Introduction To Econometrics: Lecture 2: Review of Basic Statistics & Randomized Controlled Trials (RCTS)

Introduction to Econometrics
Lecture 2: Review of Basic Statistics & Randomized Controlled

Trials(RCTs)
Zhaopeng Qu
Business School,Nanjing University
Sep. 19th, 2019
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Zhaopeng Qu (Nanjing University) Introduction to Econometrics Sep. 19th, 2019 1 / 129
Outlines
1 An Brief Review of Basic Statistics
2 Review: Random Experiment as the Research Design
3 What is an RCT?
4 Randomization: The Cornerstone of RCTs
5 How to Run RCTs?
6 Limitations of RCTs
7 An Example of Randomized Controlled Trials . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .
An Brief Review of Basic Statistics
An Brief Review of Basic Statistics
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
An Brief Review of Basic Statistics Basic Concepts
Population and Sample(总体与样本）
A population is a collection of people, items, or events about

which you want to make inferences.
Population always have a probability distribution.
A sample is a subset of population, which draw from population
in a certain way.
The sample could also follow a probability distribution.
To represent the population well, a sample should be randomly
collected and adequately large.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Random Sample(随机样本) and i.i.d(独立同分布）
Definition
The r.v.s are called a random sample of size n from the population
f(x) if X1 , ..., Xn are mutually independent and have the same
p.d.f/p.m.f f(x). Alternatively, X1 , ..., Xn are called independent,
and identically distributedrandom variable with p.d.f/p.m.f ,
commonly abbreviated to i.i.d. r.v.s.
eg. Random sample of n respondents in a survey.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Statistic(统计量) and Sampling Distribution
Definition
X1 , ..., Xn is a random sample of size n from the population f(x). A
statistic is a real-valued or vector-valued function fully depended on
X1 , ..., Xn , thus
T = T(X1 , ..., Xn )
A statistic is only a function of the sample (统计量是样本的函

数).
The probability distribution of a statistic T is called the
sampling distribution (抽样分布）of T.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Sample Mean(样本均值) and Sample Variance(样本方差）
Definition
The sample average or sample mean, X, of the n observation X1 , ..., Xn
is
1∑
n
1
X̄ = (X1 + X2 + ... + Xn ) = Xi
n n
i=1
Accordingly, the sample variance is
1 ∑
n
S2 = (Xi − X)2
n−1
i=1
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Sample Mean(样本均值) and Sample Variance(样本方差）
As we know that if Xi is a random variable(r.v.), then f(Xi ),

which is a function of Xi , is also a r.v..(随机变量的函数还是随
机变量）
∑
So if Xi is a r.v., then Xi is also a r.v..
the sample mean and the sample variance are also a
function of sums, so they are a r.v. too.
there are some certain probability functions which can
describe distributions of the sample mean and the sample
variance.
Then naturally ask: what is the expectation, variance or
p.d.f./c.d.f. of these distributions?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
A simple case of sample mean

Let {X1 , ..., Xn } ∈ [1, 100] , assume n = 2, thus only X1 and X2
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
An Brief Review of Basic Statistics Large-Sample Approximations to Sampling Distributions
Sampling Distributions
There are two approaches to characterizing sampling
distributions:
exact/finite sample distribution: The sampling distribution
that exactly describes the distribution of X for any n is
called the exact/finite sample distribution of X.
approximate/asymptotic distribution: when the sample size
n is large, the sample distribution approximates to a certain
distribution function.
Two key tools used to approximate sampling distributions when
the sample size is large, thus assume that n → ∞
The Law of Large Numbers(L.L.N.): when the sample size
is large, X will be close to µY , the population mean with
very high probability.
The Central Limit Theorem(C.L.T.): when the sample
size is large, the sampling distribution of the standardized . . . . . . . . . . . . . . . . . . . .
sample average,(Y−µ
Zhaopeng Qu (Nanjing University)
Y )/σ
Y to, Econometrics
Introduction
is approximately normal.
. . .
Sep. 19th, 2019
. . . .
10 / 129
. . . . . . . . . . . . .
Convergence in probability(概率收敛）
Definition
Let X1 , ..., Xn be an random variables or sequence, is said to converge
in probability to a value b if for every ε > 0,
P(| Xn − b |> ε) → 0
p
as n → ∞. We write this Xn −
→ b or plim(Xn ) = b.
it is similar to the concept of a limitation in a probability way.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Law of Large Numbers(大数定律）
Theorem
Let X1 , ..., Xn be an i.i.d draws from a distribution
∑ with mean µ and
finite variance σ 2 (a population) and X = n1 ni=1 Xi is the sample
mean, then
p
X− →µ
Intuition: the distribution of Xn “collapses”on µ.

