See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/308416439

PROBABILITY & STATISTICS- CHAPTER SIX- 200916_06

Working Paper · September 2016

CITATIONS READS

0 532

1 author:

Mohammed Dahman
Kadir Has University
29 PUBLICATIONS   9 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Multivariate Statistical Modelling View project

PROBABILITY & STATISTICS View project

All content following this page was uploaded by Mohammed Dahman on 22 September 2016.

The user has requested enhancement of the downloaded file.

 Kadir Has University
School of Science and Engineering, Graduate Programs
Ph.D in Management Information System
mohammedr.dahman@stu.khas.edu.tr

Abstract. In this chapter we procced the progress of exploring the last item in the list “inference”. We need to
use a sample to draw a conclusion about the population from it was drawn. The specific form of inference
called for depends on the type of variables involved—either a single categorical or quantitative variable, or a
combination of two variables whose relationship is of interest. To get it started, I should mention three forms
of statistical inference which I explained in length, point estimation, interval estimation, and hypothesis. Next
and last part of this paper work will concern inference relationships.

Introduction
So far we have examined at length the Explanatory Data Analysis & Data Producing /Sampling and
probability. As you most probably know that we have identified four steps to understand the behaviour
of a population or system. To recall the four steps identified as:

a) Sampling,
b) Explanatory Data Analysis,
c) Probability,
d) Inference,

We have accomplished the examination of the first three parts in the list. In this chapter we will start
the fourth part of the process and the final section of work paper.

Before we navigate into this part I would like to stress the goal of this work paper. We need to use a
sample to draw a conclusion about the population from it was drawn. The specific form of inference
called for depends on the type of variables involved—either a single categorical or quantitative
variable, or a combination of two variables whose relationship is of interest. To get it started, I should
mention three forms of statistical inference which we will explain in length.

a) Point Estimation, b) interval estimation, c) Hypothesis testing,

1. Point Estimation
In general we estimate an unknown parameter using a single number that is calculated from the
sample data. There are two important categories to consider while discussing this section, if the
variable is categorical and if the variable is quantitative.

a) When the variable of interest is categorical, the population parameter that we will infer about is
the population proportion (𝑝) associated with that variable. For example, if we are interested in
studying opinions about the death penalty among U.S. adults, and thus our variable of interest is
"death penalty (in favour/against)," we'll choose a sample of U.S. adults and use the collected data
to make an inference about p—the proportion of U.S. adults who support the death penalty.
b) When the variable of interest is quantitative, the population parameter that we infer about is the
population mean (𝜇) associated with that variable. For example, if we are interested in studying
the annual salaries in the population of teachers in a certain state, we'll choose a sample from that
population and use the collected salary data to make an inference about 𝜇, the mean annual salary
of all teachers in that state.

Having we said that, let’s see how many point estimator we have:

1. We use 𝑝̂ (sample proportion) as a point estimator for 𝑝 (population proportion). It is an unbiased

estimator: its long-run distribution is centered at 𝑝 as long as the sample is random.
2. We use 𝑥̅ (sample mean) as a point estimator for 𝜇 (population mean). It is an unbiased estimator:
its long-run distribution is centred at as long as the sample is random.
(𝑥1 −𝑥̅ )2 +⋯+(𝑥𝑛 −𝑥̅ )2
3. Another example of a point estimate is using sample variance,𝑠 2 = , to estimate
𝑛−1
population variance 𝜎 . However in this work paper I would consider the above two estimators.
2

I shall investigate at length the details of each estimator above, however before I do so I shall introduce
the concept of interval estimation.

2. Interval Estimation
Point estimation is intuitive however it’s problematic too, the reason is when we estimate, say, 𝜇 by the
sample mean 𝑥̅ , we are almost guaranteed to make some kind of error. Even though we know that
the values of 𝑥̅ fall around 𝜇, it is very unlikely that the value of 𝑥̅ will fall exactly at 𝜇.

