You are on page 1of 20

Estimation of Various Population

Parameters
Point Estimation and Confidence
Intervals

Dr. M. H. Rahbar
Professor of Biostatistics
Department of Epidemiology
Director, Data Coordinating Center
College of Human Medicine
Michigan State University
Important Statistical Terms
• Population: A set which includes all
measurements of interest to the researcher

• Sample: Any subset of the population

• Parameter of interest: Characteristics of interest


in a population are called population parameters
• (e.g, mean, standard deviation, proportion)
Inferential Statistics

• Estimation includes point and interval


estimation of population parameters
• Mean = 
• Median
• Standard deviation = 
• Proportion = p
• Incidence
Central Limit Theorem

• When large samples, (n>30), are selected


from a population with mean  and
standard deviation , then
• the sampling distribution of the sample
mean is approximately normal with the
same mean ,

and the standard error, SE= n
Estimation of Population Mean

A 95% confidence interval for  based on a


large sample, when n>30, is approximately,

x  2( )
n
Estimation of Population Mean
A (1- )100% confidence interval for 
based on a large sample, when n>30, is
approximately,

x  Z / 2 ( S )
n
  0.05, Z / 2  1.96
  0.10, Z / 2  1.28
Example # 1
Suppose we want to estimate the average birth-
weight of children born to non-smoking women in
community X. Assume the standard deviation of
birth-weights is 0.5 Kg and the sample mean birth-
weight is 3.0 Kg. Find a 95% confidence interval
for .
3.0  2(0.5 )
100

3.0  0.1

(2.9, 3.1)
Interpretation of CI’s

What do we mean by (2.9 Kg, 3.1 Kg) is a 95%


CI for the mean birth-weight of children born
to non-smoking women in community X?

Does interval (2.9 , 3.1) contain ?


Confidence Interval for P
Suppose we want to estimate the prevalence of
hypertension among adults in community X.
Assume in a random sample of 100 adults from
this community the prevalence of hypertension
is 0.50. Find a 95% confidence interval for P.

0.50  2( (0.50)(1  0.50) 100 )


0.50  0.10
(0.40, 0.60)
Probability vs Odds of an Event
• An experiment is
repeated for a large
number of times. • Odds of an event E
Then the probability =P(E)/{1-P(E)}.
of occurrence of an
event E is the Example,
percentage of the time
If P(E)=0.40, then
that that particular
event has occurred in Odds of E=2/3.
repeated experiments.
• Suppose P(E)=0.40
In the following contingency table estimate the
proportion and odds of drinkers among those
who develop Lung Cancer and those without the
disease?

Lung Cancer Total


Case Control
Drinker Yes A=33 B=27 60
No C=1667 D= 2273 3940
P1=33/1700 P2=27/2300
Odds1=33/1667 Odds2=27/2273
Odds Ratio as a Measure of
Association
• Odds Ratio=ad/bc
• QUESTION: Is
there a difference • OR=1.67
in the proportion • This means that
of drinkers among the odds of
those who develop drinking among
Lung Cancer and cases is 67% more
those without the than the odds of
disease? drinking among
controls.
Relative Risk as Measure of
Association
Risk among exposed
• QUESTION:
= a/(a+b)
Are persons
exposed to Risk among non-exposed
“factor X” more = c/(c+d)
likely to develop
the disease than Relative Risk
those not so
exposed? = Risk among exposed
relative to risk among
non-exposed
QUESTION: Are persons exposed to “factor X” more
likely to develop the disease than those not so exposed?
Disease Total
Yes No
Factor X Yes A=33 B=27 60 (fixed)
No C=1667 D=2273 3940 (fixed)
Risk among exposed
= a/(a+b)=33/60=0.55
Risk among non-exposed
= c/(c+d)=1667/3940=0.42

Relative Risk
= Risk among exposed relative to risk among non-
exposed RR=1.31
95% Confidence Intervals for Odds
Ratios and Relative Risk
• NOTE: • If the 95% CI for OR
• Confidence does not include one
Intervals for Odds then we conclude a
Ratios and Relative difference in the odds
Risk are not of drinking between the
symmetrical two groups
• Odds Ratio=ad/bc
• OR=1.67 • We use Statistical
• E.g., 95% CI for OR packages to calculate
(1.2, 2.9) the CI’s for OR or RR
QUESTION: Estimate the difference
between the proportions of drinkers
among Lung Cancer cases and controls!

Lung Cancer Total


Yes No
Drinker Yes A=33 B=27 60
No C=1667 D= 2273 3940

P1- P2 = 33/1700 – 27/2300


QUESTION: Estimate the difference between
the mean blood levels of cases and controls?

G ro u p 1 G ro u p 2
D is e a s e N o D is e a s e
M ean B P M ean B P
Factors influencing the
width of the CI’s

Sample size
Variance
Confidence level = (1-)100%
Width of the CI’s

If Sample size =n , then Width 


If Variance , then Width 
If Confidence level  , then Width 
Choose ‘n’ to achieve the
desired width for the CI’s

You might also like