0% found this document useful (0 votes)
7 views7 pages

Overview of Study Designs and Methods

The document outlines various study designs in research, categorizing them into observational and experimental studies, each with specific methodologies and purposes. It discusses sampling methods, data collection approaches, and data management strategies, highlighting the differences between quantitative and qualitative research. Additionally, it emphasizes the importance of study objectives in choosing the appropriate design and the limitations of each approach.

Uploaded by

2024mmed100
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
7 views7 pages

Overview of Study Designs and Methods

The document outlines various study designs in research, categorizing them into observational and experimental studies, each with specific methodologies and purposes. It discusses sampling methods, data collection approaches, and data management strategies, highlighting the differences between quantitative and qualitative research. Additionally, it emphasizes the importance of study objectives in choosing the appropriate design and the limitations of each approach.

Uploaded by

2024mmed100
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Study Designs December 30, 1899

Study Designs

1. Observational Studies
a. Case study/Case Series. You are not sure of what you are going
to find. You are exploring. You observe a few cases and draw
conclusions… may move further to a bigger study to test
hypothesis.
b. Descriptive studies. Similar to case studies but involves more
numbers of the study population. Larger Sample
c. Ecological studies. Aggregated data for entire populations is used
to draw conclusions. Findings are not for a unit/individual.
d. Cross-sectional studies, surveys (prevalence). One point in time
surveys. Census. Give no time component. Cheaper and faster.
i. Disease description
ii. Diagnosis and staging
iii. Disease processes, mechanisms
iv. Identification of associated factors
e. Case- Control Studies (retrospective). We identify cases and
compare them to controls. We ask about their past experiences.
Good for rare diseases. Identification of risk factors. We are unable
to calculate incidence nor prevalence.
f. Cohort Studies. Participants are selected based on an exposure
of interest and followed for time to see if they develop an outcome
of interest. We are able to calculate incidence, look at natural
progression of disease, identify risk factors. Here we have a time
component. We can get good quality data. It takes long and is
expensive. There are issues of self-selection. Can be prospective or
retrospective.
2. Experimental Studies
a. Randomized Controlled Trials. Unlike in the case of cohort
studies where we sit and wait for things to happen naturally. Here
we intervene or manipulate. We randomly assign people/rats etc. to
exposures or treatments. We balance out confounders. May be
long, expensive. A lot of ethical issues.

The above arrangement is based on the strength of the study.

The choice of the study design is based on the study objectives.


 Exploring a new area and would like to formulate hypothesis
 Calculate prevalence
 Study of a rare disease

1
Study Designs December 30, 1899

 Calculate incidence
 Looking at trends, rates
 Prove delicate relationships
[Link]
Study Population/ Study Participants

What is your unit of study? Who is your unit of study?

Sampling / Subject Recruitment


1. Probability Sampling Methods (everyone has equal chance of
being selected)
a. Simple random
b. Systematic sampling (the Kth member is selected)
c. Cluster sampling (clusters)
d. Stratified random sampling (Strata)
e. Multistage sampling (levels and different methods)
2. Non-Probability sampling (participants are selected based on
non-random criteria so it is not used for statistical inference)
a. Convenient sampling/ accidental sampling. The criteria
considered most is access
b. Snowball or network sampling
c. Purposive/judgmental sampling. Based on the researcher’s
experience.

The Sample

1. What do we mean when we say… we need a representative


sample? When is a sample biased?
2. Sample Size Calculation. Why do we worry about the sample
size?

Data Collection Approaches

Quantitative Qualitative

Methods Methods

Individual interviews Focus group discussions

Observation Key informant/in-depth interviews

2
Study Designs December 30, 1899

Record review Case histories/case study

Document reviews

Observation

Tools Tools

Structured Questionnaires / Semi-structured interview guides to


interview schedules … responses stimulate discussion. Greater use of open
are coded. Checklists ended questions

Sample size Sample size

Needs a large size, depending on Can make conclusions from a small


the prevalence of the variable of sample. These are usually not
interest. representative of the population

Usually selected using probability Sampling is usually purposive. A criteria is


sampling methods to represent the laid down to select who should be in the
population of interest. study. The common selection criteria
variables include … age, education, social
economic status and place residence
Findings can be generalized
The general rule is that you continue to
recruit participants until you are getting no
new information or are no longer gaining
new insights. This is called saturation

Validity of qualitative data is greatly


improved by using a combination of data
collection methods (using more than one
study population), a process known as
triangulation and independent analysis of
the data by more than one researcher.