直观解释：选择的样本量越大，其平均值越接近总体平均值
（抽样分布更紧凑）
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
A simple case
Example
Suppose X has a Bernoulli distribution if it have a binary values
X ∈ {0, 1} and its probability mass function is
{
0.78 if x = 1
P(X = x) =
0.22 if x = 0
then E(X) = p = 0.78 and Var(X) = p(1 − p) = 0.1716.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . .... .... .... . . . . .
Convergence in Distribution(分布收敛)
Definition
Let X1 , X2, ... be a sequence of r.v.s, and for n = 1, 2, ...let Fn (x) be
the c.d.f of Xn . Then it is said that X1 , X2 , ...converges in
distribution to r.v. W with c.d.f, FW if
lim Fn (x) = FW (x)

n∞
d
which we write as Xn −
→ W.
Basically: when n is big, the distribution of Xn is very similar to

the distribution of w.
Common to standardize a r.v. by subtracting its expectation and
dividing by its standard deviation
X − E[X]
Z= √ . . . . . . . . . . . . . . . . . . . .
Var[X] . . . . . . . . . . . . . . . . . . . .
The Central Limit Theorem(中心极限定理)

Theorem
Let X1 , ..., Xn be an i.i.d draws from a distribution with sample size n
with mean µ and 0 < σ 2 < ∞, then
Xn − µ d
σ/√n
∼ N(0, 1)
Because we don’t have to make specific assumption about the

distribution of Xi , so whatever the distribution of Xi , when n is
big,
d
the standardized Xn ∼ N(0, 1)
d 2
or Xn ∼ N(µ, σn )
直观理解：选取的样本量越大，样本均值的分布越趋于正态
分布。 . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . .... .... .... . . . . .
. . . .... .... .... . . . . .
How large is “large enough”?
How large is large enough ?

how large must n be for the distribution of Y to be
approximately normal?
The answer: it depends.
if Yi are themselves normally distributed, then Y is exactly
normally distributed for all n.
if Yi themselves have a distribution that is far from normal,
then this approximation can require n = 30 or even more.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
An Brief Review of Basic Statistics Statistical Inference: Estimation, Confident Intervals and Testing
Statistical Inference
Inference
What is our best guess about some quantity of interest?
What are a set of plausible values of the quantity of interest?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Inference: from Samples to Population
Our focus: {Y1 , Y2 , ..., Yn } arei.i.d. draws from f(y) or F(Y),

thus population distribution.
Statistical inference or learning is using samples to infer f(y).
Normally, we don’t need to know everything of the population,
just some measures (the Moment) enough to describe the
characteristics of the population.
Point estimation: providing a single “best guess”as to the

value of some fixed, unknown quantity of interest, θ, which is is a
feature of the population distribution, f(y).
µ =?E[Y]
σ 2 =?Var[Y]
µy − µx ? = E[Y] − E[X]
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Three Characteristics of an Estimator

Let µ̂Y denote the some estimation value of the population
moment, µY and E(µ̂Y ) is the mean of the sampling distribution
of µ̂Y ,
1 Unbiasedness: the estimator of µY is unbiased if
E(µ̂Y ) = µY
2 Consistency:the estimator of µY is consistent if
p
µ̂Y −
→ µY
3 Efficiency:Let µ̃Y be another estimator of µY and suppose that
both µ̃Y and µ̂Y are unbiased.Then µ̂Y is said to be more
efficient than µ̂Y
var(µ̂Y ) < var(µ̃Y )
Comparing variances is difficult if we do not restrict our
attention to unbiased estimators because we could always . . . . . . . . . . . . . . . . . . . .
use aUniversity)
Zhaopeng Qu (Nanjing trivial estimator with
Introduction variance zero that Sep.
to Econometrics is biased.
. . .
19th, 2019
. . . .
22 / 129
. . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Properties of the Sample Mean
Let µY and σY2 denote the mean and variance of Y,(总体的均

值和方差）
∑
Let Y = n1 ni=1 Yi is the sample mean of Yi (样本均值）
Then the expectation of the sample mean(样本均值的期望) is
1∑
n
E(Y) = E(Yi ) = µY
n i=1
so Y is an unbiased estimator of µY .
p
Based on the L.L.N., Y −
→ µY , so Y is also consistent.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Variance of sample mean(样本均值的方差）

( n )
1∑ 1 ∑
n
σ2
Var(Y) = var Yi = 2 Var(Yi ) = Y
n i=1 n i=1 n
σY
Then, the Standard Deviation of the sample mean is σY = √
n
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Follow the C.L.T, the
d σY2
Y ∼ N(µY , )
n
And let Z be the standardized Y, then
Y − µY d
Z= σY ∼ N(0, 1)
√
n
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Properties of the Sample Variance
Let µY and σY2 denote the mean and variance of Yi , then the
sample variance：
1 ∑
n
S2Y = (Yi − Y)2
n − 1 i=1
Then it is easy to prove that

