Lecture 2 (sept 13th) Flashcards
Study Goal vs Objective
Goals are just ideas, objectives are actionable and planning related
Theoretical framework
a set of established models in the published literature that can
inform the components and flows of the conceptual framework for a new research study
We use models in health research
To organize our thoughts
To explore determinant/health relationships in a systematic manner
As a guide for analysis
To identify ‘modifiable’ factors for intervention
Lalonde’s Model
Title: A New Perspective on the Health of Canadians.
Considered to be first modern government doc in Western
context to propose the health field look beyond the
biomedical health care system.
Lalonde model: Proposed health field
= human biology + environment +
lifestyle + health care organization
Evans and Stoddart model
See Tutorial 2
Eco-social Model/the Big picture
Individual
Sex, age, health, etc.
Eco-social Model/the Big picture Microsystem
Family
school
health services
peers
Church community
Neighbourhood play area
Exosystem
social welfare services
mass media
legal services
friends of family
neighbours
Macrosystem
ideologies of the culture and attitude
conceptual model
A model that a researcher sketches using boxes and arrows to illustrate the various relationships
that will
Study goal
Objectives
Models
Methods
> ideas
actionable / Plan
a framework
blueprint / instructions
Probability Sampling
Involves using selection techniques wherein the probability of selecting
each sampling unit is known
Prob (sampled) = n/N should equal 1
Probability Sampling Methods
- Simple random sampling
- Stratified random sampling
- Cluster sampling
- Multistage sampling
Simple Random Sampling
A simple random sample is a randomly selected subset of a population.
ex. 12 out of 36 people randomly
selected
Stratified Sampling
Random samples from distinct groups
> Geography (urban, rural, suburban)
> Sex (male, female, other)
Ex. 3 from each stratum
are randomly selected
Cluster Sampling
Researchers divide a population into smaller groups known as clusters. They then randomly select among these clusters to form a sample.
> Natural clusters (schools, neighbourhoods)
rather than individual units are selected
> Observation are made on all units within a
cluster
> Cluster sampling of schools
> All eligible children within each school included
Multistage Sampling
> Primary sampling units are selected
(e.g. municipalities)
Secondary units selected within
primary units (e.g. individuals)
May have more levels ex:
> Secondary units: city blocks
> Tertiary units: individuals
Other Sampling Techniques : Nonprobability-based sample
Convenience Sample:
Purposive sampling:
Convenience Sample:
Selection from a nonprobability-based
source population due to ease of access to those individuals,
schools, workplaces, organizations, or communities
> Use with caution
> Often systematically different from the target and source
populations they are intended to represent
Purposive sampling:
Recruitment of the participants for a
qualitative study based on the special insights they can provide
(Nov. 15)
Key informants
Are individuals selected to participate in a
qualitative study because they have expertise relevant to the
study question
Other Sampling Techniques
Errors Associated with Sampling
> Does the sample represent the target
population?
External validity
Sampling bias
> Don’t confuse with selection bias (when
healthier or educated people are more likely to
volunteer for research, September 27)
Ethical issues in sampling
Vulnerable Populations
Definition: Those who might have limited ability to make an autonomous decision about
volunteering to participate in a research study (more next class on Wednesday)
Vulnerable Populations examples
Young children
Some individuals with serious health issues
People in prison and some other socially marginalized populations
Limited ability to make an independent decision about volunteering because of other reasons
Vulnerable populations
Should not be selected unless absolutely necessary
At the same time, systematically exclusion is also unethical
Their health issues should be studied
Extra consideration of the potential risks of research to participants (stricter ethical
considerations)
The Importance of Sample Size
We want to recruit just the right number of participants
Recruiting too many participants wastes resources
Recruiting too few participants makes the study invalid
The desired sample size (the number of observations in a data set) for a quantitative study
is based on statistical estimations about how many data points are required in order to
answer the study question with a specified level of certainty
Proper Sample Size can be Calculated Based on
The effect estimate of exposure on the outcome:
The amount of errors we can accept
Variation of the outcome in the population
The effect estimate of exposure on the outcome:
For larger effect smaller sample is required
The amount of euro(s) we can accept
For smaller desired error larger sample is
required
Variation of the outcome in the population
Complex, largest when prevalence is 50%
Type 2 error (β)
False negative
You’re not pregnant when you actually are.
