12-Jan-25

Methods of data collection

Research Methodology The choice of methods of data collection is based on:
 The accuracy of information they will yield
LM 238  Practical considerations, such as, the need for personnel,
ES 491 time, equipment and other facilities, in relation to what
is available.
 Some tools for data collection;  Accuracy and “practicability” are
1. Questionnaires often inversely correlated.
2. Interviews Therefore, accuracy must be
INSTRUCTOR(S) 3. Schedules balanced against practical
• Dr. J. Tibaijuka 4. Observation Techniques considerations (resources and
• Dr. Taifa 5. Rating Scales other practical limitations)
6. Experimental methods

Questionnaire design
• A questionnaire is a formalized set of questions for
obtaining information from respondents.
 A device for securing answers to questions by using a
form which the respondent fills in himself.
 Normally used where one cannot see personally all
the people from whom he desires responses or where
there is no particular reason to see them personally.
 Can be used to collect:
 Beliefs: what is right or wrong?
 Attitude: what they want?
 Behavior: what do they do?
 Characteristics of users

Questionnaire Objectives
Characteristics of a Good Questionnaire:
 It must translate the information needed into a set 1. It seeks only that data which cannot be obtained from the
of specific questions that the respondents can and resources like books, reports and records.
will answer. 2. It is as short as possible, only long enough to get the essential
data.
 A questionnaire must uplift, motivate, and
3. It is attractive in appearance, nearly arranged and clearly
encourage the respondent to become involved in
duplicated or printed.
the interview, to cooperate, and to complete the 4. Directions are clear and complete, important terms are clarified.
interview. 5. The questions are objective, with no clues, hints or suggestions.
 A questionnaire should minimize response error. 6. Questions are presented in a order from simple to complex.
7. Double negatives, adverbs and descriptive adjectives are
avoided.
8. Double barreled questions or putting two questions in one
question are also avoided.
9. The questions carry adequate number of alternatives.
5

1
 12-Jan-25

Merits of Questionnaire Method:

1. it’s very economical.
2. It’s a time saving process.
3. It covers the research in wide area.
4. It’s very suitable for special type of responses.
5. It is most reliable in special cases.

Demerits of Questionnaire Method:

1. Through this we get only limited responses.
2. Lack of personal contact.
3. Greater possibility of wrong answers.
4. Chances of receiving incomplete response are more.
5. Sometimes answers may be illegible.
6. It may be useless in many problems.

Questionnaire Design
Questions may take two general forms:
 “Open ended” questions, which the subject answers in
his own words, or
 “closed” questions, which are answered by choosing
from a number of fixed alternative responses.

2
 12-Jan-25

Types of Question (Structured or closed form)

DICHOTOMOUS YES/NO

MULTIPLE ONE/MANY
CHOICE RESPONSES

ODD/EVEN
RATING SCALES
NUMBERS

OPEN-ENDED FREE RESPONSE

Types of Question (Structured or closed form)

Types of Questions
1. Pre-coded, single-choice questions
RANKING 1st/2nd ETC. 2. Open-ended questions
3. Presence-absence questions
MANY
RATIO DATA 4. Rank-ordering questions
RESPONSES
5. Likert-type questions
DISTRIBUTING
6. Index development
CONSTANT SUM
MARKS

MARKS OUT
RATIO DATA
OF TEN
16

Pre-Coded, Single-Choice Open-Ended Questions

• Pre-coded, single-choice questions ask respondents to • Open-ended questions do not have pre-set answers.
indicate which one category applies Excellent way to explore new areas
• Answers are pre-coded for easy data entry
• Use open-ended questions when:
• If not all options can be listed, include a category • Too many response categories (year of birth)
entitled “Other” with a space to indicate what the • You don’t wish to impose categories on respondents
“Other” category implies • “Really” consulting respondents
• “Please specify” ______________ • Qualitative – source of quotations
• Categories should be mutually-exclusive • Determining appropriate categorization
• i.e., no categories overlap with one another • You want a change in pace for respondents
E.g. What year are you in?
Freshman 1 E.g. What is the one thing you would like to see changed at the
Sophomore 2 university?...................................................................
Junior 3 Limitations: Are time-consuming to code, May generate responses
Senior 4 14-17 17
that are inconsistent and Are more likely to be left blank 18