2 2 2
1 E(S ) = σ , thus S is an unbiased estimator of σ
2
Y Y Y which is
also the reason why the average uses the divisor n − 1
instead of n.
2 P
Y −→ σY2 , thus the sample variance is a consistent
2 S
estimator of the population variance.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Standard Error
Recall: the standardized sample mean will be approximately

follow a standard normal distribution when n is large.
Y − µY d
Z= σY ∼ N(0, 1)
√
n
But in general σY , the standard deviation of population is

unknown, so we have to use sample to estimate it.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Standard Error
σY
Let σY = √ n
, because S2Y is an unbiased and consistent
SY
estimator of the σY2 , then we can use √ n
as an estimator of the
standard deviation of the sample mean, σY .
It is called the standard error(标准误）of the sample mean
SY
SE[Y] = σ̂Y = √
n
Equivalence to the “standard deviation”（标准差）of the sample

distribution which measures the deviations of the sample mean.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Application: Sample Size and Standard Error(Population)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Sample Size and Standard Error(n=250)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Recall: The Chi-Square Distribution
Let Zi (i = 1, 2, ..., m) be independent random variables, each

distributed as standard normal. Then a new random variable
can be defined as the sum of the squares of Zi :
∑
m
X= Z2i
i=1
Then X has a chi-squared distribution with m degrees of

freedom.
Then, it can be prove that a variation of the sample variance will
follow a Chi-Square distribution
(n − 1)S2Y
∼ χ2n−1
σ2
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Student-t Distribution

The Student t distribution can be obtained from a standard
normal and a chi-square random variable.
Let Z have a standard normal distribution, let X have a
chi-square distribution with m degrees of freedom and assume
that Z and X are independent. Then the random variable
Z
T= √
X/n
has has a t-distribution with m degrees of freedom, denoted as

T ∼ tm .
Then, the Z will follow a student t distribution.
Y − µY
Z= SY
∼ t(n − 1)
√
n . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Student-t Distribution
But it does not matter

a lot in the large
sample.
As the degrees of
freedom get large
which is highly
correlated with the
sample size n, the
t-distribution actually
approaches the
standard normal
distribution.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
An Brief Review of Basic Statistics Interval Estimation and Confidence Intervals
Interval Estimation
A point estimate provides no information about how close the

estimate is “likely”to be to the population parameter.
We cannot know how close an estimate for a particular sample is
to the population parameter because the population is never
unknown.
A different (complementary) approach to estimation is to
produce a range of values that will contain the truth with some
fixed probability.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is a Confidence Interval?
Definition
A 100(1 − α)% confidence interval for a population parameter θ is an
interval Cn = (a, b) , where a = a(Y1 , ..., Yn ) and b = b(Y1 , ..., Yn )
are functions of the data such that
P(a < θ < b) = 1 − α
In general, this confidence level is 1 − α ; where α is called

significance level.
The key is how to obtain or construct the values of a and b.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Interval Estimation and Confidence Intervals

Suppose the population has a normal distribution N(µ, σ 2 ) and
let Y1 , Y2 , ..., Yn be a random sample from the population.
Then the sample mean Y has a normal distribution:
2
Y ∼ N(µ, σn )
The standardized sample mean Z is given by:
Z = Y−µ
σ/√n ∼ N(0, 1)
Then let θ = Z, then P(a < θ < b) = 1 − α turns into

Y−µ
a< σ/√n
<b
then it follows that

P(Y − aσ/√n < µ < Y + bσ/√n) = 1 − α
The random interval contains the population mean with a
probability 1 − α. .
.
.
.
.
. . . . .
. . . .
. . . .
. . . .
. . . .
. . . . .
.
.
.
.
.
.
.
.
.


Two cases: σ is known and unknown
When σ is known, for example,σ = 1, thus Y ∼ N(µ, 1),
σ2 1
Y ∼ N(µ, = )
n n
From this, we can standardize Y, and, because the standardized
version of Y has a standard normal distribution, and we let
α = 0.05, then we have
Y−µ
P(−1.96 <
1/√n
< 1.96) = 1 − 0.05
The event in parentheses is identical to the event

Y − 1.96/√n ≤ µ ≤ Y + 1.96/√n, so
P(Y − 1.96/√n ≤ µ ≤ Y + 1.96/√n) = 0.95
The interval estimate of µ may be written as . . . . . . . . . . . . . . . . . . . .
[Y − 1.96/√n, Y + 1.96/
Zhaopeng Qu (Nanjing University)
√
n]
Introduction to Econometrics
. . . . . . . . . . . . . . .
Sep. 19th, 2019
. . .
41 / 129
. .
When σ is unknown, we could use an estimate of σ,thus

SY
SE[Y] = σ̂Y = √ n
, the standard error, replacing unknown σ
thus
Y − µY Y − µY
Zt = SY =
√
n
SE(Y)
We just prove that it follows a student t distribution.
Zt ∼ tn−1
Definition
The t-statistic or t-ratio:
Y − µY
∼ tn−1
SE(Y)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
To construct a 95% confidence interval, let c denote the 97.5th

percentile in the tn−1 distribution.
P(−c < t ≤ c) = 0.95
where cα/2 is the critical value of the t distribution.

