Memory Quiz: Stats

p̂

sample statistic

p

population proportion

po

parameter used in null hypothesis

z

standardized score

z*

upper p-critical value (right tail)

-z*

lower p-critical value (left tail)

Ho

null hypothesis

Ha

alternate hypothesis

n

sample size

N

population size

np̂

Number of successes expected in sample

nq^

Number of failures expected in sample

npo

Number of successes observed

nqo

Number of failures observed

P(A) = P(A/B)

Test for independence

b = r(Sy/Sx)

slope in linear regression

y = a + bx

equation for Least Squares Regression Line (LSRL)

p̂q/n

standard error of sample proportion

z* p̂q/n

Margin of error for 1-Prop Z-Interval

p̂ +- z* p̂q/n

1 Prop Z-Interval

test distribution for z distribution

z-value in 1 Prop Z test

(value-mean)/SD

z-score

How do you find outliers in a set of data? What is the formula for IQR?

IQR: Q3-Q1

lower: Q1- 1.5IQR

upper: Q3 + 1.5IQR

What is the most important difference between an SRS and a stratified random sample?

SRS: Every element of population has some probability of being drawn

stratified: creates homogeneous groups then takes SRS

How do you tell the difference between an observational study and an experiment?

observational: observes only; no treatment

experiment: treatment

Template answer to describe the slope of the LSRL in context:

For every 1 unit increase in x-variable, the y is predicted to increase/decrease according to slope

Template answer to describe the y-intercept of LSRL in context:

The predicted value of y when x-variable = 0

Template answer to describe what R² is:

The % of variability of y that can be explained by the LSRL relating x and y

What are the Z-scores for 90%, 95%, 98%, and 99% confidence intervals

90%: 1.645 98%: 2.326

95%: 1.96 99%: 2.576

What does the p-value mean in context?

Probability the observed statistic or one more extreme will happen by natural variation is true

Conditions for T-Tests:

random

10%

Large Enough (n>30 or graph)

Conditions for z-tests:

random

10%

Large Enough (S/F)

Conditions for x² tests:

What is a Type I error and a Type II error?

Type I: reject null when true

Type II: fail to reject null when false

What is meant by the Power of the Test?

Probability of correctly rejecting null when false

What are the 2 most effective ways to increase the power of a test?

-Increase sample size

- Increase significance level (alpha)

Formulas for ALL confidence intervals

statistic +- critical value x (SX of statistic)

Formulas to find the test statistic for ALL tests:

(statistic-parameter)/(SX of statistic)

What is ALWAYS the null hypothesis for a matched pairs t-test?

Mdiff= 0

What are the null hypotheses for the 3 different x² tests?

GOF: Ho= sample data match population

Independence: Ho= 2 variables are independent

Homogeneity: Ho= There is NO difference in group proportions