Research Methods in Human

Nutrition
Instructor: Fentanesh Nibret ( Ph.D Associate Professor )

Department of Applied human nutrition,

Faculty of food and chemical engineering,

Bahirdar Institute of technology,

Bahirdar University

Email: fantshnibret@[Link]

Fentanesh N.(PhD.) 1
 Data collection

Fentanesh N.(PhD.) 2
Some Issues to Consider in Choosing Data Collection
Methods

o Cost
o Geographic location of respondents
o Length and complexity of questionnaire
o Sensitivity of topic
o Accuracy
o Error
o Coverage error
o Sampling error
o Nonresponse error
o Measurement error

Fentanesh N.(PhD.) 3
 Data collection techniques

Quantitative Data collection  Qualitative data

 Using available information collection
 Observing  Focus group discussions

 Interviewing  In-depth interview

 Self-administering written  Observation

questionnaires  Key informant interview

Fentanesh N.(PhD.) 4
 Quantitative Data collection

1. Available data
 Clinical records and other personal records, death
certificates, published mortality statistics, census
publications, etc

 Advantages:
o Documents can provide ready made information relatively
easily
o The best means of studying past events

Fentanesh N.(PhD.) 5
 Disadvantages of available data

1) Problems of reliability

2) There is a possibility that errors may occur when the

information is extracted from the records

3) The information required may not be recorded at all, or

only partly recorded.

Fentanesh N.(PhD.) 6
 2. Observation

 Observation is a technique that involves systematically

selecting, watching and recording behavior and
characteristics of living beings, objects or phenomena

 Observation can give additional, more accurate

information on behavior of people than interviews or
questionnaires

Fentanesh N.(PhD.) 7
 2. Observation cont.---

 Observations of human behavior can form any type of

study, but as they are time consuming

 Observations can also be made on objects the presence

or absence of latrines and the state of cleanliness may be
observed

Fentanesh N.(PhD.) 8
 3. Interviewing

 It involves oral questioning of respondents, either

individually or as a group

 Answers can be recorded by writing or by tape-recording

the responses, or by a combination of them

 Interviews can be conducted with varying degree of

flexibility (high degree of flexibility Vs low degree of
flexibility)
 High degree of flexibility: Usually used when studying sensitive
issues or when the researcher has little understanding of the
problem

Fentanesh N.(PhD.) 9
 Interviewing cont.

 When studying sensitive issues (e.g. abortion ) the

investigator may use a list of topics rather than fixed
questions

 The sequence of topics should be determined by the flow

of discussion

 It is often possible to come back to a topic discussed

earlier in a later stage of the interview

Fentanesh N.(PhD.) 10
 4. Self-administered questionnaire

 Written questions are presented that are to be answered

by the respondents in written form

 A written questionnaire can be administered in different

ways, such as by:
 Sending questionnaires by mail
 Gathering all or part of the respondents in one place at one
time
 Hand-delivering questionnaires to respondents and
collecting them later

Fentanesh N.(PhD.) 11
 Qualitative data collection

1. Focus group discussions(FGD)

 FGD can be used when:
 Limited resources (time, manpower, finances)
 The phenomena being researched requires a collective
discussion in order to understand the circumstances,
behaviour or opinions
 Greater insights may be developed of the group dynamic

 FGDs allow a group of 6-12 informants to freely

discuss a certain subject with the guidance of a
facilitator or reporter
 The members of FGD should be homogeneous

Fentanesh N.(PhD.) 12
 Qualitative data collection…

2. In depth interview
 Qualitative interviews should be fairly informal and
participants feel they are taking part in a
conversation or discussion rather than in a formal
question and answer situation.

Fentanesh N.(PhD.) 13
 Qualitative data collection…

3. Observation
 Observation can sometimes obtain more reliable
information about certain things
 for example, how people actually behave
 Observation can also serve as a technique for verifying of
nullifying information provided in face to face encounters

 People or environment can be observed.

Fentanesh N.(PhD.) 14
 Qualitative data collection…

4. Key informant interview

 Obtaining information from a community resident who is in a position
to know the community as a whole, or a particular area of interest

 The person can be a profession person who works with the population
you want information about, or a member of the target population

 Provides in-depth information about causes of the problem

 Informants may give you their own impression and biases

Fentanesh N.(PhD.) 15
 Questionnaire Design

 Purpose: To get reliable and valid data/reports

from respondents
 Good survey questions provide reliable (consistent)
and valid (accurate) measures

 When 2 respondents are in the same situation they

should answer the question the same way

 The answers to questions should correspond to

what they are intended to measure

Fentanesh N.(PhD.) 16
 Before Writing Questions
 Decide the study’s purpose (aims, research
questions, hypotheses)

 Identify the variables to be measured

 Develop a preliminary analysis plan

Fentanesh N.(PhD.) 17
 Process of Questionnaire Design
 Decide what information (variables) is needed

 Draft or obtain questions to elicit that information

 Put questions in meaningful order

 Add other elements of questionnaire

 Pretest questionnaire

Fentanesh N.(PhD.) 18
 Types of Questions-
Open-ended
 Responses are not provided to the respondent

Advantages:
-Respondent not influenced by specific alternatives
suggested

-Respondent can reveal what is most salient/relevant

-Can be used to build rapport in interview

Fentanesh N.(PhD.) 19
 Open-ended questions

 Disadvantages:
o Large amount of information revealed, may be vague or
irrelevant

o Respondents may vary in ability to articulate

o Respondents may be reluctant to reveal detailed

information or socially unacceptable opinions or
behaviors

o Difficulties in recording and in reducing and coding

material

Fentanesh N.(PhD.) 20
 Types of Questions –
Close-ended
 A list of acceptable responses is provided to the respondent
o Advantages:
-Easier for respondent
-Communicates same frame of reference to respondents
-Standardization
-Less variability in interviewer performance
-Less time to administer and record response