The condence interval may be written as [Y ± cα/2 S/√n]
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
A simple rule of thumb for a 95% confidence interval
Because as the degrees of freedom get large which is highly

correlated with the sample size n, the t-distribution approaches
the standard normal distribution.
And Φ(1.96) = 0.975, so a rule of thumb for an approximate
95% confidence interval is
[Y ± 1.96 × SE(Y)]
Or
[Y ± 2 × SE(Y)]
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
An Brief Review of Basic Statistics Hypothesis Testing(假设检验）
Hypothesis Testing
Definition
A hypothesis is a statement about a population parameter, thus θ.
Formally, we want to test whether is significantly different from a
certain value µ0
H0 : θ = µ0
which is called null hypothesis. The alternative
hypothesis(two-sided) is
H1 : θ ̸= µ0
If the value µ0 does not lie within the calculated confidence

interval, then we reject the null hypothesis.
If the value µ0 lie within the calculated confidence interval, then
we fail to reject the null hypothesis. . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Introduction
In criminal law, institutions in most countries follow the rule:

“innocent until proven guilty”(疑罪从无)
The prosecutor wants to prove their hypothesis that the
accused person is guilty.
However, the burden is on the prosecutor to show guilt.
The jury or judge starts with the “null hypothesis”that
the accused person is innocent.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Introduction
In program evaluations,instead of “presumption of innocence,”

the rule is: “presumption of insignificance”
Policymaker’s hypothesis: the program improves learning.
Evaluators approach experiments using the hypothesis:
there is zero impact of the program
Then we test this “null hypothesis”
The burden of proof is on the program
it should show a statistically significant impact.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Two Type Errors(两种错误）

In both cases, there is a certain risk that our conclusion is wrong
Type I Error
A Type I error is when we reject the null hypothesis when it is in fact
true.(“left-wing”)
A Type II error is when we fail to reject the null hypothesis when it is
false.(“right-wing”)
In Criminal trial Case

The Type I : the judge reject the null hypothesis when the
suspect is actually no guilty.
“宁可错杀一千，不能放过一个”
The Type II: the judge fail to reject the null hypothesis
when the suspect is actually guilty.
“宁可放过一千，不能错杀一个” . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Significance level(显著性水平)
Definition
The significance level or size of a test is the maximum probability
for the Type I Error
P(Type I error) = P(reject H0 | H0 is true) = α
Usually, we has to carry the "burden of proof,"

We would like to prove that the assertion of H1 is true by
showing that the data rejects H0 .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Testing procedure
The following are the steps of the hypothesis testing:

1 Specify H and H .
0 1
2 Choose the significance level α.
3 Define a decision rule (critical value).
4 Given the data compute the test statistic and see if it falls
into the critical region.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Decision Rule
The decision rule that leads us to reject or not to reject H0 is

based on a test statistic, which is a function of the data
Tn = T(Y1 , ..., Yn )
Usually, one rejects H0 if the test statistic falls into a critical

region(rejection region). A critical region is constructed by
taking into account the probability of making wrong
decisions,thus α.
By convention, α is chosen to be a small number, for example,
α = 0.01, 0.05, or 0.10
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
P-Value
To provide additional information, we could ask the question:

What is the largest significance level at which we could carry out
the test and still fail to reject the null hypothesis?
Or in other word, given the data, the smallest significance level
at which the null can be rejected.
We can consider thep-valueof a test
1 Calculate the t-statistic t
2 The largest significance level at which we would fail to reject
H0 is the significance level associated with using t as our

critical value
p − value = 1 − Φ(t)
where Φ(t) denotes the standard normal c.d.f.(we assume
that n is large enough)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
P-Value
Suppose that t = 1.52, then we can find the largest significance level
at which we would fail to reject H0
p − value = P(T > 1.52 | H0 ) = 1 − Φ(1.52) = 0.065
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Hypothesis Test of of Ȳ
Specify H0 and H1
H0 : E[Y] = µY,0 H1 : E[Y] ̸= µY,0
Choose the significance level α and define a decision rule (critical

region or critical value)
eg. if we choose α = 0.05, then the critical value is 1.96,
then the region is (−∞, −1.96] and [1.96, +∞)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Given the data compute the test statistic

Step1: Compute the sample average Ȳ
Step2: Compute the standard error of Ȳ
sY
SE(Y) = √
n
Step3: Compute the t-statistic
Ȳ − µY,0
tact = ( )
SE Ȳ
Step4: Reject the null hypothesis if

| tact |> critical value
or if p − value < significance level
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Review: Random Experiment as the Research Design
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Recall the last lecture
The Core of Empirical Studies: Causality v.s. Forecasting

The Central Question of Causality?
Rubin Causal Model: comparing counterfactuals or potential
outcomes.
However, we can never observe both counterfactuals —
fundamental problem of causal inference.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Recall the last lecture
Random Assignment Solves the Selection Problem.

