Data structures
What is a table?
A two-dimensional object with a finite number of rows and a finite number of columns
What is the main difference between categorical and numerical variables?
For numerical variables, we can measure the size of difference between individuals using mathematical operations. For categorical variables, we cannot.
How are units of observation and units analysis similar?
Both are some kind of concrete object, abstract object (e.g., a group), event, or process.
Where do research questions come from?
Shared gaps in knowledge in a community of research (even the experts don't know); requires familiarity with one's discipline or area of research
What is the main difference between observational and experimental study designs?
Experimental study designs deliberately intervene on some of the focal variables (specifically, explanatory variables)
Observational study designs do not deliberately alter any of the focal variables; they attempt to observe focal variables as they would exist even if we were not observing them
What do rows of data tables represent?
Individuals (aka subjects, records, etc.)
What kind of variable can we think of as an answer to a "yes/no" question?
Indicator variables (aka binary or dichotomous variables)
How are units of observation and units analysis different?
Unit of analysis: the kind of object/event/process whose variability we wish to understand.
Unit of observation: the kind of object/event/process whose variability we directly observe and record in a data set.
Where do hypotheses come from?
Hypotheses are deduced from theories that are credible in particular areas of research (Mad Lips metaphor); requires familiarity with theory in one's discipline or area of research
What are the three principles of experiments, and what are two other characteristics of experiments that improve their quality?
(1) Treatment and comparison between treatment arms
(2) Randomized assignment to treatment arms
(3) Equal treatment
(A) Blocking
(B) Replication
Every individual in a data table is the same kind of object, event, or process. What is term we give to this object, event, or process?
Unit of observation
Which kinds of variables can we use to answer "greater than, less than" questions? Which can we not?
Ordinal and numerical, but NOT nominal
What is a source population?
The population from which we draw our samples
How does the bait-and-switch metaphor apply to the research cycle?
Our research questions, hypotheses, and inferences focus on our target/study populations, but our research design, data collection, and data description focus on samples instead. If the sample isn't representative, our inferences about populations won't be valid.
What is confounding?
Anything that interferes with or prevents causal inference
What is a variable?
A characteristic, attribute, or trait that can (but need not) vary between individuals in a data table
Why do we often confuse continuous numerical variables with discrete ones?
Practically speaking, we must round continuous variables, making them appear as if they skip over values
What is a target or study population?
A population whose patterns of variability we wish to understand. Often, defined with geographic, temporal, or other conditional boundaries.
In what ways does the research cycle loop back on itself?
If our research supports our hypothesis, this often raises new research questions.
If our research does not support our hypothesis, we develop new hypotheses to answer our research questions.
If our research proves inconclusive, we return to research design.
What is a lurking variable?
A common cause of explanatory and response variables of interest that has not been observed
What is a unique identifier?
A unique combination of numbers and/or letters that helps us to avoid confusing the records of different individuals in a data set
What is operationalization?
The act of deciding how exactly to observe and record a variable.
What is an ecological study, and what is the ecological fallacy?
Ecological study: a study whose unit of observation and analysis is a group rather than an individual.
Ecological fallacy: the mistaken inference that what is true of the pattern of variability in groups is also true of the pattern of variability of individuals within those groups.
How does the the "carts before horses" metaphor apply to the research cycle?
The normal research cycle assumes we design our data collection to collect data in a way that allows us to evaluate our hypotheses.
When we use found data, we don't have this control, so the source population of and the variables constituting found data may not match the needs of our research.
What are the major challenges of descriptive statistics: what are their names, what is challenging about them, and how do we solve them?
Forest vs. Trees: Figuring out how much information about samples to convey in our descriptions; we favor summaries that emphasize broad patterns of variability rather than getting lost in the minutia of information available about our samples
Know your audience: Communicating about variability in our samples to audiences in a way they can understand and meets their needs; identify who our audiences are, what their needs and limitations are regarding statistics, and tailoring our message to these