If we are not able to estimate the extent of the error then the estimates are in themselves of limited
usefulness. So the Interval estimation addresses this issue. The idea behind interval estimation is,
therefore, to enhance the simple point estimates by supplying information about the size of the error
attached. The statement which we can use to indicate the interval estimation as:

a) If we are dealing with quantitative variables, "I am 95% confident that by using the point estimate
𝑥̅ = 115 to estimate 𝜇, I am off by no more than 3 X points. In other words, I am 95% confident
that μ is within 3 𝑜𝑓 115, or between 112 (115 − 3) 𝑎𝑛𝑑 118 (115 + 3)".
b) If we wanted to estimate 𝑝 for categorical, the population proportion, by a single number based
on the sample, it would make intuitive sense to use the corresponding quantity in the sample, the
𝑥
sample proportion 𝑝̂ = , 𝑥 𝑖𝑠 𝑡ℎ𝑒 𝑓𝑖𝑛𝑑𝑖𝑛𝑔 𝑝𝑜𝑟𝑡𝑖𝑜𝑛.
𝑛
As we have clarified all these important facts lets’ move and see how point estimation and interval
estimation works together.

Note: before I mention the third form of inference “Hypothesis Testing Part” we investigate in more
details the estimation for population mean and population proportion.

1. Confidence Interval for Population mean: Suppose a random sample of size 𝑛 is taken from a
normal population of values for a quantitative variable whose mean (𝜇) is unknown, when the
𝜎
standard deviation (𝜎) is given. A 95% confidence interval (𝐶𝐼) for 𝜇 is: 𝑥̅ +
− 2∗ , Note that for
√𝑛
now we require the population standard deviation (σ) to be known. Practically, σ is rarely known,
but for some cases, especially when a lot of research has been done on the quantitative variable
whose mean we are estimating (such as IQ, height, weight, scores on standardized tests), it is
reasonable to assume that σ is known. Eventually, we will see how to proceed when σ is unknown,
and must be estimated with sample standard deviation (s).
Other point to mention here is if I want to increase the level of my interval to 99% or to 90%, as
𝜎 𝜎
above equation for 99% 𝑥̅ +− 2.576 ∗ and for 90% 𝑥̅ +− 1.645 ∗ .
√𝑛 √𝑛
Example: The IQ level of students at a particular university has an unknown mean (μ) and known
standard deviation 𝜎 = 15. A simple random sample of 100 students is found to have a sample
mean IQ 𝑥̅ = 115 Estimate 𝜇 with a 90%, 95%, and 99% confidence interval.

Observe the wider the confidence interval for 𝜇. The 99% confidence interval is wider than the
95% confidence interval, which is wider than the 90% confidence interval.

𝜎
Now, we should look back at the equation of defining the interval 𝑥̅ +
− 𝑧∗ ∗ , where z* is a general
√𝑛
notation for the multiplier that depends on the level of confidence. As we discussed before: For a
90% level of confidence, z* = 1.645, for a 95% level of confidence, z* = 2 (or 1.96 if you want to
be really precise), for a 99% level of confidence, z* = 2.576. look at the figure below and let’s agree
to call the part “m” as “margin of error”. And it represents the maximum estimation
error for a given level of confidence.
a) The structure is made of 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒 + − 𝑚𝑎𝑟𝑔𝑖𝑛 𝑜𝑓 𝑒𝑟𝑟𝑜𝑟.
b) The margin of error, m, is therefore "in charge" of the width (or precision) of the confidence
interval, and the estimate is in charge of its location (and has no effect on the width).

𝑧 ∗𝜎 2
c) From above as well I can identify the sample size if I want 𝑛 = ( ) ,
𝑚
 d) We may also ask when it is safe to use interval confidence. The sample must be random.
Assuming that the sample is random, recall from the Probability unit that the Central Limit
Theorem works when the sample size is large (a common rule of thumb for "large" is n > 30),
or, for smaller sample sizes, if it is known that the quantitative variable of interest is distributed
normally in the population. The only situation in which we cannot use the confidence interval,
then, is when the sample size is small and the variable of interest is not known to have a normal
distribution. In that case, other methods, called nonparametric methods. This can be
summarized in the following table:

e) What if σ is unknown? The good news is that we can easily replace the population standard
deviation, 𝜎, with the sample standard deviation, 𝑠. The bad news is that once σ has been
replaced by 𝑠, we lose the Central Limit Theorem, together with the normality of 𝑥̅ , and
therefore the confidence multipliers z* for the different levels of confidence (1.645, 2, 2.576)
are (generally) not accurate any more. The new multipliers come from a different distribution
called the "t distribution" and are therefore denoted by t* (instead of z*). We will discuss the t
𝑠
distribution in more detail when we talk about hypothesis testing. 𝑥̅ + ∗
− 𝑡 ∗ .
√𝑛
𝜎 𝑠
f) Important notes: this quantity is know as the standard deviation of 𝑥̅ , and this quantity
√𝑛 √𝑛
is know as the standard error of 𝑥̅ .
g) It turns out that for large values of n, the t* multipliers are not that different from the z*
multipliers.

2. Confidence Interval for Population proportion 𝒑

̂: having we understood the confidence interval
for population mean it would be much easier to understand how to construct a confidence interval
for population proportion. Few points to make:
𝑝(1−𝑝)
a) The standard deviation of 𝑝̂ is √ ,
𝑛
b) Since, like the sampling distribution of 𝑥̂, the sampling distribution of is normal, the confidence
multipliers that we'll use in the confidence interval for p will be the same z* multipliers we use
for the confidence interval for 𝜇 when 𝜎 is known (using exactly the same reasoning and the
same probability results). The multipliers we'll use, then, are:
1.645, 2, 𝑎𝑛𝑑 2.576 𝑎𝑡 𝑡ℎ𝑒 90%, 95% 𝑎𝑛𝑑 99% confidence levels, respectively.
c) If we're trying to estimate the unknown population proportion 𝑝, we'll replace 𝑝 with its sample
𝑝̂(1−𝑝̂)
counterpart, , and work with the standard error of 𝑝̂ = √ ,
𝑛

𝑝̂(1−𝑝̂)
d) Now we're done. The confidence interval for the population proportion 𝑝 is: 𝑝̅ +
− 𝑧∗√ ,
𝑛
e) Example: Researchers from Dartmouth Medical School conducted a study in 2003 to look at
the connection between watching actors smoking in movies and smoking initiation among
adolescents. In the study, 6,522 U.S. adolescents ages 10-14 who had never tried smoking were
randomly selected. Of those who subsequently tried smoking for the first time, 38% did so
 because of exposure to smoking in the movies. Estimate the proportion of all U.S. adolescents
ages 10-14 who started smoking because of seeing actors smoke in movies by constructing a
95% confidence interval.