We should treat experimental design as a Benchmark.
To construct the counterfactuals, two broad categories of
empirical strategies.
Random Controlled Trials/Experiments:
it can eliminates selection bias which is the most
important bias arises in empirical research. If we could
observe the counterfactual directly, then there is no
evaluation problem, just simply difference.
Program Evaluation Econometrics:
The various approaches using naturally-occurring data
provide alternative methods of constructing the proper
counterfactual.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
What is an RCT?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
Randomized Controlled Trials(RCTs)(随机可控试验）
In essence, an RCT is an experiment carried out on two or

more groups where participants are randomly assigned to
receive an intervention or not.
Participants are randomly assigned to either an treatment
group who are given the intervention, or a control group
who are not..
In RCTs, each group is tested at the end of the trial and the
results from the groups are compared to see if the intervention
has made a difference and achieved its desired outcome. If the
randomized groups are large enough, you can be confident that
differences observed are due to the intervention and not some
other factor.
RCTs are considered the gold standard for establishing a causal
link between an intervention and change.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCT in History
First recorded RCT was done in 1747

by James Lind,who was a Scottish
physician in the Royal Navy.
Scurvy(败血症) is a terrible disease
caused by Vitamin C deficiency.
Serious issue during long sea voyages.
Lind took 12 sailors with scurvy and
split them into six groups of two.
Groups were assigned:
(1) 1 qt cider(苹果酒) (2) 25
drops of vitriol(硫酸）(3) 6
spoonfuls of vinegar, (4) 1/2 pt
of sea water, (5) garlic,
mustard(芥末)and barley water
.
.
.
.
.
. . . . .
. . . .
. . . .
. . . .
. . . .
. . . . .
.
.
.
.
.
.
.
.
.

What is an RCT?
RCT in History
Only Group 6 (citrus fruit) showed substantial improvement.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCT in History
Ronald A.
Fisher(1890-1962),British
statistician and geneticist who
pioneered the application of
statistical procedures to the design of
scientific experiments.
“a genius who almost
single-handedly created the
foundations for modern
statistical science”.
“Rothamsted Experimental
Station”
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCTs in Economics
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCTs in Social Policies
According to Baruch (1978), 245 randomized field experiments

had been conducted in U.S for social policies evaluations up to
1978.
The huge effort has been prompted by the 1% part of every
social budget devoted to evaluation.
Some of them were ambitious and very costly, and affected
different kind of policies.
the Perry Preschool Program in 1961
The Rand Health Insurance Experiment from 1974-1982.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
Education: the Perry Preschool Program
123 children born between 1958 and 1962 in Michigan

Half of them (drawn at random) entered the perry school
program at 3 or 4 years old.
Education by skilled professionals in nurseries and kindergarten.
Program duration circle 30 weeks
follow-up survey (age : 14, 15, 19, 27 and 40 years old)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
Health Care: The Rand Health Insurance Experiment
5809 people randomly assigned in 1974 to different insurance

programs with 0%, 25%, 50% and 75% sharing.
They were followed until 1982.
Main results : paying a portion of health cost make people give
up some “superfluous”cares, with little harm on their health.
But some heterogeneity : not true for poor people.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCTs in China
“One egg a day”program in rural China by REAP at Stanford.

One egg a day
“Free-lunch”program in primary schools at Western China.
Free Lunch
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCT in Business
An interesting question: What is the optimal color for taxis?

Ho, Chong and Xia(2017), Yellow taxis have fewer accidents
than blue taxis because yellow is more visible than blue,
PNAS.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCT in Business
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
RCT in Business
Another Critical Question for business: Is Working at Home is

better than Working at Office?
Bloom, Liang, Roberts and Ying,(2015), “Does Working from Home
Work? Evidence from a Chinese Experiment”, The Quarterly Journal
of Economics
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
What is an RCT?
Types of RCTs
Lab Experiments
eg: students evolves a experiment in a classroom.
eg: computer game for gamble in Lab
Field Experiments
eg: the role of women in household’s decision or fake
resumes in job application
Quasi-Experiment or Natural Experiments: some unexpected
institutional change or natural shock
eg: Germany Reunion(德国统一), Great Famine in
China(1959-1961 年大饥荒）and U.S Bombing in
Vietnam(美国轰炸越南).
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Randomization: The Cornerstone of RCTs
Randomization: The Cornerstone of RCTs
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Randomization: The Cornerstone of RCTs Introduction
Introduction:
Randomization is the key element that separates RCTs from

other study designs.
Assume that our topic is about the entrepreneur or the
firm,please tell whether the following ways to assign participants
into treatment is random
the left(front)side and right(back)side in our class room.
Date of birth(生日）
Some sort of record number（登记号码奇偶数字)
Day of enrollment(参加实验的时间交替给编号）
None of these methods described above should be considered as
really generating random assignments,but rather systematic
occurrences.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Randomization: The Cornerstone of RCTs Introduction
What does random assignment mean?
More specifically, it means that all participants have a defined

probability of being assigned to a particular treatment(group).
In RCTs, neither the investigator, practitioners or participants
determine assignments and it is not predictable based on a
pattern.
So what are better potential ways to achieve the random
assignment?
tossing a coin
lottery
Most investigators use software packages to perform
randomization
There is a wide range of software programs available that
can be used, including Excel, Stata or R.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Randomization: The Cornerstone of RCTs Randomize in different levels
Which level to randomize?
1 Randomizing at the individual level