3
 12-Jan-25

Presence-Absence Questions Rank-Ordering Questions

• Presence-absence questions ask respondents to check
off which items in a list do or do not apply to them • Rank-order questions ask respondents to indicate an
• Less commonly used than other types of questions ordering of response items, usually from most
E.g. Have you ever had contact with a physically handicapped preferred to least preferred
person in any of these groups? (Circle to indicate “yes” or “no” • Must be done with great care
for each group) • Ask for only three most important items
Yes No • Must make instructions explicit
Community ----------- 1 0
Family ----------------- 1 0
• These types of questions should be avoided or
Relatives -------------- 1 0
minimized because they take time
Elementary school -- 1 0
High school ----------- 1 0
University class ------ 1 0
As a co-worker ------- 1 0
19 20

Open-ended Questions
Types of Questions in a Questionnaire  Questions that allow the respondent to answer in
1. Open-ended Questions their own words.
2. Close-ended Questions  Preferred in gathering qualitative data and provides
3. Focused Questions
richer, more detailed responses.
4. Catch-all Questions Examples:
5. Hypothetical Questions • What do you like most about our product?
6. Leading Questions
• How would you describe your experience with
customer service?

Close-ended Questions Focused Questions

 Questions with predefined answer options.  Questions that are specifically designed to focus the
Examples: respondent on a particular issue or topic.
 Do you own a smartphone? (Yes/No)  They are used To narrow the scope of responses and
 How satisfied are you with the service? (Very to ensure that the respondent’s answer is relevant to
Satisfied, Satisfied, Neutral, Dissatisfied, Very the research objectives.
Dissatisfied)"  Examples:
 Preferred in gathering quantitative data and to
o How often do you use social media for product
simplify statistical analysis.
reviews?
o What factors influence your decision when
purchasing electronics?

4
 12-Jan-25

Catch-all Questions Hypothetical Questions

 Questions that attempt to cover a broad range of  Questions that ask respondents to consider a situation that hasn’t
possible responses or opinions in a single query. occurred, typically using "what if" scenarios.
 Used to capture a wide variety of responses on a general Purpose:
topic.  To assess how respondents might behave or feel in a given
scenario.
Examples:
 Useful for predicting future actions or decision-making.
 What else would you like to share about your
Examples:
experience with our service?
 If you were offered a 10% discount on our product, would you
 Is there anything we can improve? purchase it?
 What would you do if your favorite brand raised its prices?

Leading Questions
Questionnaire – steps in designing
 Questions that suggest or imply a particular answer,
often influencing the respondent's response. Used to
guide the respondent toward a desired response. Physical
Problem Sequence characteristics
Examples:
 How much do you love the convenience of our
Target
product? respondents Wordings Approval
 Don’t you agree that our customer service is
excellent?
Number Pilot Final
of Q Type of Q study draft

28

Construction of Questionnaires Processing of Questionnaires

Steps for Effective Questionnaire Design:
 Define the objectives: Identify the purpose of the survey and
what data is needed.
 Select question types: Decide on the balance between open-
ended and close-ended questions.
 Sequence the questions: Start with simple or neutral questions,
and save complex or sensitive ones for later.
 Use clear language: Avoid jargon, ambiguous terms, or complex
sentence structures.
 Pre-test the questionnaire: Run a small pilot test to check for
clarity and effectiveness.
Tips:
 Keep the questionnaire concise.
 Avoid leading, double-barreled, or biased questions.
 Use scales or rating systems for consistency (e.g., Likert scales).

5
 12-Jan-25

Data Processing…

them
- Stages
o Types
• Field editing
• In house or central editing

Data Processing…

o Types
• Numerical
• Alphabetical
• Alpha-numerical

6
 12-Jan-25

Data Processing… Data Processing…

- contain one characteristic

- contain two characteristics
- contain more than two
characteristics

o Guidelines
• Table name and number, Headings – rows &
columns
• Unit of measurement
• Source of data and Explanatory note

7
 12-Jan-25

Graphic Presentation
• Data can be presented effectively for quick and clear
understanding in any forms of graphs.
• Pie chart
• Bar graph
• Line graph
• Histograms Pie chart
Sales
• Methods 8%
• Manual 12% 1st Qtr
60%
• Machine 2nd Qtr
20% 3rd Qtr
4th Qtr

44

Graphic Presentation… Graphic Presentation…

Bar graph Line graph Histogram

14 14
Series 3 Series 3
12 12
Series 2 Series 2
10 10
Series 1 Series 1
8 8
6 6
4 4
2 2
0 0

45 46

Data Processing - Significance DATA ANALYSIS

• Editing • Tabulation Why do we analyze data?
• Accuracy • Analysis & Interpretation  The purpose of analysing data is to obtain usable and
• Uniformity • Writing research report
• Complete • Correlation between useful information. The analysis, irrespective of
• coding variables whether the data is qualitative or quantitative, may:
• Detect error
• Coding • Easy understanding of data • Describe and summarise the data
• Classification • Find location of data • Identify relationships between variables
• Tabulation • Support written matter • Compare variables
• Classification • Graphic presentation • Identify the difference between variables
• Tabulation • Quick communication
• Speedy search of data • Effective appeal • Forecast outcomes
• Grouping • Condense large volume of
data
• Educative value
47