Let’s interpret the answer above in terms of its margin of error, as you know the margin of
0.38(1−0.38)
error above is 𝑚 = 2√ = 0.012, that’s almost 1.2%. So can tell that I’m 95% certain
6522
that the proportion we have got 0.38 is within 1.2% of (or more that 1.2% away from) the
proportion of U.S. adolescents ages 10-14 who started smoking because of seeing actors
smoke in movies.
f) There is a trade-off between level of confidence and the width (or precision) of the confidence
interval. The more precision you would like the confidence interval for p to have, the more you
have to pay by having a lower level of confidence.
g) Since n appears in the denominator of the margin of error of the confidence interval for p, for
a fixed level of confidence, the larger the sample, the narrower, or more precise it is. This
brings us naturally to our next point.
h) To estimate the sample size it’s very similar to what we have done above where we can just
𝑝̂(1−𝑝̂) 𝑧 2 𝑝̂(1−𝑝̂)
pull the equation 𝑚 = 𝑧 ∗ √ , and solve for 𝑛 = , but observe the fact that the
𝑛 𝑚2
expression above for determining the sample size depends on 𝑝̂ is problematic. The way to
1
overcome this problem is to take the conservative approach by setting 𝑝̂ = , It is conservative
2
1
because the expression that appears in the numerator 𝑧 𝑝̂ (1 − 𝑝̂ ), is maximized when 𝑝̂ = .
2
2
1
In this case our formulation would be 𝑛 = ,
𝑚2
i) Example: A few days before an election, a media outlet would like to estimate 𝑝, the proportion
of eligible voters who support the Democratic candidate. The media outlet would like the
estimate to be within 1% (that is, 0.01) of the true proportion. What is the sample size needed
1
to achieve this in a poll? Set 𝑛 = = 10000,
0.012
j) Another observation to make, that the margin of error if we adopt the conservative approach
1
we should have 𝑚 = , Example, a random sample of 2,500 U.S. adults was chosen to
√𝑛
participate in a public opinion survey about different issues related to crime. What is the margin
1
of error of this survey? 𝑚 = = 0.02, The survey has a margin of error of 2%. This means
√2500
that for each of the questions asked, the obtained sample proportion will be within 2% of the
proportion among all U.S. adults.
k) When is it safe to use confidence interval? As we did for the mean, the assumption we made
in order to develop the methods in this section was that the sampling distribution of the sample
proportion,𝑝̂ , is roughly normal. Recall from previous chapter “Probability” that the conditions
under which this happens are that 𝑛. 𝑝 ≥ 10 and 𝑛. (1 − 𝑝) ≥ 10. Since 𝑝 is unknown, we will
replace it with its estimate, the sample proportion, and set 𝑛. 𝑝̂ ≥ 10 and 𝑛. (1 − 𝑝̂ ) ≥ 10.
3. Hypothesis
We are now moving to the other kind of inference, hypothesis testing. This form is unlike the inferential
methods we presented so far, where the goal was estimating the unknown parameter, the idea, logic
and goal of hypothesis testing are quite different. Let’s see how it works.

3.1. General Idea: the general idea of hypothesis that we have a claim of something and we have the
opposite of that claim. Each one of the fold has a name. Claim 1 is called the null hypothesis
(denoted "𝐻0 "), and Claim 2 plays the role of the alternative hypothesis (denoted "𝐻𝑎 ").
There are four important steps to construct the platform of hypothesis, as:
a) Stating the claims,
b) Choosing a sample and collecting data,
c) Assessing the evidence,
d) Making conclusion,
I will go through each points through an example to clarify the idea.
Example, A certain prescription allergy medicine is supposed to contain an average of 245 parts per
million (ppm) of a certain chemical. If the concentration is higher than 245 ppm, the drug will likely
cause unpleasant side effects, and if the concentration is below 245 ppm, the drug may be ineffective.
The manufacturer wants to check whether the mean concentration in a large shipment is the
required 245 ppm or not. To this end, a random sample of 64 portions from the large shipment is
tested, and it is found that the sample mean concentration is 250 ppm with a sample standard
deviation in the entire shipment is 12 ppm.
a) Stating the claims: Let's write the two claims notation:
a. The null hypothesis 𝐻0 : the mean concentration in a large shipment is the required 245
ppm,
b. The alternative hypothesis 𝐻𝑎 : the mean concentration in a large shipment is NOT the
required 245 ppm,

Observe that "There is nothing unusual about this shipment, the mean concentration is the
required 245 ppm." This claim is challenged by the manufacturer, who wants to check whether
that is, indeed, the case or not”.

Always we should remember that the null hypothesis is the original claim and the alternative one
is the opposite or the challenged one.