2 Randomizing at the group level
Have to consider
What unit does the program target for treatment?
What is the unit of analysis?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Unit of Randomization: Individual
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Unit of Randomization: Individual
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Unit of Randomization: Class
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Unit of Randomization:Class
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Unit of Randomization:School
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Unit of Randomization:School
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Constraints in reality
1 Simple random sampling. Most basic of the probability sampling

methods —every unit has an equal probability of being selected in the
study. Easiest to do.
2 Stratified sampling. Sample each subpopulation of interest
independently and then reweigh after data collection. Statistically
efficient while allowing for adequately powered subgroup analyses.
3 Cluster sampling. Sample by a "cluster" such as geographical location
or unit. For example, sampling entire schools at random. Statistically
efficient if more variance between clusters than within clusters.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Other methods of randomization
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Political Advantages:
Lotteries are simple, common and transparent.
Randomly chosen from applicant pool
Participants know the “winners”and “losers”
Simple lottery is useful when there is no a priori reason to discriminate
Perceived as fair.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Avoid contamination?
Spillovers
Crossovers
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Randomization: The Cornerstone of RCTs Multiple Treatments
Introduction
Some times core question is deciding among different possible

interventions.
Can randomize these programs
any one more intervention is benefit enough
Remember: you should always have a control group
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Multiple treatments: can’t work
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Varying levels of treatment:
Treatment 1: Some schools are assigned full treatment: All kids get
pills
Treatment 2: Some schools are assigned partial treatment: 50% kids
get pills
Control: Some schools are assigned zero treatment: no kids get pills
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Should the assignment make the number of participants

equal to each group?
In general, participants are typically allocated in equal numbers to
different treatments.
50% of participants go to the treatment group and the other 50% to
the control group
However, there are a number of practical reasons for which you may
choose to have a different allocation ratio, i.e. an unequal allocation.
resource constraints or costs: a fixed quota of participants it needs to
assign to the treatment group.
an allocation ratio of 3:1, where the intervention group has three times
as many participants as the control group.
limited resources to provide the intervention (including the availability
of staff delivering it), or treatment is significantly more expensive.
an allocation ratio of 1:2 (twice as many participants in the control
group compared to the intervention group)
In most cases, we don’t even to adjust the weight of the sample.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
How to Run RCTs?
How to Run RCTs?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
How to Run RCTs? The Process of RCTs
Step 1:Asking the Right Question

Choosing a right topic: Check whether the answer to the question
you are planning to study already exists.
There are many different types of questions that you might be
interested in,ensure that you are asking the type of question that
requires an RCT.
Strategic Question: What areas of innovation policy should we focus
our attention on
Descriptive question: What are the key challenges that SMEs in certain
geographic areas are facing.
Process-related questions: Is the young entrepreneurs mentoring
program using the resources dedicated for program delivery?
RCTs can answer questions about impact:
Did the program or policy work?
Which components of the intervention were most crucial for achieving
impact?
Try answer one question one time at first. . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Step 2: Finding a specific context
Transform into a measurable empirical question

Empirical problem: Class size and educational output.
Policy question: What is the effect on test scores (or some
other outcome measure) of reducing class size by one
student per class? By 8 students/class?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Step 3: Making a experiment design
Specifying eligibility criteria for participants

Specifying the treatment
Specifying the outcomes
Sample sizes
Randomization
Specifying the analyzing method
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Step 4: Piloting the treatment
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Step 5:Random assignment(Baseline)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Step 6:Implementing the program
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Step 7: Measure participant outcomes
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Step 8: Analyzing the data and Writing the report
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
How to Run RCTs? An assuming case: the California School
A Case: the California School
Draw schools (n = 420) randomly from all school in California

Variables:
5th grade test scores (Stanford-9 achievement test, combined math
and reading), district average
Student-teacher ratio (STR) = no. of students in the district divided
by no. full-time equivalent teachers
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Summary Table: Descriptive Statistics
Does this table tell us anything about the relationship between test
scores and the STR?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Scatterplot: test score v. student-teacher ratio
What does this figure show? and it may suggest...? .

