination of Sample Size (Contd.)
43. 43
Taro Yamane’s Approach
43
Determination of Sample Size (contd)
This implies n = N[Z (d-c)/(2e)]2 / [N – 1 + (Z (d-c)/(2e))2]
Where,
N is the population from where the sample is meant to be selected
‘Z’ is z-score for the confidence interval selected.
‘d’ is the upper limit of the range of occurrence of the event
‘c’ the lower limit of the range of occurrence of the event and,
‘n is the sample size
44. 44
Given the population of Osun State as 3,423,535 distributed into 1,677,532 and 1,746,003, male and female
respectively, The estimate sample size will be:
n = 1,677,532 [1.96 *( 1 - 0)/(2 * 0.05)]2 / [1,677,532 - 1 + (1.96(1 - 0)/(2 * 0.05))2]
or
1,677,532 [1.96 *( 1 - 0)/(2 * 0.05)]2
[1,677,532 - 1 + (1.96(1 - 0)/(2 * 0.05))2]
= 384
44
Determination of Sample Size (contd)
45. 45
Exercise 1
Q1. We’ve just started a new educational TV Advert that teaches viewers all about research methods. We know from
past educational TV programs that such an advert would likely capture 2 out of 10 viewers on a typical night. If we
want to be 99% confident that our obtained sample proportion of viewers will differ from the true population
proportions by not more than 5%, what sample size do we need?
Solution (Q1)
Where there is information on the Mean:
45
Where,
n = The require Sample Size
P = Prob. of success
q = Prob. Of failure
E = Error margin i.e. how much error you are willing to accept (5% = 0.05)
Z = Value of Standard Normal Distribution
Z at 99% confidence = 2.58
Sample Size Without Sampling Approach
46. 46
Exercise 2
Q2. Also, if we wish to determine the required sample size with 95% confidence and 5% error tolerance
where the 40% of population preferred the TV Advert, what would be required sample size?
Solution (Q2)
Where there is information on the Mean:
46
Where,
n = The require Sample Size
PC = Percentage of Population (interested in)
E = Error margin i.e. how much error you are willing to accept (5% = 0.05)
Z = Value of Standard Normal Distribution
Z at 95% confidence = 1.96
Sample Size Without Sampling Approach (Contd.)
48. 48
• If you desire a C% confidence interval for a population proportion p with an
accuracy specified by you, how large does the sample size need to be?
• We will denote the accuracy by Margin of Error (ME).
Determining Sample Size to Estimate p
To Estimate a Population Proportion p
49. 49
*
*
*
ˆ
ˆ
2
2
p ME
pq
CI for p : p z
n
pq
set z ME and solve for n :
n
z pq
n = ;
(ME)
Required Sample Size n to Estimate a Population Proportion p
51. 51
What About p and q=1-p?
*2
2
z pq
n = ; we don't know p or q;
(ME)
TWO METHODS :
1: if prior information is available concerning
the value of p, use that value of p to calculate
n;
2 : if no prior information about p is available,
to o
btain a conservative estimate of the
1
required sample size, use p q
2
52. 52
Example: Sample Size to Estimate a Population
Proportion p
• The U. S. Crime Commission wants to estimate p = th
e proportion of crimes in which firearms are used to w
ithin .02 with 90% confidence. Data from previous ye
ars shows that p is about .6
53. 53
Example: Sample Size to Estimate a Population
Proportion p (cont.)
*
*
2
2
2
2
z pq
n = ; ME .02; p is estimated to
(ME)
be about .6 from previous years' data;
90% z 1.645
(1.645) (.6)(.4)
n 1,623.6; n 1624
(.02)
54. 54
Example: Sample Size to Estimate a Population
Proportion p
The Curdle Dairy Co. wants to estimate the
proportion p of customers that will purchase its
new broccoli-flavored ice cream.
Curdle wants to be 90% confident that they have
estimated p to within .03. How many
customers should they sample?
55. Example: Sample Size to Estimate a Population
Proportion p (cont.)
• The desired Margin of Error is ME = .03
• Curdle wants to be 90% confident, so z*=1.645; the required sa
mple size is
• Since the sample has not yet been taken, the sample proportion
p is still unknown.
• We proceed using either one of the following two methods:
*2
2
z pq
n =
(ME)
2
2
(1.645) pq
n
(.03)
56. Example: Sample Size to Estimate a Population
Proportion p (cont.)
• Method 1:
– There is no knowledge about the value of p
– Let p = .5. This results in the largest possible n needed for a 90% confidence inter
val of the form
– If the proportion does not equal .5, the actual ME will be narrower than .03 with t
he n obtained by the formula below.
03
.
p̂
2
2
1.645 .5 .5
751.67 752
.03
n
2
2
1.645 .2 .8
481.07 482
.03
n
• Method 2:
– There is some idea about the value of p (say p ~ .2)
– Use the value of p to calculate the sample size
*2
2
z pq
n =
(ME)
57. 57
l 57
Types of Survey Designs
• Cross-sectional longitudinal
• Longitudinal
trend
cohort
panel
58. l 58
Types of Cross-sectional Longitudinal Survey Designs
Time of Data Collection
Study Over Time Study at One Point in Time
Longitudinal Cross-sectional
Changes
in a
sub-population
group identified
by a common
characteristic
over time
Changes
in the
same
people
over time
Trends in
the same
population
over time
Attitudes
and
Practices
Community
Needs
Program
Evaluation
Trend Cohort Panel
Group
Comparisons
National
Assessment
59. 59
l 59
Key Characteristics of Survey Designs
• Sampling from a population
• Collecting data through questionnaires or interviews
• Designing instruments for data collection
• Obtaining a high response rate
• Designing and using a mailed questionnaire
• Conducting an interview survey
60. 60
l 60
Collecting Data Through Questionnaires
• Mailed questionnaires
• Electronic questionnaires
• One-on-one interviews
• Focus group interviews
• Telephone interviews
61. 61
l 61
Designing Instruments For Data Collection: Types of Questions
• personal
• attitudinal
• behavioral
• sensitive
• scale
• open-ended