b) Choosing sample & collecting data: in this step we look at sampled data in order to draw
conclusions about the entire population. In order to summarize the data we used simple
sample statistics such as the sample proportion (𝑝̂ ), sample mean (𝑥̅ ) and the sample standard
deviation (𝑠). We will go a step further and use these sample statistics to summarize the data
with what's called a test statistic. Let’s find the details of this step as the example stated:
A sample of 𝑛 = 64 portions is chosen and after summarizing the data it is found that the
sample concentration is (𝑥̅ = 250) and the sample standard deviation is 𝑠 = 12.
Is the fact that (𝑥̅ = 250) is different from 245 strong enough evidence to reject claim 1 and
conclude that the mean concentration in the whole shipment is not the required 245? In other
words, do the data provide strong enough evidence to reject claim 1? That’s what next step
would tell us,
c) Assessing the evidence: this is the step where we calculate how likely it to get data like that is
observed when 𝐻0 true. Since we draw our conclusions based on this probability. If this
probability is very small, then that means that it would be very surprising to get data like that
observed if 𝐻0 were true. The fact that we did observe such data is therefore evidence against
𝐻0 , and we should reject it. On the other hand, if this probability is not very small this means
that observing data like that observed is not very surprising if 𝐻0 were true, so the fact that we
observed such data does not provide evidence against 𝐻0 . This crucial probability, therefore,
has a special name. It is called the p-value of the test.
Obviously, the smaller the p-value, the more surprising it is to get data like ours when 𝐻0 is
true, and therefore, the stronger the evidence the data provide against 𝐻0 .
These days we can easily use some software to find the p-value which is the step to complete
the test statistics as we mentioned above. As I introduce the last part of the four steps I will
explain the steps of test statistics.
Coming back to our example: In order to assess whether the data provide strong enough
evidence against claim 1, we need to ask ourselves the following question: If the mean
concentration in the whole shipment were really the required 245 ppm (i.e., if claim 1 were
true), how surprising would it be to observe a sample of 64 portions where the sample mean
concentration is off by 5 ppm or more (as we did)? It turns out that it would be extremely
unlikely to get such a result if the mean concentration were really the required 245. There is
only a probability of p-value= 0.0007 (i.e., 7 in 10,000) of that happening.
d) Making conclusion: Since our conclusion is based on how small the 𝑝 − 𝑣𝑎𝑙𝑢𝑒 is, or in other
words, how surprising our data are when 𝐻0 is true, it would be nice to have some kind of
guideline or cutoff that will help determine how small the p-value must be, or how "rare"
(unlikely) our data must be when 𝐻0 is true, for us to conclude that we have enough evidence
to reject 𝐻0 .
This cut-off exists, and because it is so important, it has a special name. It is called the
significance level of the test and is usually denoted by the Greek letter α. The most commonly
used significance level is α = 0.05 (or 5%). This means that:
1) If the p-value < α (usually 0.05), then the data we got is considered to be "rare (or
surprising) enough" when 𝐻0 is true, and we say that the data provide significant evidence
against 𝐻0 , so we reject 𝐻0 and accept 𝐻𝑎 .
2) If the p-value > α (usually 0.05), then our data are not considered to be "surprising enough"
when 𝐻0 is true, and we say that our data do not provide enough evidence to reject 𝐻0 (or,
equivalently, that the data do not provide enough evidence to accept 𝐻𝑎 ).
Notes:
1) Another common wording (mostly in scientific journals) is:
"The results are statistically significant" - when the p-value < α.
"The results are not statistically significant" - when the p-value > α.
2) I reject 𝐻0 and accept 𝐻𝑎 (when the p-value is smaller than the significance level) or I
cannot reject 𝐻0 (when the p-value is larger than the significance level).
3) In particular, note that in the second type of conclusion we did not say: "I accept claim
1," but only "I don't have enough evidence to reject claim 1. Saying (by mistake) "I don't
have enough evidence to reject 𝐻0 so I accept it" indicates that the data provide
evidence that 𝐻0 is true, which is not necessarily the case.
 Going back to our example and make a conclusion: Here, it is pretty clear that a sample like
the one we observed is extremely rare (or extremely unlikely) if the mean concentration in the
shipment were really the required 245 ppm. The fact that we did observe such a sample
therefore provides strong evidence against claim 1, so we reject it and conclude with very little
doubt that the mean concentration in the shipment is not the required 245 ppm.
3.2. Test Statistics: as I have mentioned above I will introduce the idea of finding the test statistics. Let’s
mention this in parts as:
a) Finding the test has many ways, in my work papers I mention the popular one p-value.
However as you will most probably know that for any method there is rule as:
Hypothesis If population parameters If sample statistics Notes
Null hypothesis 𝐻0 𝑃 = # (categorical) 𝑃 = # (categorical)
̂ Always using equal sign
𝜇 = # (Numerical) 𝑥̅ = # (Numerical)
𝜎=# 𝑠=#
Null hypothesis 𝐻𝑎 𝑃 < 𝑜𝑟 > 𝑜𝑟 ≠ # 𝑃̂ < 𝑜𝑟 > 𝑜𝑟 ≠ # Always using more, less or
𝜇 < 𝑜𝑟 > 𝑜𝑟 ≠ # 𝑥̅ < 𝑜𝑟 > 𝑜𝑟 ≠ # not equal sign
𝜎 < 𝑜𝑟 > 𝑜𝑟 ≠ # 𝑠 < 𝑜𝑟 > 𝑜𝑟 ≠ #
b) Level of significance alpha (α), can be 0.1, 0.05, or 0.10,
c) For alternative hypothesis it has left, right or two-tail test, depending on the sign, look at the
figure where I assume I’m dealing with population parameter 𝜇.

