stats present sample size
TRANSCRIPT
-
7/29/2019 Stats Present Sample Size
1/38
Sample Size and PowerCalculations
Marcia A. Ciol
04/09/08
-
7/29/2019 Stats Present Sample Size
2/38
What resources do I need? How long will it take to conduct the study?
I need 50 participants in my study
About 5 individuals per year will be enrolled Therefore, it will take 10 years to finish the study
How much money do I need? I will follow a cohort of 500 individuals
A lab test that costs US$100 will be conducted foreach person
Therefore, I will need US$50,000 just for labtests
-
7/29/2019 Stats Present Sample Size
3/38
Am I going to reach my
objective? I have 2 years to finish my thesis, of
which one year is for data collection
I think I can get data on 50 people inthat year
Is 50 a sufficient number of people to
test my hypothesis with the significancelevel I want?
-
7/29/2019 Stats Present Sample Size
4/38
Why to calculate sample size
and power? To show that under certain conditions, the
hypothesis test has a good chance ofshowing a desired difference (if it exists)
To show to the funding agency that the studyhas a reasonable chance to obtain aconclusive result
To show that the necessary resources(human, monetary, time) will be minimizedand well utilized
-
7/29/2019 Stats Present Sample Size
5/38
What do I need to know to
calculate sample size? Most Important: sample size calculation is
an educated guess
It is appropriate for studies involvinghypothesis testing
There is no magic involved; only statisticaland mathematical logic and some algebra
Researchers need to know something aboutwhat they are measuring and how it variesin the population of interest
-
7/29/2019 Stats Present Sample Size
6/38
Factors related to the
sample size Population factor (cannot be controlled by
researcher)
Characteristics of the study design
Quantities related to the research question(defined by the researcher)
-
7/29/2019 Stats Present Sample Size
7/38
Where do we get this
knowledge? Previous published studies
Pilot studies
If information is lacking, there is nogood way to calculate the sample size!
-
7/29/2019 Stats Present Sample Size
8/38
Population factor Variance of the measure (outcome)
within the population
-
7/29/2019 Stats Present Sample Size
9/38
0.00
0.02
0.04
0.06
0.08
0.10
x-20 -10 0 10 20 30
-
7/29/2019 Stats Present Sample Size
10/38
0.00
0.02
0.04
0.06
0.08
0.10
x-20 -10 0 10 20 30
-
7/29/2019 Stats Present Sample Size
11/38
Study DesignType of response variable or outcome
Number of groups to be compared
Specific study designType of statistical analysis
In conjunction with the research question, thetype of outcome and study design willdetermine the statistical method of analysis
-
7/29/2019 Stats Present Sample Size
12/38
Quantities related to the research
question (defined by the researcher)
= Probability of rejecting H0 when H0 is true is called significance level of the test = Probability of not rejecting H0 when H0 isfalse
1- is called statistical powerof the test
-
7/29/2019 Stats Present Sample Size
13/38
Quantities related to the research
question (defined by the researcher) Size of the measure of interest to be detected
Difference between two or more means
Odds ratio Change in R2, etc
The magnitude of these values depend on
the research question and objective of thestudy (for example, clinical relevance)
-
7/29/2019 Stats Present Sample Size
14/38
Example: test of difference of
means in two populations Researcherfixes probabilities of type I and II
errors
Prob (type I error) = Prob (reject H0 when H0 is true) = Smaller error greater precision need more information
need larger sample size
Prob (type II error) = Prob (dont reject H0 when H0 isfalse) =
Power =1- More power smaller error need larger sample size
-
7/29/2019 Stats Present Sample Size
15/38
Example: test of difference of
means in two populations The equation for sample size is derived from
the equation for the statistical test
In a t-test the equation for the test ist = (x1 - x2) - (m1- m2)
(s12n + s2
2n)12
The derived equation for sample size is
n = (z1-/2+ z1-)2(s1
2+ s22)
(m1- m2)2
-
7/29/2019 Stats Present Sample Size
16/38
Using PASS: t-test example Question: does exercise help to decrease
body weight?
