Estimation and confidence, significance test and p values

Question 1

How to deal with uncertainty

Answer 1

confidence interval, statistical significance test (if testing hypothesis)

Question 2

What is standard error of mean and what does it do

Answer 2

standard deviation of sample means, provide estimate of precision of sample mean. SD/squareroot of n. If SE decreases, precision greater for larger samples

Question 3

How to work out confidence interval

Answer 3

mean-multiplierStandard error of mean to mean+multiplierStandard error of mean
Multiplier=1.64 for 90% CI, 1.96 for 95% Ci, 2.58 for 99% CI

Question 4

Standard error of proportion, how calculated and when used

Answer 4

squareroot (p(1-p)/n), large sample size

Question 5

95% confidencce interval for a proportion from large sample

Answer 5

p-1.96standard error of proportion to p+1.96*standard error of proportion

Question 6

What is classified as a large sample

Answer 6

Mean: sample size of 100 is considered large of sample mean
Proportions: for sample proportion, sample size considered large if r and n-r are both greater than 5. If not, binomial confidence interval can be calculated

Question 7

2 tailed and 1 tailed test

Answer 7

2-alternative hypothesis is general and allows the difference to be in either direction
1-doesn’t distinguish between no difference and harmful effect.
2 sided tests should always be used unless justification

Question 8

Error in significance testing

Answer 8

type 1: getting a significant result in sample when null hypothesis is in the underlying population (false significant)
type 2-getting a non significant result in sample when null hypothesis is false in the underlying population

Question 9

What is statistical significance (p value) influenced by

Answer 9

size of real effect in population sampled, smaple size, variability of measure involved

Question 10

Does statistical significance always link to clinical significancee

Answer 10

no