d) I will mention here the traditional method and the p-value method:
𝑃̂ −𝑃
a. Traditional method: 𝑧 = , this is depends on the critical value, in this method we
𝑃(1−𝑃)
√
𝑛

concern categorical variables.

Example ONE: Has the proportion of defective products been reduced as a result of
the repair?
𝐻0 : p = 0.20 (No change; the repair did not help).
𝐻0 : p < 0.20 (The repair was effective).
We collect sample n=400, found defect of 64. 𝑃̂ = 0.16, we calculate the standard
𝑃(1−𝑃) 0.80
deviation as 𝑠 = √ = 0.20 ∗ √ = 0.02, then we are ready to find 𝑧 =
𝑛 400
 𝑃̂−𝑃 0.20
= .16 − = −2, from z table we obtain value 𝑃(𝑧 ≤ −2) = 0.0228. thus we
𝑃(1−𝑃) 0.02
√
𝑛

draw a conclusion that 0.0228 < 0.05(α), then we have enough evidence to reject 𝐻0 .

Example TWO: in previous example we discussed when p< value so it was a left tail
test, in this example I will discuss when 𝑃 ≠ 𝑣𝑎𝑙𝑢𝑒 so it’s the case of two test tail. Next
point will discuss when P>value.
The dean of students in a certain liberal arts college was interested in whether the
proportion of students who use drugs in her college is different than the proportion
among U.S. college students in general. Suppose that in a simple random sample of
400 students from the college, 76 admitted to marijuana use. Do the data provide
enough evidence to conclude that the proportion of marijuana users among the
students in the college (p) differs from the national proportion, which is 0.157? (This
number is reported by the Harvard School of Public Health.)
𝐻0 : 𝑃 = 1.57
𝐻𝑎 : 𝑃 ≠ 1.57
Since we have the data (76 marijuana users out of 400), we have the sample proportion
and the same test statistic: 𝑃̂ = 0.19, 𝑧 = 1.81,
Since the calculation of the p-value depends on the type of alternative we have, here
is where things start to be different. Statistical software tells us that the 𝑝 −
𝑣𝑎𝑙𝑢𝑒 𝑖𝑠 (2 ∗ 0.35 = 0.070).
If we use the 0.05 level of significance, the p-value we got is not small enough (0.07 >
0.05), and therefore we cannot reject Ho. In other words, the data do not provide
enough evidence to conclude that the proportion of marijuana smokers in the college
is different from the national proportion (0.157).

𝑥̂−𝜇
b. P-value method: 𝑧 = , here we concern a numerical value. The p-value I obtain
𝜎/√𝑛
from the formula/ test statistic, according to our information above 3.1 from the
255−245
example 𝑧 = = 3.2, note that the p>value so we are looking at right tail. We
12/√64
have 𝑃(𝑧 ≥ 3.2) = 0.0007, from this point I can make the conclusion as our finding
above.
Thus the value 0.0007<0.05, that is the reason to reject the null hypothesis.

