Study Design, Measurement of Variables & Data Collection for Public Health Flashcards
How do people come up with research questions?
They have ideas about things around them and have questions to answer them
What does randomization do?
It limits the selection bias
What is a variable?
Something that we can assess in a study. It measures the attributes or characteristics of a study.
-can be a person, place, or thing
Ex: age, weight, height
Nominal scales?
Categorical values
-used w/ variables that refer to 2 or more classes, groups, categories, etc.
-attributes are only named; weakest
Ex: 0 Blonde 1 Brunette 3 Redhead; 0 Muslim 2 Catholic 3 Christian; The numbers mean nothing; just for data
Ordinal scales?
Ordinal variables
-values assigned based on rank. Think of a race or competition
-attributes can be ordered
Ex: Ellen 1, Tom 2, Kristen 3 means that Ellen is ahead or was first
Interval scales?
Continuous variables
-equal difference between variables have equal meaning
-values are measured on a metric
-distance is meaningful
Ex: yr. of birth: 1992, 1997, 2002 all have the same difference between them (5 years)
-most psychosocial states are measured this way
Ratio scales?
Continuous variables; the same as interval scales
- absolute zero
- keeps count
Likert Scale?
Range of answers that are equal distant from each other
-developed by Rensis Likert
-used to assess attitudes, beliefs, behaviors, etc.
Ex: Strongly Disagree to Strongly Agree
What is the logic scheme for research studies?
- Observation
- Research question
- Study Design
- Data collection
- Data analysis & results
What is AD de Groot (Empirical Cycle)?
a cycle
-Observation to Induction to Deduction to Testing to Evaluation. Over and over
What is empirical research?
a way of gaining knowledge. It answers a research question
Ex: Does sex education in elementary schools lead to less sexual risk taking in adolescence among both men and women?
Characteristics of research questions?
indicate the possibility of for testing
- associations between 2 or more variables
- associations between exposure and disease states
- associations between independent/dependent variables
What is a subjective question?
non-researchable; based on what individuals may know
-aim to measure feelings, attitudes, and perceptions of something
Ex: Should sex ed. be provided in elementary schools?
Criteria for research questions?
- expresses the relation between 2 or more variables
- implies that there must be a manner to test this relation
- should be able to test the variables
- asked in a question format
- specifies the population format
Types of research questions?
- Cause and effect (what is the effect of X on Y?)
- Association (is X related to Y?)
- Differences (does Y differ among groups of X?)
Does causation imply association? Vice-versa?
No
Variables can be measured. T or F?
True
- age can be measured in years or in categories
- gender can be male, female, etc.
- depression can be on a scale of the BDI
How do variables have value?
The value can diffrentiate based on what the variable is towards. Ex: 100 on an exam doesn’t have the same meaning as 100 on a thermometer
-Values are important bc it gives a proper way for things to be measured
Ex; Anxiety level of a person
Ex: iTune single download chart
Ex: Race/ethnicity
What are continuous variables?
variables that continue counting…
Ex: age; a person could be 25 yrs, 10 months, 2 days, etc
Ex: how many vaccination doses have you had?
-distance is equally spaced; space has meaning
What are categorical variables?
-variables can be put in categories
Ex: 0 brunette/1 redhead/2 blonde
-space doesn’t have meaning
What are ordinal (rank) variables?
Same as categorical variables, but w/ a difference.
Ex: Canada 1, USA 2, Russia 3 indicates a rank
-space doesn’t have meaning
Scales?
A set of numbers/values meant to represent the values of a variable
Ex: a ruler that measures height, length, width
Types of scales?
- Nominal
- Ordinal
- Interval
- Ratio
What converts data into information?
Statistics
Methods for data collection? (Interpretive)
- interviewing
- group discussions
- open/closed ended questions
- direct observation
- scales and measures
How can direct observation be overt?
when the subject or individuals in the environment know the purpose of the observation; open
How can direct observation be covert?
when the subject and individuals are unaware of the purpose; closed
What are closed ended questions?
Asks for a simple response
- yes/no
- multiple choice
- likert scale
Multiple linear regression?
if the Y/D/outcome/criterion is is continuous
Logistic regression?
if the outcome is categorical
What are the 2 main types of designs?
- Observational
- Experimental
Types of observational studies?
- Case-control
- Cohort
- Cross sectional
Types of experimental studies?
-RCT (Randomized control trial)
Cross sectional study?
observational study that involves data collection from a pop., at one specific point in time; asks of disease status and exposure
Ex: the gross annual income for each of 1000 randomly chosen households in New York City for the year 2000
Cohort study? (panel study)
observational study that is longitudinal and involves data collection from a group of people over many points in time to examine factors related to an outcome; evaluates exposure over time and looks at disease status later
Ex: students at HHS from 2010-2015 who may have gotten sick from eating the cafeteria lunch. record the outcomes, meet w/ them occasionally, etc.
Case-control study?
observational study in which two existing groups differing in outcome (disease) are identified and compared on the basis of some supposed causal attribute (exposure)
Ex: one group w/ heart palpitations will have ectasy to experiment and a control group also w/o heart palpitations will also be effected w/ ectasy
RCT?
directly comparing 1 treatment w/ another to determine which would have greater benefit
Observational studies can be done by?
-surveys
-electronic charts
-other approaches*
Experimental studies can be done by?
-clinical trial
-other systematic ways of data collection
Comparison may be done in both observational and experimental studies
T or F?
True
