Stat 350 exam 2 Cheat Sheet.

Cheat Sheet > Stat 350 exam 2 Cheat Sheet.

Purdue University STAT 350 Harsh Patel Chao 2:30 Section 60/61 Types of studies… Experimental: controlling variables Observational: just observing them, no control Sampling methods… Simple Random Sample (SRS): everything has a chance of being picked Stratified Random Sample: divide into groups then SR ...[Show More]

Purdue University
STAT 350
Harsh Patel

Chao 2:30

Section 60/61

Types of studies…

Experimental: controlling variables

Observational: just observing them, no control

Sampling methods…

Simple Random Sample (SRS): everything has a chance of being picked

Stratified Random Sample: divide into groups then SRS for each group

Matched Pair: experimental unit is matched with another one

Managing Bias and Variability…

To reduce bias: use random sampling

To reduce variability: use SRS and a larger sample

Parameters of sampling distributions:

Confidence interval: CI conclusion:

We are 95% (100C%) confident that the population (true)

mean of […] is covered by the interval (a,b) [or between a

and b].

Choosing sample size:

Confidence Bounds:

***Pay attention to Z_A***

Z vs T distributions:

**use Z if you know population STD

**use T if you know sample population STD

Summary of CI – t dist.

N < 15: distribution should be normal

15 < N < 40: mild skewness is ok

N > 40: procedure is valid

TYPES OF ERRORS:

Type I = reject H(null), but H(null)

is really true

Type II = fail to reject H(null), but

H(null) is actually false

CI and HT on two samples Independent vs paired samples (2-sample)

2-sample Independent: the objects in sample 1 have

no effect/relation to those in sample 2

2-sample Paired: dataset is similar, sample 1 has

effect/relation on sample 2

2 – sample independent (Z) HT

2 – sample independent (Z) CI

2-sample independent (T) HT

2 – sample independent (T) CI

Assumptions for

ANOVA:

1). K independent

SRSs

2). Normal Dist,

with unknown mew

3). Same variance,

(sigma^2)*below*

S_max

---------- <= 2

S_min

Association doesn’t mean causation

Q: A small town will rezone is the average property tax is > $25k. A Type

1 error is… the inference says that the tax is more than $25k, but it is

really less than $25k

Q: 2-sample independent Z test can be used it both population variances

are known…True

Tukey’s multiple comparison

*the CI is significant if it d