Fentanesh N.(PhD.) 21
 Close-ended Questions

Disadvantages:
o Lack of spontaneity permitted respondent

o Respondent may be forced into an unnatural frame of

reference

o May suggest response categories respondent has not

thought of

o Respondent may not feel as involved or motivated by

questionnaire

o Need to know appropriate response categories in

advance
Fentanesh N.(PhD.) 22
 In questionnaire design remember to:

 Make the wording clear and unambiguous in meaning

 Use familiar and appropriate language

 Write questions like people talk not like people write

 Start with simpler questions

 Ask the same question to all respondents

 Provide sufficient space for the response (For open ended

questions)

Fentanesh N.(PhD.) 23
 In questionnaire design remember to:

o Avoid abbreviations, double negatives, etc.

o Avoid two elements to be collected through one question

o Avoid leading questions

o Avoid embarrassing and painful questions

o Provide other, or don‟t know options where appropriate

Fentanesh N.(PhD.) 24
 In questionnaire design remember to:

o Provide the unit of measurement for continuous variables

(years, months, kgs, etc.)

o Arrange questions in logical sequence

o Pre-code the responses to facilitate data processing

o Pretest the questionnaire

Fentanesh N.(PhD.) 25
 Question Order

o General questions before specific

o Beginning – inviting, interesting, non-threatening

questions

o Middle – most important, put difficult and sensitive issues

o Closing – easy questions again, often routine,

background

Fentanesh N.(PhD.) 26
 Why respondents report questions/events with
less than Perfect Accuracy

o They do not know the information

o They cannot recall it, although they do know it

o They do not understand the question

o They do not want to report the answer in the interview context

Fentanesh N.(PhD.) 27
 Pretesting the Questionnaire

o A pre-test is where a questionnaire is tested on a small

sample (5-10%) of respondents before a full-scale study

o Purpose: to identify any problems such as unclear

wording, question variation, meaning, and respondent
interest and attention

Fentanesh N.(PhD.) 28
 Issues to be addressed in the pretest

o Is each question measuring what it is intending to

measure?

o Is the wording understood by all respondents, and is

the understanding similar for all respondents?

Fentanesh N.(PhD.) 29
 Pretesting the Questionnaire

Fentanesh N.(PhD.) 30
 Questionnaire layout

o The questionnaire should be visually easy to read and comprehend

o Lower case letters should be used for texts
o The covering letter
o It is customary for the first few lines of a questionnaire to include:
o –the title of the study and a brief introduction
–the label „confidential‟,
– the respondent‟s serial (identification) number (to preserve
anonymity),

Fentanesh N.(PhD.) 31
 Question forms

 Structured questionnaires: involve the use of fixed

questions or scales which are presented to respondents in
the same way, with no variation in question wording and
with closed questions

 Some structured questionnaires will also include open

ended questions, to enable respondents to reply in their
own words

Fentanesh N.(PhD.) 32
 Question forms

 Semi structured interviews include fixed questions but

with no, or few, response codes, and are used flexibly, often
in no fixed order, to enable respondents to raise other
relevant issues not covered by the interview schedule

 Unstructured interviews are comprised of a checklist of

topics, rather than fixed questions, and there are no pre
codes

 The more structured approach is only suitable for topics

where sufficient knowledge exists for largely pre coded
response formats to be developed

Fentanesh N.(PhD.) 33
 Data Quality Control Issues

Describe/provide
 Operational definitions of crucial concepts

 Selection and training of field workers/research staffs

 Field testing the research methods and tools (pretest)

Fentanesh N.(PhD.) 34
 Bias, Validity , and Reliability

 In research bias occurs when systematic error is

introduced into sampling or testing by selecting or
encouraging one answer over others

 Validity in data collection means that your findings truly

represent the phenomenon you are claiming to measure

 Reliability is also a way of assessing the quality of the

measurement procedure used to collect data in research

 In order for the results from a study to be considered

valid, the measurement procedure must first be reliable

Fentanesh N.(PhD.) 35
 Assignment 6.
o Prepare appropriate data collection tool for your proposed
research topic?

Fentanesh N.(PhD.) 36
 Data Analysis and Presentation

Fentanesh N.(PhD.) 37
 Data Analysis and Presentation

 Types of statistical Software

 For data entry:
 Epidata
 EPI Info
 For data analysis
 SPSS
 SAS
 STATA
 R software

Fentanesh N.(PhD.) 38
 Data management

 Data processing refers to:

 Data entry into a computer

 Data cleaning

 Data checks and correction

 Variable coding
 The aim of this process is to produce a relatively „clean‟
data set

Fentanesh N.(PhD.) 39
 Data management cont…

Data Editing
 After data collection

 Check for an answer code for each variable

 Editing for special codes

 Missing values occur when measurements were not taken, or respondents

did not answer questions

 In general, missing values should not be entered as a „ blank‟, because

some statistical packages interpret blanks as zero

 Ideally, a code should be chosen to denote a missing value (e.g. a code „9‟
or ‟99‟

Fentanesh N.(PhD.) 40
 Data management cont. …
Data entry
 Data entry concerns the transfer of data from a questionnaire to a
computer file

 It is converting data into a form that can be read and manipulated

by computers used in quantitative data analysis

 Before analysis, data must be checked for errors, variables

requiring coding must be coded, and missing values must be
appropriately addressed

Fentanesh N.(PhD.) 41
 Data management cont…

Data Cleaning

 Once data have been gathered, they need to be entered

into a computer data file and checked for errors

 No matter how carefully the data have been entered some

errors are inevitable

 The aim of this process is to produce a clean set of data for

statistical analysis

Fentanesh N.(PhD.) 42
 Data management cont…

Data coding
 Coding is assigning a separate (non-overlapping) numerical code for
separate answers and missing values
 Close-ended: Best to assign numeric codes to each category
Example: Eg. Instead of using „Male‟ and „Female‟ for
Male…………….1
Female……………..2
 In general computers are at their best with numbers

 Some statistical packages cannot analyze alphabetic codes, some

cannot understand open ended responses

 You must translate variables through coding

Fentanesh N.(PhD.) 43
Data management cont…

Data Modification
o Create new variables by:
o Recoding response categories
o Combining several variables that make up a scale (e.g., 9 items
to measure food security )
o Combining several variables to make one new variable (e.g
BMI=weight (kg)/height (m)2

