03 - Chapter 3: Measurements, Mistakes, and Misunderstandings

A set of flashcards covering key concepts from Chapter 3: Measurements, Mistakes, and Misunderstandings, focusing on types of variables, survey bias, measurement reliability/validity, and related concepts.

Variables

Information (data) collected from subjects.

Categorical (qualitative) variables

Divide subjects into groups; do not support meaningful arithmetic.

Measurement (quantitative) variables

Responses are meaningful numeric values; allow arithmetic.

Nominal variables

Categorical with no natural ordering (e.g., major, ethnicity, home state).

Ordinal variables

Categorical with natural ordering (e.g., Freshman–Senior; Likert scales).

Interval variables

Differences are consistent; ratios are not; zero is arbitrary (e.g., temperature in Fahrenheit).

Fahrenheit–Celsius conversion

C = (F – 32) × 5/9; demonstrates that zero and ratios may be arbitrary in interval variables.

Ratio variables

Ratios are meaningful; zero means none (e.g., age, height, distance, count of chairs).

Discrete variables

Countable number of possible values.

Continuous variables

May take any value within an interval; every fraction is possible.

Example: Political party membership

Categorical (nominal).

Example: Number of credit hours

Discrete measurement (also ratio).

Example: Distance from home to campus

Continuous measurement (also ratio).

Example: Age

Continuous measurement (also ratio).

Example: Age at most recent birthday

Discrete measurement (also ratio).

Survey bias

Systematic prejudice in one direction; question wording can bias responses.

Biased survey question (A)

Should the State of Missouri increase funding for the University of Missouri? (leading wording).

Biased survey question (B)

To ensure a brighter future for our state, should the State of Missouri increase funding for the University of Missouri? (framing bias).

Anchoring

Information in one question may influence responses to a later question; most likely when related questions are near each other.

Open vs closed form questions

Open questions (essays/short answers) vs closed questions (multiple-choice).

Survey problem: Asking the uninformed

Respondents may answer about topics they know little about; include a 'no opinion' option.

Validity vs Reliability

Validity: measures what it claims to measure. Reliability: yields similar results on repeated trials.

Confidentiality vs Anonymity

Confidential: identities known but protected. Anonymous: identities unknown.

Measurement issues: Variability

Discrepancies between repeated measurements.

Natural variability

Variability that cannot be explained or predicted.