Stats Midterm 1

Front 1

IV- Omitted Variables

Back 1

Events other than treatment that provide alternative explanation for results

Front 2

IV: Trends in Outcomes

Back 2

Observation change as a function of time not because our treatment necessary

Front 3

IV: Misspecified Variance

Back 3

Overstating statistics test because of omitting group error

Front 4

IV: Mismeasurement

Back 4

Changing on measurement that can cause differences

Front 5

IV: Political Economy

Back 5

Endogenous policy changes due to governmental responses to variables associated with past or expected future outcomes. Ex. Government lowering interest in response to an incoming recession

Front 6

IV: Simultaneity

Back 6

Joint determination of outcomes

Ex. Interest rates of 5 and 10 year bonds being correlated- no causal relationship.

Front 7

IV: Selection

Back 7

Assignment of treatment are done in a way that could provide correlation for other reasons

Front 8

IV: Attrition

Back 8

Loss of participants

Front 9

IV: Omitted Interactions

Back 9

Differential trends in treatment and control groups

Front 10

EV: Unrepresentative group

Back 10

The treatment and control don’t represent the entire population of interest

Front 11

EV: Geographic

Back 11

Does not generalize across population in geography

Front 12

EV: History

Back 12

Does not generalize across time

Front 13

Ethics: Beneficence

Back 13

Maximize potential benefits of the study for participants

Front 14

Ethics: Non-Maleficence

Back 14

Minimize the potential for harm or potential risks- physical, psychological, and social

Front 15

Ethics: Autonomy

Back 15

Right to self determination, Informed Consent, Capacity, free power

Front 16

Ethics: Respect for Persons

Back 16

Every person deserves the respect for being an individual

Privacy and confidentiality

No legal protection for privacy

Front 17

Ethics: Vulnerable Populations

Back 17

Federal law protects: children, elderly, mentally disabled, and prisoners

States or universities may protect: minorities, fetuses, AIDS patients, economically disavtaged

Front 18

Ethics: Justice

Back 18

Fairness- will participants gain the benefits if they are taking the risk

Front 19

Central Limit Theorem

Back 19

The sampling of a population with a mean mu will be a normal distribution when n is large.

Front 20

Law of Large Numbers

Back 20

As sample size increase, a random sample producing a mean approaching the population mean

Front 21

Point Estimate property: unbiased

Back 21

Biased parameter will systematically under or over estimate the true value

Variance can be proven to be unbiased when standard deviation is not.

Front 22

Point Estimate property: efficiency

Back 22

When choosing between multiple unbiased estimators (mean or median) or techniques (maximum likelihood, method of moments), we prefer the sampling distribution with the smallest variance

Front 23

Point Estimate property: Consistency

Back 23

Point estimator remains close to the value when the parameter increase its sample size

Front 24

Point Estimate property: sufficiency

Back 24

Uses all the information in a sample

Useful when raw data is too much and statician can confirm that it will not impact the results

Front 25

Point Estimate property: Robust

Back 25

Extent to which an estimation are adversely effected by violations of the underlying assumption

Assuming our data is normal when it is not

Front 26

Binomial Distribution: Description

Back 26

Occurs when

1. There are a fixed number of n observations and k trials

2. The trials are independent

3. The outcome is binary (success/fail)

4. The probability of success is the same for each trial

Ex. A coin toss

Hint 26

4

Front 27

Binomial Distribution: Model

Back 27

Back (Definition)

Hint 27

n = number of trials

k = number of successes

p = probability of a single trial being a success

Front 28

Binomial Distribution: Mean

Back 28

Back (Definition)

Front 29

Binomial Distribution: Variance

Back 29

Back (Definition)

Front 30

Poisson Distribution: Description

Back 30

COUNT DATA

Concerned with the average number of successes per observation

Occurs when:

1. The number of successes are independent

2. The probability of success is the same for all observations of equal size and proportional to the size of the unit

3. The probability of a success that 2 or more successes will occur in a unit approaches zero as the size decreases

Hint 30

3

Front 31

Poisson Distribution: Model

Back 31

Hint 31

Mu = mean number of successes per failure

K = number of successes

Front 32

Poisson Distribution: Mean

Back 32

Front 33

Poisson Distribution: Variance

Back 33

Front 34

Normal Distribution: Description

Back 34

1. Bell shaped

2. Symmetric

3. Mean, median, and mode are all identical

4. Frequencies gradually decreasing at both ends of the curve

5. The mean and the variance completely define the distribution

6. The mean and the variance are independent

Hint 34

6

Front 35

Normal Distribution: Model

Back 35

Front 36

Standard Normal Distribution: Model

Back 36

Front 37

Standard Normal Distribution: Mean

Back 37

Front 38

Standard Normal Distribution: Variance

Back 38

Front 39

Uniform distribution: Description

Back 39

Spreads the probability for being chosen equally across a range from a to b

Front 40

Uniform distribution: model

Back 40

Else 0

Front 41

Uniform distribution: Mean

Back 41

Front 42

Uniform distribution: Variance

Back 42

Front 43

Negative Binomial Distribution

Back 43

How many binary outcomes trials must I conduct to get the keg success. Ex. How many times do I need to flip a coin to get 5 heads

Front 44

Multinomial Distribution

Back 44

Binomial distribution where there are more than 2 possible outcomes

Front 45

Gamma Function

Back 45

A large number of distribution related to the gamma function

Recursive

Contains many asymptotes

Front 46

Chi-Squared Distribution

Back 46

Used in sampling theory

Uses degrees of freedom

A normal distribution times itself

Front 47

Front (Term)

Back 47

Intersection

Where both event may occur

Front 48

Front (Term)

Back 48

Union

Where either event does occur

Ex. A or B

Front 49

Front (Term)

Back 49

B is a subset of A

Front 50

Front (Term)

Back 50

Disjoint or mutually exclusive

Front 51

Means in layman terms

Back 51

What is the probability of B if we know A has occurred

Front 52

Front (Term)

Back 52

Back (Definition)

Front 53

Front (Term)

Back 53

Back (Definition)

Front 54

Front (Term)

Back 54

Back (Definition)

Front 55

Front (Term)

Back 55

Back (Definition)

Front 56

Front (Term)

Back 56

Back (Definition)

Front 57

Front (Term)

Back 57

Back (Definition)