Sampling Distributions
Statistical Inference
By,
Yaseen Ahmed Meenai
Faculty, DMS-FCS-IBA
Sampling and Sample Estimates
• Sampling is the
Procedure of getting a
Sample from a Population
• Sample results are ESTIMATES [ Statistic(s) ] while
Population results are PARAMETERS
• For e.g. (a sample mean) is an Estimate or
Statistic and “” (a population mean) is a
Parameter.
Repeated Sampling
• Repeated Sampling can produce several
Samples from a single population.
POPULATION
Sample1 Sample2 Sample3 ... Sample m
Mean1 Mean2 Mean3 …….. Mean m
Sample Mean (Estimate) itself became a DATA
Sampling Distribution
• Distribution of Sample Estimate is called a
Sampling distribution…
• For e.g. Sampling Distribution of Mean,
Sampling Distribution of Variance etc.
Class Work
• Determine the Sampling Distribution of
Sample Means data gathered in the class
• Observed its SHAPE and comments.
Data Tabulation (Grouping Exercise)
Step # 01: Class Intervals Frequency
Finding the range Min ______ Min+h
Range = Max. – Min
Step # 02: Min+h ___ Min+h+h
Finding the number of classes Min+h+h _______ ….
No. of classes = 1 + 3.3 log(n)
…
Step # 03:
Finding the width or height (h) … ____________ Max
h = Range/No. of classes
Mean and Variance of the
Sampling Distribution of Mean
Class Intervals fi xi fi xi fi (xi-mean)2
Min ______ Min+h f1 x1 f1 x1 ..
Min+h ___ Min+h+h f2 x2 f2 x2 ..
Min+h+h _______ …. f3 x3 … ..
… .. .. … ..
… ____________ Max fm xm … ..
f x fx f (x
i i -mean)2
The Central Limit Theorem
• If ‘X’ is a random variable follows a Normal
Distribution with mean= and standard
deviation=; we can show it as:
X ~ N(,)
• Similarly, distribution of sample mean ഥ
𝒙 also
Normal with mean= and standard error =
/n (where ‘n’ is the sample size)
~ N( , /n)
Empirical Proof of CLT
• Obtain the Population Mean () and the
Population Standard Deviation () from the given
population data:
10 12 3 5 7 8 13 5 2 4 6
• Now Obtain the Mean 𝐸(𝑋) ത and standard
deviation of Means’ data
• Now compare : ത
= 𝐸(𝑋)
=
Normal Distribution v/s Central Limit
Theorem
• The only difference b/w Normal Distribution
case and C.L.T. is the standard error.
• In Normal Distribution, the standard error is
‘’ and in CLT, it’s ‘/n’ (which is
comparatively lesser)
• What could be the impact
C.L.T.
Normal
on a Normal Curve when Distribution
we minimize the standard error???
A Normal Case v/s Central Limit
Theorem
Normal Distribution Exercise:
Q) Daily Profit of a Grocery store owner follows a
Normal Distribution with mean=64 hundred
Rs/- and the standard deviation=12 hundred
Rs/-.
If any day is being selected at random;
determine the probability that his profit on that
day will be less than 60 hundred Rs/-?
P(X<60) = ?
P = P(Z < - 0.33) = 37%
A Normal Case v/s Central Limit
Theorem
Central Limit Theorem Exercise:
Q) Daily Profit of a Grocery store owner follows a
Normal Distribution with mean=64 hundred Rs/- and
the standard deviation=12 hundred Rs/-.
in 6 working days; determine the probability that his
average profit will be less than 60 hundred Rs/-?
P( 𝑋ത < 60) = ?
P P(Z < -0.82) = 21%
A Normal Case v/s Central Limit
Theorem
Normal Distribution Exercise:
P(X<60) = ?
𝑋−𝜇 60−64
P <
𝜎 12
P( <60) = 21%
P(Z < - 0.33) = 37%
Central Limit Theorem Exercise:
P(X<60)=37 %
P(𝑋ത < 60) = ?
P = 21%
A Normal Case v/s Central Limit
Theorem
Central Limit Theorem (Contd…)
If we select a sample 6 working days; determine
the Probability that his Average Profit will fall
within the range of 60 to 70 hundred Rs/-?
P(60 < < 70) = ??