Study design: participants will be randomizedinto two groups (exercise and control)
Outcome: change in weight
Want to detect: a change of at least 15
pounds Known: from past studies, the standard
deviation varies between 10 and 15 pounds.
-
7/29/2019 Stats Present Sample Size
17/38
-
7/29/2019 Stats Present Sample Size
18/38
-
7/29/2019 Stats Present Sample Size
19/38
-
7/29/2019 Stats Present Sample Size
20/38
-
7/29/2019 Stats Present Sample Size
21/38
Example: One-way ANOVA
Number of Groups: 4Hypothesized means: 35, 20, 25, 18 (possibly
from a pilot study)Sample size pattern: same number in eachgroupSD of subjects: 18 (from a previous study)
= 0.01 and 0.05Find power for sample sizes from 5 to 30 pergroup (increments of 5)
-
7/29/2019 Stats Present Sample Size
22/38
-
7/29/2019 Stats Present Sample Size
23/38
-
7/29/2019 Stats Present Sample Size
24/38
-
7/29/2019 Stats Present Sample Size
25/38
Example: Linear Multiple
RegressionResearch Question: is depression score an importantfactor in explaining pain ratings, after adjusting for ageand sex?
Statistical question: does adding depression scoreincrease the explained variation of pain ratings, in alinear regression model that already has age and sex in
it and has R2 =.2?
Suppose I may have sample sizes of 20, 30, 50, 70,and 100. What is the minimum R2 change I can detect
with power .8?
-
7/29/2019 Stats Present Sample Size
26/38
-
7/29/2019 Stats Present Sample Size
27/38
-
7/29/2019 Stats Present Sample Size
28/38
Other Types of Hypothesis
Tests Different methods of data analysis
require different input for sample size
calculations
-
7/29/2019 Stats Present Sample Size
29/38
Cox Regression (Survival
analysis)
-
7/29/2019 Stats Present Sample Size
30/38
Logistic Regression
-
7/29/2019 Stats Present Sample Size
31/38
Repeated measures
-
7/29/2019 Stats Present Sample Size
32/38
Simple designs may not
require complex calculations Read chapter 2 ofStatistical Rules of
Thumb, by Gerald van Belle (2002,
John Wiley and Sons)
Using specialized software is useful if
many calculations will be performed
-
7/29/2019 Stats Present Sample Size
33/38
Important to remember Pilot studies do not need sample size
calculation!!!
There is no point in doing power analysis afterthe study is done
Sample size is an educated guess, and itworks only if: The study samples comes from the same or similar
populations to the pilot study populations
The population of interest is not changing over time
The difference or association being studied exists
-
7/29/2019 Stats Present Sample Size
34/38
How about Effect Size? Most common definition
E = m1
- m2
spooled
If we change de value of E, how do weknow what we changed in the formula?
-
7/29/2019 Stats Present Sample Size
35/38
Some situations I have
encountered Question: How many more people do I need to
enroll in the study (already in progress) to show
statistical significance? Answer: It depends If the two populations
have the same mean, increasing the samplesize will not help!
Since when is the objective of a study to find astatistically significant result??
-
7/29/2019 Stats Present Sample Size
36/38
Some situations I have
encountered Researcher is interested in outcome A, which
differs very little for two treatments
Sample size needed is around 3000!!
Researchers changes the outcome to B, wheresample size is smaller
B does not answer the researchers question
and he needs to accept that his new treatmentis not really different (clinically speaking) fromthe already existent treatment
-
7/29/2019 Stats Present Sample Size
37/38
Some situations I have
encountered Researcher is interested in comparing
two groups regarding prediction of
outcome A by using a regressionanalysis (using several variables)
He uses the only available formula fromhis statistical book (for a t-test)
Wrong! He should find a software thatcan calculate the sample sizeappropriately
-
7/29/2019 Stats Present Sample Size
38/38
Summary Define research question well
Consider study design, type of response
variable, and type of data analysis Decide on the type of difference or change you
want to detect (make sure it answers yourresearch question)
Choose and Use appropriate equation sample size
calculation