Flashcards on Nonparametric Statistical Methods
Nonparametric Statistical Methods: A Student's Guide
Tap to flip · Swipe to navigate
Nonparametric rank tests
45 cards
Card 1
Question: What is the main idea behind nonparametric tests compared to parametric tests?
Answer: They make fewer assumptions about the population distribution (e.g., only continuity), often use ranks, and do not rely on a specific parametric model
Card 2
Question: When are nonparametric tests particularly useful?
Answer: When sample sizes are small, distributional assumptions (like normality) are questionable, measurements are ordinal or ranks, or when robustness to me
Card 3
Question: List key assumptions that parametric tests (like the paired t-test) rely on which nonparametric tests relax.
Answer: Parametric assumptions include normally distributed differences, interval-scale measurement, independence of observations, and (for two-sample tests)
Card 4
Question: How do nonparametric tests typically handle data values?
Answer: They often convert data to ranks and base test statistics on those ranks instead of the raw measurements.
Card 5
Question: What is a potential consequence if parametric test assumptions are violated?
Answer: The test may be invalid, leading to an actual Type I error probability greater than the nominal α (increased chance of incorrect rejection).
Card 6
Question: What are general steps followed when performing a nonparametric hypothesis test?
Answer: 1. State the null hypothesis. 2. Calculate the test statistic (often based on ranks). 3. Apply a decision rule to reject or not reject based on the st
Card 7
Question: Why might nonparametric tests be considered easier to learn than parametric tests?
Answer: Because they require fewer assumptions about underlying distributions and often use simpler rank-based procedures.
Card 8
Question: In what measurement situations are nonparametric tests especially appropriate?
Answer: When data are inherently rank-ordered or measured on an ordinal/nominal scale, or when interval-scale measurement cannot be assumed.
Card 9
Question: What is the basic idea behind nonparametric rank tests?
Answer: They replace original measurements with ranks and use rank-based statistics to test hypotheses without assuming parametric distributions for the data.
Card 10
Question: Why use rank tests instead of completely randomized designs when subjects are highly variable?
Answer: Rank tests (and blocking) reduce the impact of large subject-to-subject variability that can blur treatment differences in completely randomized desig