Term 2 Week 1: Confidence intervals

Question 1

Confidence interravla take into account the _____ associated with statsitcal ______ by butting CL/B around PE

Answer 1

uncertainty
results
confidence limits/ bounds
point estimates

Question 2

What is a poitn estimate

taking the ____ found in your ______ as a signle best _____ of the pp

they are subject to SE

Answer 2

reuslt
sample
popuation parameter

sampling error

Question 3

Why use CIs?

sample, chance

what do CIs set? to amount of error our measures are subject to

Answer 3

randomly sampling from a popualtion or randomingly assinging to experiement groups = risk of sampling error

ppt by chance more or less representative of the population we are studing

therefore the statistical results are subject to error

probable limits

Question 4

What infleuces the size of error? (SD)

Answer 4

Study design

taking bigger samples reduces teh CIS haves a more precise estimate

Question 5

What do 95% CIs mean?

range of ____ that we can be 95% certain contains the ______ ________ (e.g. ________ _____)

Answer 5

values
population parameter
population mean

Question 6

What is central limit theorem?

Answer 6

the greater the sample size the comler the sampling distrigution of the mean will ressemble the normal distribution

If the popualtion distribution is NOT normal the smapling distribution will not be nirmaly espcially if the sample size is small

n> 30 sampling distriution of the mean will be close to normal

Question 7

What does the central limit theorem Imply?

Answer 7

The SD of the M for a sufficiently large sample is

a normal distribution
the mean = the popualtion (mu) mean
a standard devition (SD of the pop (sigma)/ sqaure root of the sample size (n))

Question 8

The SD of the smapling distribution is also know as the XX

because it is the average X we make when we use sample mean to estimate popualtion mean

Answer 8

standard error

error

Question 9

Estimation and confidence intervals

what is a CI?

Answer 9

gives a range of values that is more liekly to contain the true poit estimate alone = interval estimation

95% of sample means fall within 1.96 SE either side of the mean
(see card 7 for formula)

Question 10

Confidence intervals allow us to specify a pm of error around an estimate

a wide CI signifies X uncertainty about the X

Answer 10

probable margin

greater estimate

Question 11

Cis only take into about SE (or E due to RA)

Not

• NRB
• PSS

Shouldn’t use CIs when

Answer 11

sampling error
or error due to random allocation

non response bias
poor sampling strategy

sampling isn’t random

Question 12

Hypothesis testing

H0
H1

Answer 12

Ho = null H
H1= alternative H (direction or non directional)

Question 13

If H0 true….

95% of z values lie (-1.96 & +1.96)

5% of z values < or > +/- 1.96

Answer 13

Standard normal distribution

mean = 0 SD = 1

1.96 critical values

z 1.96 = critcal region/ region of rejection

Question 14

Tpoes of errors in H testing

Answer 14

alpha - acpection H1 when there is no affect

Beta error rejection the alternative hypothesis where there is an affect

Question 15

ONe tailed hypothesis

Answer 15

if strong theorectical reason to H a particualr direction will be found

Larger or smaller

Question 16

two tailed hypothesis

Answer 16

non direction H - there will be a difference

Question 17

What to remember when HTesting

Answer 17

before collecting tdata must:

clearly specify H
determien the level of significane

decider weren one tail or 2 tail

H0 until proven H1