Statistical tests

Question 1

Compares two independent samples

Eg: To compare girls’ heights with boys’ heights

Answer 1

Parametric: Two-sample (unpaired) t test

Nonparametric: Mann-Whitney U test

Question 2

Compares two sets of observations on a single sample

E.g To compare weight of infants before and after a feeding

Answer 2

Parametric: One-sample (paired) t test

Nonparametric: Wilcoxon matched pairs test

Question 3

Compares three or more sets of observations made on a single sample

e.g To determine whether plasma glucose level is higher 1 hr, 2 hr, or 3 hr after a meal

Answer 3

Parametric: One-way analysis of variance ( F test) using total sum of squares

Nonparametric: Kruskal-Wallis analysis of variance by ranks

Question 4

Compares three or more sets of observations made on a single sample, but tests the influence (and interaction) of two different variables

eg. To determine whether plasma glucose level is higher 1 hr, 2 hr, or 3 hr after a meal, looking at difference in males vs females

Answer 4

Parametric: Two-way analysis of variance (ANOVA)

Nonparametric: Two-way analysis of variance by ranks

Question 5

Tests the null hypothesis that the distribution of a categorical variable is the same in two (or more) independent samples

E.g. To assess whether acceptance into medical school is more likely if the applicant was born in Britain

Answer 5

Parametric: χ2
(chi square) test

Nonparametric: Fisher exact test

Question 6

Assesses the strength of the straight-line association between two continuous variables

e.g. To assess whether and to what extent plasma HbA1C concentration is related to plasma triglyceride concentration in patients with diabetes

Answer 6

Parametric: Product moment correlation coefficient (Pearson r )

Nonparametric: Spearman rank correlation coefficient (rσ)

Question 7

Describes the numeric relation between two quantitative variables, allowing one value to be predicted from the other

E.g To see how peak expiratory flow rate varies with height

Answer 7

Parametric: Regression by least squares method

Nonparametric: Nonparametric regression (various tests)

Question 8

Describes the numerical relationship between a dependent variable and several predictor variables (covariates)

e.g To determine whether and to what extent a person’s age, body fat, and sodium intake determine his or her blood pressure

Answer 8

Parametric: Multiple regression by least squares method

Nonparametric: Nonparametric regression (various tests)