Fentanesh N.(PhD.) 44
 Data Analysis and Presentation

 Analysis is usually a description of the data for each

variable for all the study units included in the sample

 The goal of data analysis is to discover useful information,

suggesting conclusions, and supporting decision-making

 It has multiple approaches ranging from descriptive to

inferential analysis

Fentanesh N.(PhD.) 45
 Data Analysis and Presentation…

 Descriptive analysis: are used to describe the basic

features of the data in a study
 What does the data look like

 They provide simple summaries about the sample and the

measures
 Frequency, proportion/percentage
 Table/cross-tabulation
 Diagram
 Summary measures (e.g mean, median)

Fentanesh N.(PhD.) 46
 Inferential statistics

 The goal is to make conclusions about a phenomenon

based on observed data

 In this way, statistical inference may help the researchers

in suggesting or verifying scientific hypotheses, examines
relationship/differences or decision makers in improving
their decisions

 There are many Statistical modes of performing inference

 Good statistical inference never move away very far

from the data
Fentanesh N.(PhD.) 47
 Choosing Statistical models

 The choice of statistical models for the specific data depends

on different properties:
 Objective of the study
 Study Design
 Nature of the variable
 What is the nature of the outcome and explanatory variables: is it a
continuous measurement, a count, a proportion, a category, or a time-at-
death?
 Distribution of the variable
 The nature of the data
 The number of group we want to compare
 Sample size

Fentanesh N.(PhD.) 48
 Statistical model

Or Chi-
Square test

Fentanesh N.(PhD.) 49
 Comparing two groups

 The outcome variable is continuous (e.g., BMI, weight,

hemoglobin level) and we want to compare means among
two groups
 t test (3 types of t test)
 One sample t test: Compares the mean of a single group to a
known or hypothesized population mean
 E.g: WHO reference mean hemoglobin for non-pregnant women = 12
g/dL
o You collect data from 100 women in Bahir Dar city
o Sample mean hemoglobin = 11.2 g/dL
o Is the mean hemoglobin of these women significantly different from
12 g/dL?

Fentanesh N.(PhD.) 50
 Comparing two groups

 Independent sample t test: Compares the means of two independent

groups to determine if there is a statistically significant difference
between them

 E.g: Mean birth weight of infants born to:

o Mothers who received nutrition education
o Mothers who did not
o Is the mean birth weight significantly different between the two
groups?

 Paired sample t test : Compares the means of two related

measurements from the same individuals groups
 E.g: Mean weight of women:
 Before nutrition intervention
 After nutrition intervention

Fentanesh N.(PhD.) 51
 Summery table t-test

Type of t test Groups compared Example

One group vs known Mean Hb vs WHO

One-sample
value standard
Independent Two different groups Urban vs rural BMI
Same group measured
Paired Weight before vs after
twice

Fentanesh N.(PhD.) 52
 Comparing two groups

 Both the outcome variable and explanatory variables are

categorical

 Example obesity among age group

 Categorical outcome + categorical explanatory variable →
Chi-square test

Fentanesh N.(PhD.) 53
 Comparing more than two groups

 To compare among more than two groups

o ANOVA
o Example: Weight among mothers age group (<20 years, 20-30,
and >30 years)
 Assumption
o The outcome variable is continuous

o The error term is assumed to be normally distributed

o Variance is assumed to be constant

o Factor variable is categorical with more than two groups

Fentanesh N.(PhD.) 54
 Hypothesis to be tested

 The hypothesis to be tested are

 H0: µ 1 = µ 2 = µ 3 =….= µ k
Versus
 H1: µ i ≠ µ j , j≠j
 The test statistics is F
 The tabulated value test of F at α level of significant and
df1 and df2 degree of freedom will be obtained from the
table where
 df1= is degree of freedom of the numerator

 df2= is degree of freedom of the denominator

Fentanesh N.(PhD.) 55
 Hypothesis test

 The calculated value of the test statistics can be obtained

from the ANOVA table

 Then if Fcal>Ftab then reject the null hypothesis

Otherwise, accept the alternative hypothesis

Fentanesh N.(PhD.) 56
 ANOVA-Test
SS (Sum of Squares): Measures the variation.
Treatments SS = variation between groups
Error SS = variation within groups (residual)
DF (Degrees of Freedom):
Treatments df = number of groups − 1 = 3 − 1 = 2
Error df = total observations − number of groups = 15 − 3 = 12
MS (Mean Square): Average variation
MS=SS/DF
Treatments MS = 27.897 ÷ 2 = 13.949
Error MS = 17.452 ÷ 12 = 1.454F
(F-statistic):𝐹=𝑀𝑆Treatments/𝑀𝑆Error=13.949/1.454≈9.59

 The test statistic is the F value of 9.59. Using an α of 0.05, we

have F0.05;2,12 = 3.89.
 So 9.59> 3.89, we reject the null hypothesis of equal population means
and conclude that there is a (statistically) significant difference among
the population mean

Fentanesh N.(PhD.) 57
 Statistical model
 Linear regression predicts a continuous outcome: e.g., weight,
blood pressure.
simple linear regression

yi = β0 + β1xi + ei
Where;
• yi​ –is the outcome we are trying to explain
• β0​ – intercept. the expected value of yi​ when xi=0i
• β1– slope coefficient. It shows the change yi​ for a one-unit
increase in xi​. xi​ – independent variable. It’s the variable we
believe has an effect on yi​.
• ei​ – Terror term (residual). It represents the difference between the
observed value yi​ and the value predicted by the model. It
captures randomness, measurement error, or other factors
affecting yi​

Explanation: How does (the mean of) y change with x ?

Prediction: What is the expected value of y, and how much
variation is there in the value of y, at a particular value x ?

Fentanesh N.(PhD.) 58
 Statistical model

 Multiple linear regression

yi = β0 + β1xi + β2zi + ei

 Explanation: How does (the mean of) y change with x and z?