Type one error (α)
False positive
example : you’re pregnant and you’re actually not
Type 1 error (α) occurs when
A study population yields a
statistically-significant test result even though a significant difference or association does not actually exist in the source population
Type 2 error (β) occurs when
A statistical test of data from
research finds no significant result even though a
significant difference or association actually exists in the source population
Amount of random α and β errors is
a function of sample size
(1- β) is usually referred as the
The power of the study, i.e. how much the study is able to detect a
true relationship, can be calculated post hoc (after the study is complete
Data Collection in Health Often (NOT ALWAYS) is conducted by
interviewing (either self or by an interviewers) participants via questionnaires or survey instrument
Other methods of Data Collection in Health:
Direct measurement of physical functioning (measurement of height, weight,
blood pressure)
Blood work (measurement of hormones, sugar)
Sampling other tissues
Imaging (X-ray, CT scan, MRI)
Interview is …
…the process of verbally asking a participant questions and recording that
person’s responses
Trained interviewers record the responses, and they can ensure the accuracy and
completeness of each questionnaire
Interviews may require major time commitments from study personnel
Self-administered survey
is a questionnaire form that participants complete
by themselves, using either a paper-and-pencil version or online
In-depth Interview for Qualitative
Studies
Spending 1 or 2 hours interviewing a key informant (not sampled) using open-ended questions
> semi-structured interview
>Probing
In a semi-structured interview
the interviewer starts with a list of open-ended
questions that will be asked of each participant, but these questions or lists of topics
are merely starting points for eliciting responses from participant
Probing
Is an interviewing technique that prompts an interviewee to provide a more complete or specific response
Used in qualitative studies
specifically interviews
Interviewers can also record their observations
of the body language and other
nonverbal communication conveyed by interviewee
Interview pros and cons
Pros: we can train interviewers to ensure the accuracy and
completeness of each questionnaire
Cons: may require major time
commitments, expensive
Self-administered survey pros and cons
Pros: cost- and time-efficient
Possibly of approaching a large number of participants
The best way to get honest answers to sensitive questions
Cons: Problematic for low literacy populations, and those
who have limited Internet access or be unfamiliar with
computers
page 184 of the textbook
Training Interviewers
> Standard protocols will be followed to ensure the interview process is
the same for all study participants
All interviewers should undergo role-specific training and have an
opportunity to practice their interview skills, pilot testing, quality control to avoid interviewer bias
Interviewer bias
Occurs when interviewers systematically question cases and controls or exposed and
unexposed members of a study population differently, such as probing only individuals they believe to
have the disease or exposure of interest for more information
Uniformity
is easiest to accomplish when all interviewers are
provided with the tools they need to follow a standardized set of procedures
Questionnaire
A series of questions used as a tool for systematically gathering data from
study participants
Systematically Gathering Data for Focused and specific content steps (3)
1) List the topics that the survey instrument must cover
Theoretical frameworks can inform the diversity of questions that may be relevant
to include in the survey instrument
2) The first set of questionnaire items typically are ones that enable the
researcher to confirm that participants meet the eligibility criteria for the
study, exclude ineligible people
3) Several questions may be required in order to accurately assign participants to
key exposure and disease categories
Types of questions
Closed-ended questions
Open-ended questions
Open-ended questions
(also called free response questions) allow an unlimited
number of possible response
Closed-ended questions
allow a limited number of possible responses
> Date and time questions
> Numeric questions
> Categorical
> Ranked (ordinal) categorical answers have an inherent order
> Unordered (nominal) categorical answers do not have any built-in order
> Paired comparisons
Practical Considerations for questionnaires
1) Order of questions
2) Getting the correct answer
3) Data recording method
4) Layout and formatting
5) Back translation
1) Order of questions
> It is better to start with easy or at least general questions before moving to more
difficult or sensitive questions
It is often best to group similar questions with similar response types, so that
they are asked consecutively
2) Getting the correct answer
Sometimes it is better to mix up such questions to prevent habituation, an error that occurs when
participants become so accustomed to giving a particular response (like “agree… agree… agree…”) that
they continue to reply with the same response even when that does not match their true perspectives
Survey developers must carefully consider how previous questions could taint the answers to later ones
