Flashcards on Nonparametric Statistical Methods

Nonparametric Statistical Methods: A Student's Guide

1 / 45

What is the main idea behind nonparametric tests compared to parametric tests?

They make fewer assumptions about the population distribution (e.g., only continuity), often use ranks, and do not rely on a specific parametric model

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