.
.
.
.
. . . . .
. . . .
. . . .
. . . .
. . . .
. . . . .
.
.
.
.
.
.
.
.
.
The California Test Score
We need to get some numerical evidence on whether districts with

low STRs have higher test scores.
But how?
1 Compare average test scores in districts with low STRs to those with
high STRs (“estimation”)
2 Test the “null”hypothesis that the mean test scores in the two types
of districts are the same, against the “alternative”hypothesis that
they differ (“hypothesis testing”)
3 Estimate an interval for the difference in the mean test scores, high v.
low STR districts (“confidence interval”)
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The California Test Score
Compare districts with “small”and “large”class sizes:
Small v.s. Large

Class Size Average score(Y) Standard deviation N
Small(STR < 20) 657.4 19.4 238
Large(STR ⩾ 20) 650.0 17.9 182
1 Estimation of ∆= difference between group means

2 Test the hypothesis that ∆ = 0
3 Construct a confidence interval for ∆
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
How to Run RCTs? Comparing Means from Different Populations
An Example: Comparing Means from Different Populations
In an RCT, we would like to estimate the average causal effects

over the population
ATE = ATT = E{Yi (1) − Yi (0)}
We only have random samples and random assignment to

treatment, then what we can estimate instead
difference in mean = Ytreated − Ycontrol
Under randomization, difference-in-means is a good estimate for

the ATE.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Hypothesis Tests for the Difference Between Two Means
To illustrate a test for the difference between two means, let mw

be the mean hourly earning in the population of women recently
graduated from college and let mm be the population mean for
recently graduated men.
Then the null hypothesis and the two-sided alternative
hypothesis are
H0 : µm = µw
H1 : µm ̸= µw
Consider the null hypothesis that mean earnings for these two
populations differ by a certain amount, say d0 . The null
hypothesis that men and women in these populations have the
same mean earnings corresponds to H0 : H0 : d0 = µm − µw = 0
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Difference Between Two Means

Suppose we have samples of nm men and nw women drawn at
random from their populations. Let the sample average annual
earnings be Ym for men and Yw for women. Then an estimator
of µm − µw is Ym − Yw .
Let us discuss the distribution of Ym − Yw .
σ2 σ2
∼ N(µm − µw , m + w )
nm nw
2 2
if σm and σw are known, then the this approximate normal
distribution can be used to compute p-values for the test of the
null hypothesis. In practice, however, these population variances
are typically unknown so they must be estimated.
Thus the standard error of Ym − Yw is
√
s2m s2
SE(Ym − Yw ) = + w
nm nw .
.
.
.
.
. . . . .
. . . .
. . . .
. . . .
. . . .
. . . . .
.
.
.
.
.
.
.
.
.
The Difference Between Two Means
The t-statistic for testing the null hypothesis is constructed

analogously to the t-statistic for testing a hypothesis about a
single population mean, thus t-statistic for comparing two means
is
Ym − Yw − d0
tact =
SE(Ym − Yw )
If both nm and nm are large, then this t-statistic has a standard
normal distribution when the null hypothesis is true,thus
Ym − Yw = 0.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Confidence Intervals for the Difference Between Two

Means
the 95% two-sided confidence interval for d consists of those

values of d within ±1.96 standard errors of Ym − Yw , thus
d = µm − µw is
(Ym − Yw ) ± 1.96SE(Ym − Yw )
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Hypothesis Test of the Difference Between Two Means
Reject the null hypothesis if

Ym −Yw −d0
| tact |=| SE(Ym −Yw )
|> critical value
or if p − value < significance level
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Limitations of RCTs
Limitations of RCTs
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Limitations of RCTs
RCT are far from perfect!
High Costs, Long Duration

Potential Ethical Problems: “Parachutes reduce the risk of
injury after gravitational challenge, but their effectiveness has not
been proved with randomized controlled trials."
Milgram Experiment
Stanford Prison Experiment
Monkey Experiment
Limited Generalizability
RCTs allow us to gain knowledge about causal effects but
without knowing the mechanism.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Limitations of RCTs
Potential Problems in Practice
Small sample: Student Effect

Hawthorne effect(霍桑效应）：The subjects are in an experiment
can change their behavior.
Attrition（样本流失）：It refers to subjects dropping out of the
study after being randomly assigned to the treatment or control
group.
Failure to randomize or failure to follow treatment protocol:
People don’t always do what they are told.
Wearing glasses program in Western Rural China.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Limitations of RCTs
Program Evaluation Econometrics(项目评估计量经济学）

Question: How to do empirical research scientifically when we
can not do experiments? It means that we always have selection
bias in our data, or in term of “endogeneity”.
Answer: Build a reasonable counterfactual world by naturally
occurring data to find a proper control group is the core of
econometrical methods.
Here you Furious Seven Weapons in Applied Econometrics(七种
盖世武器）
1 Random Controlled Trials (RCT)（随机实验）
2 OLS(最小二乘回归）
3 Decomposition（分解）
4 Instrumental Variable（工具变量）
5 Differences in Differences（双差分）
6 Matching and Propensity Score（匹配）
7 Regression Discontinuity（断点回归）
8 Synthetic Control （合成控制法） . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
An Example of Randomized Controlled Trials
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Working from Home(WFH) v.s Working from Office
“Does Working from Home Work? Evidence from a Chinese