 Prediction: What is the expected value of y, and how

much variation is there in the value of y, at a particular
combination of x and z?

Fentanesh N.(PhD.) 59
 Logistic regression

 Outcome variable: categorical

 Research question: Prediction
 Statistical model: Logistic regression
 Logistic regression predicts a probability of a categorical
outcome: e.g., probability of a child being malnourished
(yes/no), probability of disease (1/0). Where P= Probability
 Simple logistic regression that the outcome = 1
(event occurs)
α=Intercept (log-odds
when x = 0)
Β1=Effect of predictor x
on the log-odds of the
outcome
X= Predictor variable
​
Fentanesh N.(PhD.) 60
 Logistic Regression

Fentanesh N.(PhD.) 61
 Extension of logistic regression

 When the assumption of ordinary logistic regression

didn’t work, we use its extensions
Multinomial logistic regression
 Ordinal logistic regression
 Mixed effect logistic regression
 Generalized linear mixed model
 Generalized estimating Equation

Fentanesh N.(PhD.) 62
 Poisson regression

• Outcome variable: Count (discrete data with non-

negative integer values)
• Explanatory variables, X = (X1, X2, … Xk), can be
continuous or a combination of continuous and
categorical variables can be all categorical.
• Research question: Prediction
•

Assumption: Mean=Variance

Fentanesh N.(PhD.) 63
 Poisson regression

• Poisson Regression helps us analyze count data by allowing us

to determine which explanatory variables (X values) have an
effect on a given response variable (Y value, the count)

• Overall goodness of fit statistics of the model are the same as

for any GLM:
o Pearson chi square statistic
o Deviance
o Likelihood ratio test, and statistic

Fentanesh N.(PhD.) 64
 Hypothesis test

• The hypothesis to be tested are

•

• We use 95% CI and check whether the interval

includes one

• We can also use p value whether it is lower than

0.05

Fentanesh N.(PhD.) 65
 Survival Analysis

• Outcome variable: time to event

• The time from entry into a study until a subject has a particular
outcome
• Example: time from entering a study until death, recovery, or
any other event of interest

• Censoring: Subjects are said to be censored if they are lost to

follow up or drop out of the study, or if the study ends before
they die or have an outcome of interest

• Statistical approach: survival analysis

o Notation:
• T = survival time of a selected individual
• t = a specific point in time.
• (S t ) = Survival Function
• h (t) = instantaneous failure rate among survivors at time t also called hazard
function

Fentanesh N.(PhD.) 66
• The goal of survival analysis is to estimate and
compare survival experiences of different groups
• The survival function can be formulated as

• Life table is the probability calculated from the survival function

• It takes censoring in to account to compute probabilities
of the life table
Fentanesh N.(PhD.) 67
 •

Data analysis using SPSS

Fentanesh N.(PhD.) 68
 Analyze: Descriptive Data Analysis

Analyze

Fentanesh N.(PhD.) 69
 Chi-square test

Fentanesh N.(PhD.) 70
 Outputs

Chi-Square Tests
Value df Asymp. Sig. (2-
sided)
a
Pearson Chi-Square 3869.306 6 .000
Likelihood Ratio 3896.071 6 .000
Linear-by-Linear Association 1031.032 1 .000
N of Valid Cases 15683
a. 0 cells (0.0%) have expected count less than 5. The minimum expected
count is 27.98.

Fentanesh N.(PhD.) 71
 Independent sample t-test

• To compare the outcome variable (continuous)

among two levels of factor variable
• Eg. Comparing birth weight among sex of the child

Outcome
variable

(Independent
(group variable)

Level values
Fentanesh N.(PhD.) 72
 output

Fentanesh N.(PhD.) 73
 Analysis of Variance (ANOVA)

• Analysis of Variance (ANOVA) is used to compare

between groups if more than two
• AnalyzeCompare means One way ANOVA

Dependent variable
Fentanesh N.(PhD.)
Factor Variable 74
Fentanesh N.(PhD.) 75
 Linear regression

Dependent variable

Fentanesh N.(PhD.) Independent variables 76

 output

Fentanesh N.(PhD.) 77
 Logistic regression

 Now, the outcome variable is categorical

 If the categorical outcome variable is binary (only 2

possible values), then binary logistic regression is
fitted
 If the outcome variable has more than two levels of
nominal, then multinomial logistic regression

 If the outcome variable has ordered levels, then

ordinal logistic regression is fitted

Fentanesh N.(PhD.) 78
 Binary logistic regression

 Dependent variable: Birth weight ( normal=1, underweight

=0)
 Explanatory variable : Age of mother, sex of the child, parity,
marital status, and so on
 Bivariate logistic regression : Fitting logistic regression for
each explanatory variable
 Odds ratio, 95% CI and p value will be computed
• A variable will be taken to multivariable logistic regression if
p value is less than 20% (0.2)

Fentanesh N.(PhD.) 79
 Binary logistic regression

o Reference level should be defined by clicking “categorical” and

choose either last or first We can change the
reference level by
choosing either last or
first, then click “change”

Fentanesh N.(PhD.) 80
 Binary logistic regression

Fentanesh N.(PhD.) 81
 Bivariate logistic regression

 First run the frequency and check the distribution of

each explanatory variable over the dependent
variable
 This tells us whether the Variable is eligible for regression
 Example: dependent variable Anemia independent Education level
anemia_ordinal
Frequen Percent Valid Cumulative
cy Percent Percent
.00 10560 67.3 72.9 72.9
Valid 1.00 3929 25.1 27.1 100.0
Total 14489 92.4 100.0
Missin
System 1194 7.6
g
Total 15683 100.0
Categorical Variables Codings
Frequency Parameter coding
Dependent Variable Encoding
(1) (2) (3)
Original Value Internal Value
No education 6596 1.000 .000 .000
No 0 Primary 4862 .000 1.000 .000
Yes 1
Highest educational level
Secondary 2003 .000 .000 1.000
Higher 1028 .000 .000 .000
Fentanesh N.(PhD.) 82
 Odd
ratio
Variables in the Equation

B S.E. Wald df Sig. Exp(B) 95% [Link] EXP(B)

Lower Upper

V106 318.800 3 .000

V106(1) 1.059 .091 136.113 1 .000 2.882 2.413 3.443

Step 1a V106(2) .539 .093 33.366 1 .000 1.714 1.428 2.058

V106(3) .295 .104 8.084 1 .004 1.343 1.096 1.646

Constant -1.721 .087 391.901 1 .000 .179

a. Variable(s) entered on step 1: V106.

