Lesson0.1: Research Methods Flashcards
What is a theory?
A theory is a well-sustained explanation for a natural phenomenon that is broader than a single hypothesis
What do experiments do to theories?
Expand, support, or refute them
What is a requirement of a hypothesis?
It must be falsifiable
What is operationalization?
The process of determining how things will be measured and what those data will represent
What are independent variables also known as?
A factor, explanatory variable, predictor
What are levels?
Potential values of a factor/independent variable. Each factor has two or more levels.
What are combinations of factor levels called?
Treatments
What are experimental units?
The things being experimented on.
Animals: subjects
People: participants
What are confounding variables?
Unmeasured variables that may also affect the dependent variable
When should risks to participants be taken?
Only if it is necessary and outweighed by the expected benefits of the study
What should researchers do in a study where deception is necessary?
They should tell participants as much as possible about the expriment beforehand, should reveal its true purpose afterwards, and should give participants the opportunity to ask questions
What is a vulnerable population and what are some examples?
A population that is not able to protect its members’ self-interests in the course of being test subjects.
Examples include people who…
-lack the cognitive capacity to understand the risks involved
-lack the linguistic abilities to understand what they’re consenting to
-are in such poverty that monetary incentives would disproportionately convince them to undergo particularly risky work
What kind of work is ideal with vulnerable populations?
Work for vulnerable populations. For example, developing a pharmaceutical that may more quickly cure jaundice in newborns.
What are categorical variables and what are the two types?
Variables that are categories (favourite colour, university major, preferred mode of transportation)
2 types:
1) Nominal
2) Ordinal
What is a quantitative variable and what are the two types?
Variables that are numbers and behave like numbers (height, weight, salary)
2 types:
1) Interval
2) Ratio
What is a nominal variable?
A type of categorical variable that names something but has no meaningful order. For example, eye colour
What is an ordinal variable?
A categorical variable that classifies data into ordered categories but does not convey the degree or magnitude of difference between the categories. For example, pain scales; the difference between moderate and severe pain is not conveyed
What is an interval variable?
A quantitative variable that measures the intervals between values on a scale. It has an arbitrary zero point, so the ratios between values are not
meaningful. Valid operations include addition and subtraction. For example, temperature
What is a ratio variable?
A type of quantitative variable that has a meaningful zero and equal distance between points. Valid operations include addition, subtraction, multiplication, division, powers, logs, and everything else. For example, kinetic energy
What are negative controls in experimental studies?
Procedures not expected to produce results (a placebo)
What are positive controls in experimental studies?
Procedures with well-understood, usually positive effects (e.g., a previously established vaccine)
What are the 3 stages of experimental design?
1) Recruit
2) Sort
3) Repeat
What is a population vs a sample?
A population is the group that we wish to know about. A sample is the group of participants from the population
What are the 4 types of recruitment methods?
1) Simple random sampling
2) Cluster random sampling
3) Stratified sampling
4) Snowball sampling
What is simple random sampling?
Selecting participants purely randomly
What is cluster random sampling?
Selecting clusters (e.g., townships) of participants randomly -This is a convenient method
What is stratified sampling?
Sorting a population into subpopulations (demographics), then randomly sample proportionately from those subpopulations
What is snowball sampling?
Used when the sample is hard to recruit. Initial participants are found, then they refer researchers to other participants
What is a block design?
Where participants are grouped first (e.g., by gender), then randomized making sure equal numbers of each group go into each of the experimental groups
What is a matched pair design?
Each participant has a partner that matches them in a target variable(s) that the experimenters think might be meaningful (e.g., age, geneder, smoking status). Then, those partners are separated into different groups
What is blinding and describe the 3 types
Blinding refers to who knows whether a participant is in the experimental or control group.
1) Unblinded: everyone knows which participants are in which group
2) Single-blind: the participant doesn’t know which group they’re in, but the researcher does
3) Double-blind: neither the participant nor the researcher knows which group the participant is in; only a third party (a research coordinator) knows this
What are observational studies?
Studies observing variables without manipulating them
What are correlational studies?
Studies that seek to correlate two or more variables. For example, observational studies
What are cross-sectional studies?
Studies that study a sample(s) at one point in time
What are longitudinal studies?
Studies that study one sample over a length of time
What are quasi-experimental studies?
Studies where there are indepedent and dependent variables, but no random assignment
What are case-control studies?
Studies employed when studying something rare. Studies two groups: one that has a certain ouctome (e.g., skin cancer) and one that doesn’t. Then, the values of another variable (e.g., sun exposure) are compared between groups
What are case studies?
Studies that dig deeply into a few cases. These are often used in the development of models, which are often tested with other methods
What is a mixed methods study?
An approach that combines multiple methods in a way that paints a more complete picture than any of the individual methods
Populations have _________, samples have _________
Paramters, statistics