Note: an important note to make, if the standard deviation is unknown, so we can use the standard
deviation of the sample we have collected then instead of using Z table we use t-test table. The
𝑥̂−𝜇
formula is 𝑡 = . Few observations to make here:
𝑠/√𝑛
1. You can see in the picture that the t distribution has slightly less area near the expected central
value than the normal distribution does, and you can see that the t distribution has
correspondingly more area in the "tails" than the normal distribution does. (It’s often said that
the t distribution has "fatter tails" or "heavier tails" than the normal distribution.)
2. Therefore, the t distribution ends up being the appropriate model in certain cases where there
is more variability than would be predicted by the normal distribution. One of these cases is
stock values, which have more variability (or "volatility," to use the economic term) than would
be predicted by the normal distribution.
3. An important concept to mention once you are using t distribution is the degree of freedom.
Discussing this matter will be in upcoming work papers. See the picture below for further
understanding.

4. Remember that we said that higher degrees of freedom indicate that the t distribution is closer
to normal. So in the context of a test for the mean, the larger the sample size, the higher the
degrees of freedom, and the closer the t distribution is to a normal z distribution. This is
summarized with the notation near the bottom on the following image:

As a result, in the context of a test for a mean, the effect of the t distribution is most important
for a study with a relatively small sample size.
5. An important note to make how we calculate p-Value using t-distribution. No much changed
happened as we explained above in Z distribution. In case 𝑃(𝑡 >, <, ≠). Let’s see the three
pictures below to demonstrate this.
Note: we consider for the three cases (1) degree of freedom.
Note that due to the symmetry of the t distribution, for a given value of the test statistic t, the p-value for
the two-sided test is twice as large as the p-value of either of the one-sided tests. The same thing happens
when p-values are calculated under the t distribution as when they are calculated under the Z distribution.

Let’s find the same example we proposed above 3.1 with a little of change. Where at this time
we no longer assume that the population standard deviation is known, and instead use the
sample standard deviation of 12. Again, the problem was thus changed from a z-test problem
to a t-test problem. So there is no changes about the hypothesis 𝐻0 = 250, 𝐻𝑎 ≠ 250.
Observe the condition to use t distributions are met.
a) Sample is random,
b) Sample is large enough to follow Central Limit Theorem and ensure that 𝑥̅ is normally
distributed.
𝑥̂−𝜇 250
c) The 𝑡 = = 247 − 12 = −2.5,
𝑠/√𝑛
√100

d) The data (represented by the sample mean) are 2.5 standard errors below the null value.
e) Finding the p-value

To find the p-value we use statistical software, and we calculate a p-value of 0.014 with a 95%
confidence interval of (244.619, 249.381). For comparison purposes, the output we got when
we carried out the z-test for the same problem was a p-value of 0.012 with a 95% confidence
interval of (244.648, 249.352). Note that here the difference between the p-values is quite
negligible (0.002). This is not surprising, since the sample size is quite large (n = 100). The p-
value is small (0.014) indicating that at the 5% significance level, the results are significant. The
data therefore provide evidence to conclude that the mean concentration in entire shipment
is not the required 250, so we can reject the null-hypothesis.
See the picture below to summarize our finding:
Contents
Introduction ............................................................................................................................................................................................................................................................1

1. Point Estimation ........................................................................................................................................................................................................................................ 2

2. Interval Estimation .................................................................................................................................................................................................................................... 2

3. Hypothesis ................................................................................................................................................................................................................................................. 6

Bibliography......................................................................................................................................................................................................................................................... 13

Bibliography
Jessica M. Utts, Robert F. Heckard, 2002. Mind on Statistics. s.l.:s.n.

George McCabe, D. S. M., n.d. Introduction to the Practice of Statistics. 5 ed. s.l.:s.n.

Indian Institute of Science, B. I., n.d. NPTEL. [Online]

Available at: http://nptel.ac.in/copyright.php

University, S., 1985. Stanford University. [Online]

Available at:
http://cl.gsb.stanford.edu/exed/landing/gen.html?utm_campaign=evergreentoplevel2016&utm_source=google&utm_medium=cpc&
utm_content=digital-ad-google-paid-search-general&atrkid=V3ADWF3CB676F_30701431240_kwd-
366977048__129420103240_g_c___1t1&gclid=CKb826mL4c4

View publication stats