P-value

 As we see from p value (<0.05) or 95% CI (not include 1), The Odds ratio here is
unadjusted , which may be affected by confounding variables
 However, in the baivariate analysis we only focus on 20% p value

Fentanesh N.(PhD.) 83
 Multivariable logistic regression

 Now we are going to include all potential explanatory variables

in the model simultaneously
Categorical Variables Codings
Frequenc Parameter coding
y (1) (2) (3) (4) (5) (6)
15-19 3172 1.000 .000 .000 .000 .000 .000

20-24 2667 .000 1.000 .000 .000 .000 .000

25-29 2652 .000 .000 1.000 .000 .000 .000

Age in 5-year groups 30-34 2091 .000 .000 .000 1.000 .000 .000
35-39 1772 .000 .000 .000 .000 1.000 .000
Independent 40-44 1212 .000 .000 .000 .000 .000 1.000

variables 45-49
Poorest 3649
923 .000
1.000
.000
.000
.000
.000
.000
.000
.000 .000

with codes Poorer 1957 .000 1.000 .000 .000

Wealth index
used for the combined Middle 1903 .000 .000 1.000 .000

analysis Richer
Richest
1948
5032
.000
.000
.000
.000
.000
.000
1.000
.000
No
6596 1.000 .000 .000
education
Highest educational
Primary 4862 .000 1.000 .000
level
Secondary 2003 .000 .000 1.000
Higher 1028 .000 .000 .000
No or
13436 1.000
Currently pregnant unsure
Yes 1053 .000
Type of place of Urban 4709 1.000
Rural 9780 .000
Fentanesh N.(PhD.)
residence
84
 Multivariable logistic regression AOR

B S.E. Wald df Sig. Exp(B) 95% [Link] EXP(B)

Lower Upper
V106 67.740 3 .000
V106(1) .674 .102 43.908 1 .000 1.962 1.607 2.394
V106(2) .370 .099 14.068 1 .000 1.448 1.193 1.758
V106(3) .256 .106 5.858 1 .016 1.291 1.050 1.588
V013 15.306 6 .018
V013(1) .191 .093 4.176 1 .041 1.210 1.008 1.453
V013(2) .301 .093 10.502 1 .001 1.351 1.126 1.621
V013(3) .294 .091 10.502 1 .001 1.342 1.123 1.604
V013(4) .288 .093 9.712 1 .002 1.334 1.113 1.600
Step 1a V013(5) .223 .095 5.534 1 .019 1.250 1.038 1.506
V013(6) .190 .102 3.491 1 .062 1.210 .991 1.477
V102(1) .070 .080 .760 1 .383 1.073 .916 1.255
V190 256.659 4 .000
V190(1) .895 .085 110.572 1 .000 2.448 2.072 2.893
V190(2) .328 .093 12.524 1 .000 1.388 1.157 1.664
V190(3) .238 .093 6.507 1 .011 1.268 1.057 1.523
V190(4) .065 .092 .494 1 .482 1.067 .891 1.278
V213(1) -.343 .069 24.395 1 .000 .710 .619 .813
Constant -1.757 .158 123.112 1 .000 .172
a. Variable(s) entered on step 1: V106, V013, V102, V190, V213.
Hosmer and Lemeshow Test
Step Chi-square df Sig.
P-value
1 8.199 8 .414

Goodness of fit test of the model can be checked

Fentanesh N.(PhD.)
using Hosmer and Lemeshow test (higher p value 85
good model)
 Assignment 7

 Choose appropriate analysis methods for your

proposed research
 Explain your reason for choosing the model

Fentanesh N.(PhD.) 86
 Research Ethics

Fentanesh N.(PhD.) 87
 Introduction

 Ethics:- moral principles of right and wrong

 A set of standards that includes both prohibitions
against and prescriptions for specific kinds of behavior
in research
o It is not absolute; may vary by person, by time, by place
o involves a balance between and within principles and
practices
 Two things to be consider about the ethics in the
conduct of research
o Basic ethical principles underlying research
o Formulation and approval of ethical procedures for conducting
research

Fentanesh N.(PhD.) 88
 Research must respect
 Dignity
 Safety
 Rights of research participants

Fentanesh N.(PhD.) 89
 Three Broad Areas of Ethical Concerns

 Ethics of research design, data collection and

analysis and results reporting

 Responsibility to society and groups in society

regarding benefits and risks of human research
o Has a responsibility to assess the possible uses of scientific
findings, to promote their beneficial application, and to
speak out against their destructive application
 Recruitment and treatment of research participants

Fentanesh N.(PhD.) 90
Historical Development of Medical Ethical

 Tuskegee experiment on syphilis (1932 -72)

 Studied the natural history of syphilis in 400 African
American males
 Motivated by the concern that syphilis might be less
harmful in this group
 Men afflicted with syphilis to go untreated for a
period of 40 years
 Finally
o 28 persons had died from the disease,
o 100 persons had died from related diseases

Fentanesh N.(PhD.) 91
 Historical Development …
 Nazi experiments (1933–45)
 Was the first among all evil human experimentation
 During World War II, a number of German physicians
conducted painful and often deadly experiments on
thousands of prisoners without their permission

 Considering the inhumane conditions, lack of consent,

and questionable research standards, modern scientists
overwhelmingly reject the use of results from
experiments