8
 12-Jan-25

SCALES OF MEASUREMENT
Ordinal scale
 The type of analysis to use on a given set of data and the  The data can be classified into non-numerical or named
relevant forms of data presentation or display is based on the categories and inherent order exists among the response
scale of measurement of the data. categories.
 These scales includes nominal, ordinal and numerical.  Data are compared and contrasted to determine those that are
greater than(>) compared to less than (<) within the data set
Nominal scale but not equal to “=“
 Nominal scale is simply a system of assigning number symbols to  This scale have only a rank meaning and the appropriate
events in order to label them and assigning them to categories. measure of central tendency is the median while a percentile or
 Nominal scales provide convenient ways of keeping track of quartile measure is used for measuring dispersion.
people, objects and events and have no quantitative value. Numerical scale
 For example,  The numbers represent the possible response categories.
• Asign “0” to represent a male and if female assign to “1”. • Interval Scale
• Also, you can record “Yes or No” answers to a question as “0” • Ratio Scale
and “1”  Mean is the appropriate measure of central tendency, while standard
• Numbers of players jersey deviation is the most widely used measure of dispersion.

Concepts in data analysis Statistical Analysis – Tools and Techniques

Statistical Analysis – Tools and Techniques  Measures of Central Tendency (statistical averages)
• The most commonly investigated characteristic of a
• Statistical terms collection of data (or data set) is its center, or the point
• Ratio around which the observations tend to cluster.
• Proportion • The single value which is a representative of all values is
• Percentage called an average or a measure of central tendency.
• Rate  Objective
• Mean
• Arrive at single value
• Median
• Get a bird eye view
• Mode
• Standard deviation • Compare
• Trend • Precise relationship
• Know about universe from sample

51 52

Arithmetic Mean
Measures of Central Tendency  Value is obtained by dividing the sum by total
number Individual distribution

Measures
A.M = sum of values of all items
--------------------------------
Number of items

Mathematical Positional

Arithmetic Geometric Harmonic

Median Mode
mean mean mean

 The mean is sensitive to extreme values, i.e skewed distributions.

53 54

9
 12-Jan-25

Median Mode
 Is the middle value of a series when the data of a series  The value of a variable which occurs most frequently
is arranged in ascending or descending order. It divides
 Mode is particularly useful in the study of popular
the series in two parts
sizes. E.g. The most preferred size of shoes.
 Mode is a positional average and is not affected by the
values of extreme items

 Median is a positional average and is used only in the context of

qualitative phenomena
 The median is not as sensitive to extreme values as the mean
55 56

Measures of Dispersion Dispersion

• Describes how the data varies or is dispersed
(scatter or spread out) about the true value.
Measures
• Commonly used measures
• Range – Min –Max Absolute Relative
• mean deviation
• Standard deviation Lorenz
Range QD MD SD
• Objectives Curve
• Judge reliability of measures of central tendency Coeff - Coeff Coeff -
• Compare two or more series MD - QD variance
• Control variability
• Use of other statistical measure

57 58

Measures of Relationship Correlation Analysis

 Methods of determining the relationship between variables of any
populations (bivariate or multivariate) depends: • States the relationship • r>0 = positive
 Existence of association or Correlation technique between variables • r<0 = negative
correlation; • Whether the relationship • r=0 = no relation
 Cause and effect relationship; is positive or negative
• r=+1= perfect +ve
btn variables Technique of regression • The strength or degree
of relationship • r=-1 = perfect –ve
• Positive Strength
 For bivariate population, Correlation can be studied through (a) cross
tabulation; (b) Charles Spearman’s coefficient of correlation; (c) Karl Pearson’s • When both rise or fall • -1 to -0.5 strong
coefficient of correlation; whereas cause and effect relationship can be together • -0.5 to -0.3 moderate
studied through simple regression equations (‘best’ fit that a straight line).
 For multivariate population: Correlation can be studied through (a) coefficient • Negative • -0.3 to -0.1 weak
of multiple correlation; (b) coefficient of partial correlation; whereas cause • When one increase other • -0.1 to 0.1 none or v weak
and effect relationship can be studied through multiple regression equations. decrease
60