Experiment”,by Nicholas A. Bloom, James Liang, John Roberts,
Zhichun Jenny Ying The Quarterly Journal of
Economics,February 2015, Vol. 130, Issue 1, Pages 165-218.
Basic Question: WFH = SFH
SFH(Shirking from Home)?
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Working from Home(WFH) is going to a trend

internationally.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Motivations
Working from home is a modern management practice which

appears to be stochastically spreading in the US and Europe
20 million people in US report working from home at least once
per week
Little evidence on the effect of workplace flexibility
productivity
employee satisfaction
shirking
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Ctrip Experiment
Ctrip, China’s largest travel-agent, with16,000 employees, $6bn

NASDAQ.
Co-founder of Ctrip, James Liang, was an Econ PhD at
Stanford and decided to run a experiment to test WFH.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Ctrip Experiment: A call center in Shanghai

The experiment runs on airfare & hotel departments in Shanghai.
Main Work: Employees take calls and make bookings.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
The Experimental Design
Treatment: work 4 shifts (days) a week at home and to work the

5th shift in the office on a fixed day.
Control: work in the office on all 5 days.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
In early November 2010, employees in the airfare and hotel

booking departments were informed of the WFH program.
Of the 994 employees in the airfare and hotel booking
departments, 503 (51%) volunteered for the experiment.
Among the volunteers, 249 (50%) of the employees met the
eligibility requirements and were recruited into the experiment.
The treatment and control groups were then determined from
this group of 249 employees through a public lottery.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Results: the number of receiving calls
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Results: Working hours
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Results:Many Outcomes
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
Conclusion: Very positive
They found a highly significant 13% increase in employee

performance from WFH,
of which about 9% was from employees working more
minutes of their shift period (fewer breaks and sick days)
and about 4% from higher performance per minute.
Home workers also reported substantially higher work satisfaction
and psychological attitude scores, and their job attrition rates fell
by over 50%.
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .

Introduction To Econometrics: Lecture 2: Review of Basic Statistics & Randomized Controlled Trials (RCTS)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Introduction To Econometrics: Lecture 2: Review of Basic Statistics & Randomized Controlled Trials (RCTS)

Uploaded by

Copyright:

Available Formats

Introduction to Econometrics

Lecture 2: Review of Basic Statistics & Randomized Controlled

Business School,Nanjing University

Sep. 19th, 2019

1 An Brief Review of Basic Statistics

2 Review: Random Experiment as the Research Design

4 Randomization: The Cornerstone of RCTs

5 How to Run RCTs?

7 An Example of Randomized Controlled Trials . . . . . . . . . . . . . . . . . . . .

An Brief Review of Basic Statistics

Population and Sample(总体与样本）

A population is a collection of people, items, or events about

Random Sample(随机样本) and i.i.d(独立同分布）

eg. Random sample of n respondents in a survey.

Statistic(统计量) and Sampling Distribution

A statistic is only a function of the sample (统计量是样本的函

Sample Mean(样本均值) and Sample Variance(样本方差）

Accordingly, the sample variance is

Sample Mean(样本均值) and Sample Variance(样本方差）

As we know that if Xi is a random variable(r.v.), then f(Xi ),

A simple case of sample mean

it is similar to the concept of a limitation in a probability way.

The Law of Large Numbers(大数定律）

Intuition: the distribution of Xn “collapses”on µ.

then E(X) = p = 0.78 and Var(X) = p(1 − p) = 0.1716.

lim Fn (x) = FW (x)

Basically: when n is big, the distribution of Xn is very similar to

The Central Limit Theorem(中心极限定理)

Because we don’t have to make specific assumption about the

How large is “large enough”?

How large is large enough ?

Inference: from Samples to Population

Our focus: {Y1 , Y2 , ..., Yn } arei.i.d. draws from f(y) or F(Y),

Point estimation: providing a single “best guess”as to the

Three Characteristics of an Estimator

Three Characteristics of an Estimator

Three Characteristics of an Estimator

Properties of the Sample Mean

Let µY and σY2 denote the mean and variance of Y,(总体的均

Properties of the Sample Mean

The Variance of sample mean(样本均值的方差）

Properties of the Sample Mean

Follow the C.L.T, the

Properties of the Sample Variance

Then it is easy to prove that

estimator of the population variance.

The Standard Error

Recall: the standardized sample mean will be approximately

But in general σY , the standard deviation of population is

The Standard Error

Equivalence to the “standard deviation”（标准差）of the sample

Application: Sample Size and Standard Error(Population)

Sample Size and Standard Error(n=250)

Sample Size and Standard Error(n=500)

Sample Size and Standard Error(n=1000)

Recall: The Chi-Square Distribution

Let Zi (i = 1, 2, ..., m) be independent random variables, each

Then X has a chi-squared distribution with m degrees of

The Student-t Distribution

has has a t-distribution with m degrees of freedom, denoted as

The Student-t Distribution

But it does not matter