 Seventy medical researches were conducted in Nazi

concentration camps
 including:
 High-Altitude,
 Freezing experiment
 Head injury experiment
Fentanesh N.(PhD.) 92
 Historical Development …
 Cold War Human Radiation Experiment (1944-74)
 Over 4000 secret and classified radiation
experiments by
 Atomic Energy Commission (AEC)
 other government agencies
 believing low dose radiation was not harmful
ignoring
 increased risk of cancer
 heart disease
 neurological disorders
 immune system disease
 reproductive abnormalities
 birth defects, and genetic mutations

Fentanesh N.(PhD.) 93
 San Antonio Contraceptive Study (1970)
o Involved randomizing double-blind subjects
between active and placebos contraceptive pills to
identify the side effects attributed to oral
contraceptives

o The women were not informed as they were the

subjects of this type of research

o There were high numbers of unplanned

pregnancies in the placebo group

Fentanesh N.(PhD.) 94
 Response to Research Abuses

 Since 1945, Various codes have been adopted by

different organizations

 The best known are:

1. Nuremberg Code of 1947

2. Universal Declaration of Human Rights (1948)

3. Declaration of Helsinki (1964)

4. National research act (1974)

5. Belmont Report(1979)

Fentanesh N.(PhD.) 95
 1, Nuremberg Doctors Trial

 led to the formulation of the Nuremburg Code (1948)

 23 persons were charged

 Among these, 7 were sentenced to death by hanging, 9

were given prison terms, and 7 were found not guilty

 It outlined ethical principles required for research

 Now regarded as a milestone toward the

establishment of a permanent international court

Fentanesh N.(PhD.) 96
2. Universal Declaration of Human Rights (1948)

 The United Nations (UN) came into being in 1945 by

51 countries
 It is the most universal human rights document in
existence
 Consists of 30 articles
 Subsequent international agreements
 Regional human right instruments
 National constitutions and law
 Ethiopia becomes a member in 1945

Fentanesh N.(PhD.) 97
 3. Declaration of Helsinki (1964)

 World Medical Association drafted in 1964

 Has been revised 7 times and the current on is in 2013
 It consists of 2 preamble and 35 general principles
(37)
 Developed on the Nuremberg Code
 It adds three key points on NCode:
 Higher priority for the subject than society
 Every subject should get the best known treatment
 independent review of all human subject research

 It is the origin of Institutional Review Board (IRB).

Fentanesh N.(PhD.) 98
 4. National Research Act (1974)

 The Commission had four goals that it needed to analyze

1. Boundaries between biomedical and behavioral research

2. Assessing the risks and benefits of the appropriateness of

research

3. Determining appropriate guidelines for how human subjects

can be chosen for the participation of research

4. Defining what informed consent is in each research setting

 Recommendation for Belmont report

Fentanesh N.(PhD.) 99
 5. Belmont Report

 Ethical Principles and Guidelines for the protection of

Human Subjects of Research
 Attempts to summarizes the basic ethical principles
identified by the National Commission for the Protection
of Human Subjects of Biomedical and Behavioral
Research
 Consists three ethical principles:
• Respect for person
• Beneficence
• Justice

Fentanesh N.(PhD.) 100

 Ethics in Health Research

 Health research that uses human being as a

subject has benefits and risks to the society
 It is beneficial to the society as it ensures good
health
 It is risk in that the human participant of
research may be exposed to undesirable
consequences
 In order to strike a balance between these two
conflicting societal interests, countries have
developed laws and ethics to promote health
research

Fentanesh N.(PhD.) 101

 Ethics in Health…

 Risks in general fall into three categories

 Invasion of Privacy
 Breach of Confidentiality
 Study Procedures
 When assessing risk associated with participation
consider
 The Probability of Harm
 The Magnitude of Such Harm
 There are three widely recognized principles
 Respect for Persons
 Beneficence/Non-Maleficence
 Justice/Non-Exploitation

Fentanesh N.(PhD.) 102

 Respect for Persons

Autonomy: Each individual:

o Is unique and free
o Has the right and capacity to decide
o Has value and dignity
o Has the right to informed consent
Protection for vulnerable persons
o Special protections for those whose decision making capacity is impaired or
diminished
 Application: Informed Consent
o The consent process must include :
o Information
o Comprehension, and
o Voluntary participation

Fentanesh N.(PhD.) 103

 Beneficence
 Protection of the study participants is the most important
responsibility of the researcher
 Researchers must:
o Protect the physical, mental and social well-being
o Minimizes physical and social risks
o Maximize the possible benefits
o Retain the community perspective
 Application: Assessment of risks and benefits
o The nature and scope of risks and benefits must be assessed in a
systematic way

Fentanesh N.(PhD.) 104

 Justice

The principle that calls for fairness

 Research must: 5 formulations to each person:
1. an equal share
2. according to individual need
3. according to individual effort
4. according to societal contribution
5. according to merit
Application: Selection of participants
1. There must be fair procedures and outcomes in the selection
of research participants

Fentanesh N.(PhD.) 105

 Informed consent

 Informed Consent is a voluntary agreement to

participate in research

 Elements should be included in an informed

consent?
o Purpose of the research
o Procedures involved in the research
o Alternatives to participation
o All foreseeable risks and discomforts to the subject
o Benefits of the research to society and possibly to the individual
human subject
o Length of time the subject is expected to participate
o Person to contact for answers to questions or in the event of a
research-related
o Right to confidentiality and right to withdraw from the study at any
time without any consequences

Fentanesh N.(PhD.) 106

 Informed consent

 The consent form need to include the following

 Identification of the researcher
 Identification of the sponsoring institution
 Indication of how the participants were selected
 Identification of the purpose of the research

Fentanesh N.(PhD.) 107

 Consent by surrogates cont…

 Parents are considered surrogates for their minor

children
 Spouses for one another
 Adult children for parents when parents are
lacking
 Adult grand children for grand parents

Fentanesh N.(PhD.) 108

 Confidentiality

o Confidentiality is necessary in diagnostic,

therapeutic and research context

o Confidential information is both private and

voluntarily imparted in confidence and trust

o If research subject authorizes release of

the information to others, then there is no
violation of rights or confidentiality

