Data & Scales
This scale of measurement applies to categorical data that possesses a natural, inherent ranking or ordering, such as "Letter Grade (A, B, C, D, F)" or "Shirt Size (S, M, L)."
What is the Ordinal scale?
This term describes the complete, comprehensive collection of all possible subjects or individuals that a researcher aims to describe or draw conclusions about.
What is a Population?
This non-probability sampling method occurs when an investigator simply gathers data from whatever individuals happen to be closest, easiest, or most convenient to reach.
What is Convenience Sampling?
The natural, unavoidable variation that occurs by pure chance because different random samples pulled from the exact same population will yield slightly different numbers.
What is Sampling Error?
An unmeasured, background variable that correlates with both the explanatory and response variables, muddling a researcher's ability to isolate a true relationship.
What is a Confounding (or Lurking) Variable?
A quantitative variable like "exact tree height" or "race completion time" that can take on any numerical value along a continuous numerical scale, including fractions and decimals.
What is a Continuous variable?
This branch of statistics focuses entirely on organizing, summarizing, and displaying patterns directly observed within a collected sample without making claims about anyone else.
What is Descriptive Statistics?
The gold standard baseline of sampling where every single distinct member of a target population has an exact, equal chance of being selected for the study.
What is a Simple Random Sample (SRS)?
The single most direct action a researcher can take during design to actively reduce the margin of natural sampling error in their study.
What is increasing the sample size (n)?
The essential tool researchers must use when assigning subjects to treatment groups to completely disrupt and neutralize the influence of hidden confounding variables.
What is Random Assignment?
This specific column in a dataset assigned to an individual case uniquely distinguishes it from all other cases (e.g., Social Security Numbers or Flight Numbers) and should never be used to calculate an average.
What is a Unique Identifier?
While a sample statistic changes with every new sample taken, this true numerical value describes a fixed characteristic of the entire target population.
What is a Parameter?
To ensure fair representation of various demographics, a researcher breaks the population into non-overlapping groups (like age groups or socioeconomic brackets) and then takes an SRS from within each group.
What is Stratified Random Sampling?
A systematic discrepancy between a sample estimate and a population parameter caused by a fundamentally flawed selection method that leaves out certain portions of the population.
What is Sampling Bias (or Undercoverage Bias)?
Because hidden confounding variables can almost never be completely ruled out or controlled in an observational study design, statisticians replace the word "causes" with this phrase.
What is "is associated with" (or "is correlated with")?
Eye color, political party affiliation, and zip codes are classified under this scale of measurement because they categorize data without any numerical value or order.
What is the Nominal scale?
The primary objective of advanced statistical work: using data systematically gathered from a smaller sample to make a well-reasoned, science-based guess about a larger population.
What is Inferential Statistics (or Inference)?
A city quality manager divides a grid of neighborhoods into 50 geographic zones. They randomly select 5 entire zones and survey every single household inside those selected zones. Name this sampling method.
What is Cluster Sampling?
This distinct form of bias occurs when individuals are properly selected through a random process but choose not to respond, answer calls, or participate in the survey.
What is Nonresponse Bias?
A medical study monitors the health outcomes of individuals who already chose to follow a strict keto diet on their own accord. This design is classified as this type of study.
What is an Observational Study?
Counting the exact number of cars parked in a parking garage yields this specific type of quantitative variable, which can only take on distinct, isolated whole-number values.
What is a Discrete variable?
A researcher tracks a sample of 200 local voters and finds that 45% plan to vote for a school bond. State whether 45% is a parameter or a statistic, and identify its target group.
What is a statistic, and it describes the sample of 200 local voters.
An online video streaming platform posts an open link on its homepage asking users: "Rate your experience today!" Name this highly volatile sampling method.
What is Voluntary Response (or Self-Selected) Sampling?
An employer interviews workers face-to-face in the main office and asks: "Have you ever used company time to browse personal social media?" This scenario is highly vulnerable to this specific bias.
What is Response Bias (subjects are pressured to give dishonest answers to look good or avoid trouble)?
A study tests a productivity app using only highly competitive, top-tier corporate executives who volunteered to try it. Explain whether the app's true effect on the general population will be overestimated or underestimated.
What is overestimated, because highly motivated corporate executives likely possess superior baseline organizational skills and discipline compared to the general public, exaggerating the app's true utility.