10
 12-Jan-25

Distinction Testing of Hypothesis

Correlation Regression
• Relationship • Average relationship Parametric Non-Parametric
• No cause-effect • Cause-effect
• Spurious • Normal data • Nominal ordinal
• No nonsense
• Mutual dependency • Measurement – metric • Qualitative and
• Functional dependency
• Purpose –measure • Quantitative data quantitative
degree of linear • Purpose- which variable
relationship is dependent on another • Whole population • Measurement –non-
• Units of measurement- • Unit- by x=(bxy) metric
-1 to +1 • Regression
• Wider/ non-linear • Chi-square
• Scope - limited • ANOVA
• T,f,z
61 62

Parametric Test
The test you choose depends on level of measurement: T-test - sample size is less than 30
test the mean of one population against the standard
Independent Dependent Statistical Test
two independent variable
Dichotomous Interval-ratio Independent Samples t-test discrete
Dichotomous
Z test – above 30
Nominal Nominal Cross Tabs single proportion
Dichotomous Dichotomous
difference of proportion- two population
Nominal Interval-ratio ANOVA single mean on specific value of population mean
Dichotomous Dichotomous
single variance
Interval-ratio Interval-ratio Correlation and OLS Regression equality of variance
Dichotomous F test - ratio of variance of 2 samples
ANOVA
variance of two means
compare 2 population variance
63 64

Non-Parametric Test – Chi-square test Factor Analysis

• Measure which evaluates extent to which a set of the observed
frequencies of a sample deviates from the corresponding set of
expected frequencies of the samples. • Statistical method used to describe variability
• Aggregate discrepancies between actual and expected among observed, correlated variables in terms of a
frequencies potentially lower number of unobserved variables
• Goodness of fit called factors
• Independence of two attributes
• Two variable –measurement - non-metric
• 1 6 10
• 2 5 10
• 3 7 10 • Types
• 4 9 10
• 5 3 10 • Exploratory factor analysis EFA
• 30/5=6 50/5=10 • Confirmatory factor analysis CFA

65 66

11
 12-Jan-25

Trend
 A trend is a pattern of gradual change in a condition,
output, or process, or an average or general tendency of
a series of data points to move in a certain direction
over time, represented by a line or curve on a graph.
 To follow a trend you must not only be aware of what is
currently happening but also be astute enough to
predict what is going to happen in the future.

68

Calculating Trends Calculating Trends

Adults and children tested for Maralia, 2008–2011 Adults on ART and children on ART, 2011
200 200
180
# of people (Hundreds)

180
# of people (in thousands)

160 160
140
140
120
120 100
# adults on ART # adults on ART
100 80
# children on ART # children on ART
80 60
60 40
40 20
0
20
0
2008 2009 2010 2011 69
70
69 70

Types of Charts

71

12
 12-Jan-25

Interpretation of Data - Significance

 Recommendations/ suggestions - solutions
 Decision-making
 Forecasting
 Development of models
 Validate theory
 Future reference
 Motivate the researcher

76

Interpretation of Data – Precautions Ethical issues concerning research participants

• Consistency  Safety of respondents.
• Reliable data  Permission or consent.
• Accuracy  Incentives: Offering incentives, gifts, etc for seeking
• Avoid bias information is unethical and equivalent to bribing.
• Homogenous data
• Uniformity  Sensitive Information: be careful about the sensitivities
• Suitability of the participants.
• Adequacy  Confidentiality: Sharing information about a respondent
• Proper generalization
• Proper processing of data with others for purposes other than research is
• Proper analysis unethical.
• Objective in mind

77

13
 12-Jan-25

Ethical issues relating to the researcher Plagiarism

 Avoid biasness or Prejudice • Presentation of the work of another person as one’s
 Provision or deprivation of a treatment own or without proper acknowledgment
 Inappropriate research methodology • Copying word to word
 Misrepresentation of facts: report the findings in a • Images, graphs, oral presentation, computer
way that changes or slants them to serve your own assignment, tables, ideas, etc
or someone else‘s interest is unethical • Copy, paraphrase or translate without the source of
original work
 Misuse of data: The data collected has to be used
• Reasons
only for the purpose it is collected for not for • Give credibility
making unethical usage. • Help reader to go and check
• Protect yourself from plagiarism

80

Dissemination of Results
Avoid Plagiarism
• Cite
• Paraphrase
• Quoting “ “
• Citing quotes
• Citing own material – self citation
• Referencing

81

Research Report Format - (Traditional)

 Title Page
 Table of Contents
 Abstract
 Acknowledgements
 Introduction
 Theoretical Framework and Review of Related Literature
 Research Design
 Data Analysis and Interpretation
 Conclusion and Suggestions/ Recommendations for
Further Research
 References/ Bibliography
 Appendices

14