Data Management Data Management

Using computer data analysis The procedure requires listening,


packages… EPI Info, SPSS, STATA observing, paying attention to one’s own
etc. inner biases. Also requires hours of typing
up field notes and a good system for filing
Data is coded, entered and cleaned
or cataloguing data;

3
Study Designs December 30, 1899

Inscription. Mental notes, learning to


notice what is important to other people
and what one has not been trained to see,
and then write down.

Description. Write things down in


jottings, diaries, logs, field notes. Field
notes are produced in a quiet place away
from the site of observation and interaction
with people in the field. They include
reflection, preliminary analysis, initial
interpretations, identification of gaps and
new questions.

Transcription. Word for word creation of


a written text from an audio or videotaped
account given by an informant. Involves
writing down verbatim, photographing,
videotaping or copying. It is important to
record each impression as soon as possible
after it occurs to avoid erosion,
modification and even falsification of the
mental record.

Tidying up. Organising data in files,


labelling the files, putting in boxes,
indexing. It is important to keep track of
the dates when notes, instruments or tapes
were handled, and who worked with them,
regardless of how the data were
manipulated or catalogued.

Data Analysis Data Analysis

Easy …. Frequencies, proportions, Cut and paste (manual or using software)


means, medians, confidence
Read and re-read your data. Make it
intervals, proportions, statistical
familiar. Continuous reflection and
tests of significance are run.
familiarisation with the data that is key.
What are they saying and who is saying
what?

4
Study Designs December 30, 1899

The software will help you code and


organize but your eye and brain will be
needed to find quality themes and
categories

Identify the range and importance of key


issues and items

Themes and Categories. Identify areas


of interest/concern. These depend on the
study objectives

Coding. Label each piece of the text …


coding… who, where, patterns, factors, key
words, phrases, local meaning.

Codes identify variables of interest in the


data.

Cut and paste (manual or using


software)

Using highlighters for similar ideas…


different colors

Cut outs in different labelled envelopes

Cut outs on walls under different themes


(titles and sub-titles)

Cut and paste into different documents

Compare as you analyse

Compare key issues across different


segments of the population (by age,
gender, marital status, social status,
context, roles, interests,) ….

Continue the reflection. What is missing?

5
Study Designs December 30, 1899

What is not being said?

Data Presentation Data Presentation

Tables, graphs Voices: key quotes, – verbatim. Be sure


whose voice. Bring quotes together that
Means, medians, confidence
support your arguments, assertions… what
intervals
you have observed, understood, concluded
Risk ratios, odds ratios, p-values
Case stories

Ranking issues

Maps, Venn diagrams, Decision trees, Fish


bones

Flow charts… boxes and arrows to indicate


association, choices, timeframes

Making Conclusions Making Conclusions

Size/magnitude of proportion, mean Review your prior assertions


and median
Rough frequencies or counts use of
Width of the confidence interval majority, less etc

Significant associations Similarity, co-occurrences, sequence

Omissions… what is not being said

Those who are not participating


(omukamoori)

Constant comparison. When, why, under


what conditions do these themes occur in
the text?

Construct relationships … try to explain the


story, the patterns, the meaning,
associations & effects, make arguments.
Tell what is going on?

Limitations Limitations

Difficult to use when trying to Observer bias. due to observer’s


explore sensitive topics like perceptions, views, ideologies, cultural

6
Study Designs December 30, 1899

prostitution, abortion. perspectives etc

Does not explore what people think, Low reliability…. The research usually
feel and experience. cannot be replicated or repeated

You might also like