Fentanesh N.(PhD.) 109

 Confidentiality cont..

o Research may involve collecting and storing data relating

to individuals and groups

o Investigators should make arrangements for protecting

the confidentiality of such data by:

o Omitting information that might lead to identification

of individual subjects
o limiting access to the data, or by other means

o When personal identifiers remain on records used for a

study, investigators should explain why this is necessary
and how confidentiality will be protected

Fentanesh N.(PhD.) 110

 Dissemination and Utilization of Results

 Briefly describe the dissemination plan

o Feedback to the community
o Feedback to local authorities
o Identify relevant agencies that need to be informed
o Scientific publication
o Presentation in meetings/conferences
 Briefly describe how the study results can be best
translated into application

Fentanesh N.(PhD.) 111

 Work plan

 Work plan summarizes (in a table, chart, graph) the

various components of a research project and how they
fit together.
 Includes:
o Tasks to be performed
o When the task will be performed
o Who will perform the task (identify human resource needed for
each task)
o Number of staff needed to perform the task

Fentanesh N.(PhD.) 112

 Work plan cont..

 A work plan can serve as:

o A tool in planning the details of the project
activities and later the project funds
o A visual outline or illustration of the sequence of
the project operations
o A management tool for the principal investigator
and/or members of his/her team, showing what
tasks and activities are planned, their timing, and
when various members will be involved in the
tasks

Fentanesh N.(PhD.) 113

 Work plan cont..

 Ways of presenting a work plan

o Work schedule
o GANTT chart
 The Work Schedule
o Is a table
o Summarizes:
• tasks to be performed
• duration of each activity, and
• staff responsible.

Fentanesh N.(PhD.) 114

 Work plan cont..

 The GANTT Chart

o Represents graphically the order in which various
tasks must be completed and their duration of
activity
• A typical Gantt chart includes the following
information:
o The tasks to be performed
o Who is responsible for each task; and
o The time each task is expected to take
o The length of each task is shown by a bar that
extends over the number of days, weeks or months
the task is expected to take.

Fentanesh N.(PhD.) 115

 GANTT CHART

Fentanesh N.(PhD.) 116

 Budget
How should a budget be prepared?
 It is necessary to use the work plan as a starting
point

 Specify, for each activity in the work plan, what

resources are required

 Determine for each resource needed the unit

cost and the total cost

 The budget for the fieldwork component of the

work plan will include funds for personnel,
transport and supplies

Fentanesh N.(PhD.) 117

 Budget cont…

 Budget preparation
o Include a 5%-10% contingency fund
o If inclusion of a contingency fund is not
allowed, an alternative is to slightly over-
budget in major categories.

o ¨Do not box yourself in too tightly with

very detailed categories and amounts,
especially if regulations do not allow
adjustments afterwards

o.

Fentanesh N.(PhD.) 118

 Budget cont…

 Budget justification
o It is not sufficient to present a budget without
explanation
o The budget justification follows the budget as an
explanatory note justifying briefly why the various
items in the budget are required

o Make sure you give clear explanations concerning

why items that may seem questionable or that are
particularly costly are needed and discuss how
complicated expenses have been calculated

o If a strong budget justification has been prepared, it

is less likely that essential items will be cut during
proposal review

Fentanesh N.(PhD.) 119

 References

Methods of citations
1) Vancouver system
 For an article the following information should be
noted:
 Author(s)’ Surname followed by initials. Title of article. Name
of Journal. Year, Volume(number): page numbers of article.

 Example: Tiruneh FN, Individual-level and community-level

determinants of cervical cancer screening among Kenyan
women: a multilevel analysis of a Nationwide survey. BMC
women's health. 2017 Dec;17(1):109.

Fentanesh N.(PhD.) 120

 References cont..

2)The Harvard System

 In other journals and books it is common to put the year,
between brackets, straight after the name of the author(s)
 This is called the Harvard system.
 Name of the author(s) (year). Title. Place of Publication:
Publisher
 Example: Dagne, S., 2019. Assessment of breast cancer
treatment outcome at Tikur Anbessa Specialized Hospital
Adult Oncology Unit, Addis Ababa, Ethiopia. European
Journal of Oncology Pharmacy, 2(2), p.e13..
 There are more systems in use for referencing to
literature.
 Always carefully look what system is used in the journal
you are submitting an article to and follow it
systematically.

Fentanesh N.(PhD.) 121

 Annex

• Include in the appendices of your proposal any additional

information you think might be helpful to a proposal
reviewer.
• For example, include:
o The study questionnaire if you have it.
o The consent form.
o A copy of the approval from the Institutional Review Board.

Fentanesh N.(PhD.) 122

 Assignment 8.
• Write
o Ethical consideration
o Dissemination of result
o Work plan
o Budget
o Reference
For your proposed research

Fentanesh N.(PhD.) 123

Scientific Paper Writing

124
 Components of scientific Paper

 Title
 Abstract
 Introduction
 Objective
 Methods
 Results
 Discussion
 Conclusion
 Reference
 Acknowledgement
 Annex

125
 Cover page

 Should contain:
– the title
– name of the authors with their titles, positions and their
affiliation
– the month and year

126
 Title

• The title is the most important part of the paper because it

connects your research with readers
• Few words that can adequately describe the contents of the
paper
• Purpose: To provide a brief , informative summary that
will attract your target audience

127
 Abstract
 A summary of the paper, giving the reader a preview of
what's to come.

 Writing abstracts demands a systematic reflection and is

time consuming

 Allows readers to quickly scan the large scientific

literature, and decide which articles they want to read in
depth
 Purpose : to help readers decide whether to read the full
article
 Often 250-300 words, depending on the criteria set by
journal/school

128
 Components of the abstract

• Background: Brief description of the problem (Why this

study was needed)

– Main objectives of the study

• Methods: Place of the study; type of the study and

methods/analysis used

• Results: Main findings

• Discussion: Conclusion and recommendations

129
 Abstract cont…

 Abstract can be structured or unstructured (depends on

the instruction of the journal/school)

