exm 3 Flashcards

1
Q

what are the properties of a normal distribution

A

continuous, symmetrical around mean, single mode, density highest at mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

what are the 2 rules of thumb for normal distributions

A

2/3 of area under the curve lies within one sd of mean
2. 95% of the area under curve lies within 2 sd deviations of mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

observational studies

A

-easier
-cant be randomized
-can establish association, but cannot determine cause of association

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

experimental studies

A

-permits random assignment
-can determine causes of association
-minimize influence of cofounding variables

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

what are the 3 ways to reduce bias

A

simultaneous control group, randomization, blinding

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

how to reduce effects of sampling error

A

replication, balance, blocking

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

central limit theorem

A

sampling distribution of mean will always be normally distributed as long as sample size is large enough

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

z table is critical value so if 2 *p = <0.05 then we

A

we reject the null

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

how to calculate standard error for a normal population

A

s/sqrt of sample size

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

why should you use t distribution instead of Z

A

important fro confidence intervals or hypothesis due to sampling error in estimate of SE

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

t distribution versus Z

A

t has fatter tails and a lower peak due to sampling error

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

what does normal distribution describe

A

sampling distribution of many estimates

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

assumptions of one-sample t test

A

must be a random sample from population, variable must have a normal distribution in population

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

paired versus two independent sample

A

2 sample- measurements are independent
paired- more powerful- less variation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

how to deal with violation of assumptions

A

ignore, transform, non-parametric tests, permutation tests

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

what are the ways to detect deviations from normality

A

graphical methods- histogram, normal quantile plot
formal test- shapiro wilks

17
Q

why do you need to be cautious about interpreting shapiro-wilk test results for small sample sizes and for large sample sizes

A

if sample size is too low- power is low.
as n gets larger assumption of normality becomes less important

18
Q

when to ignore violations

A

when n>30 and sd is less than 3