 Remember that only few words are allowed to be included

in the abstract

 Emphasize on what is new and useful

130
 An Abstract should

 State the principal objectives and scope of the investigation

 Describe methods employed
 Summarize the results
 State principal conclusions
 Provide a balanced summary of what was done and found.
 Common Mistakes
 Too much background or methods information
 References to other literature
 Abbreviations or acronyms

131
Example

132
 Introduction (why did you start?)

 The introduction should prepare readers to understand why and

how you did your research and tell them what to expect
 Purpose: to explain why your study is needed
 When using abbreviations, always define them the first time
they are used.
 Introduction should always
 Provide the background/rationale
 Review pertinent literature
 Objectives are usually the last part of the introduction

133
 Components of introduction
• Background statement: orient readers to the area of
interest and give them the background they need to
understand the problem.
• Problem statement:
– describe the nature and importance of the problem
– What do we need to learn
• Literature review: What do we already know
• Justification/activity statement:
– What will we learn from the study
– What you did to address the problem;
– How you studied the problem,
– Why you approached the problem the way you did.

134
 The purpose of the study

o The purpose of the study must be clear.

o The purpose is never use “to study”, “to evaluate”, “to

compare” or “to investigate”,

o Instead, the purpose is “to determine”, “to verify”, “to

describe” or “to select”, which refer specifically to what
you sought to accomplish with the research

135
 Methods (what did you do)
 Purpose →to describe how you collected, organized and
analyzed data that are relevant to answering your research
question
 Organize into logical subsections that illustrate the steps
you took to collect, organize and analyze the data

 Describe what you did in your research

 Describe original methods in detail; otherwise give

references

136
 Possible subsection of the methods

o Study design

o Study area and period

o Study population

o Sample size and sampling method(s)

o Variables of the study

o Data analysis

o Ethical considerations

137
 Possible subsection of the methods

o Study area and period: Location(s) (number of health

facilities, schools) , relevant dates,

o Give the dates of the data collection period and places

o Reporting possibly relevant events and can provide

additional information, such as how long it took to
reach the targeted sample size.
138
 Possible subsection of the methods

o Study population: Eligibility criteria, methods of selection

o Sample size and sampling method
– The nature and size of the sample and how it was selected are critical in
any research study

– The sample has to be large enough to provide estimates with enough

accuracy to be useful

– Larger samples provide more precise estimates in the form of narrower

confidence intervals.

– P values, essentially measures of chance as an explanation for the

outcome and that have no biological meaning
139
 Possible subsection of the methods

o Variables of the study

o Clearly identify your explanatory/independent and
outcome/dependent variables

o Operationally define (measure) outcome, explanatory,

potential confounders, and effect modifiers

o Data collection procedures: use appropriate data collection

tool/s

140
 Possible subsection of the methods

• Data analysis
– Explain how variables were handled in the analysis, i.e. continuous,
categorical, if so what were the categories
– Statistical methods you used
• Ethical considerations
– Informed consent
– Respect for anonymity and confidentiality
– Respect for privacy
– Approval of ethical issues

141
 Results (what did you find)

 Purpose →to describe the results of data analysis that are

relevant to the study
 Systematic presentation of your findings in relation to the
research objective is crucial
 Present the results for the primary endpoint/outcome first
 Results may presented with tables and figures, can also with
texts
 Result section should include:
 Characteristics of study participants (usually Table 1)
 Number of missing
 Main results
 Additional analyses, such as stratified analyses, interactions,

142
 Sequence of presentation of findings

 The 1st section is usually description of the study

population

 Provide information on the problem you studied

(prevalence, distribution, characteristics)

 In analytic study, the degree to which different

independent variables influence the problem

143
 Use of tables & figures

o Tables & figures need numbers & clear title

o Include only those tables & figures that present main

findings & need more elaborate discussion in the text

o Use tables to highlight individual values

o Use figures to highlight trends/relationships

o In tables, keep lines to a minimum;

o Use footnotes to clarify points of potential ambiguity

o Define abbreviations, even if they have been defined in the

manuscript.
144
 Text

o Describe what you found

o Present results in a logical sequence

o Less is often better ( a table or figure is worth thousand

words)

o Make sure all numbers in text are consistent with

tables/figures

145
 Discussion (What does it mean)

 This is your moment to shine and show your knowledge of

the prior literature.
 Summarize key results in reference to study objectives
 Summarize the study and the main findings in a few
sentences, instead of repeating individual results
 Interpretation
 Give an attention overall interpretation of results
considering the objectives, limitations, results from
similar studies, and other relevant evidence
 Mention the significance of your study
146
 Guidelines for constructing the discussion

 Focus on the main results

 Emphasize on what is new or different
 Interpret the result in context of literature /compare
 Interpret your results and suggest an explanation(s) for them
 Describe why you think you found what you found or did not find

 Critique your study (Strengths and Limitations)

Potential bias
Potential imprecision

147
 Discussion, con‟t.

 Whenever discussing your limitations try to put a positive

spin on what you have to say.
 For example, Measures of height and weight were
self-report, likely resulting in non-differential
misclassification. Nonetheless, such misclassification
was not an important cause of concern as some
found self-reported height and weight to be
reasonably accurate (citation)
 Give directions for Future researches

 Study strengths
148
 Conclusions & recommendations

o Conclusions should be short, as they have been elaborately

discussed in the discussion section

o Listing your conclusions will force you to be specific about

what you found

o Every conclusion needs to be supported with the data

presented in the results

o Recommendations are usually placed in roughly the same

sequence as the conclusion

149
 Conclusions & recommendations

 Recommendations may at the same time be summarized

according to the groups towards which they are directed

 Speculate on your findings.

What are their implications for public health,
– What are the policy implications?

150
 References

• Use sofwares like Endnote for citations.

• Reference style is dependent on the journal/school where

you want to submit your manuscript

151
 Annex
• Findings which are less important to be presented in the
main section of the paper will be presented this section

152
The End. Thank You!

153