statistic introduction
TRANSCRIPT
-
8/9/2019 Statistic Introduction
1/68
-
8/9/2019 Statistic Introduction
2/68
WE RESEARCHERSUSE STATISTICS THE
WAY A DRUNKARDUSES A LAMP POST,
MORE FOR SUPPORT
THAN
ILLUMINATION.
WE RESEARCHERSUSE STATISTICS THE
WAY A DRUNKARD
USES A LAMP POST,
MORE FOR SUPPORT
THANILLUMINATION.
-
8/9/2019 Statistic Introduction
3/68
StatisticsStatistics
The branch of mathematicsThe branch of mathematics
that deals with thethat deals with the
collectioncollection,,organizationorganization,,
analysisanalysis, and, andinterpretationinterpretation
of numerical data.of numerical data.
-
8/9/2019 Statistic Introduction
4/68
StatisticsStatistics
is especiallyis especially
useful in drawinguseful in drawing
general conclusionsgeneral conclusions
about a set of dataabout a set of data
from a sample of the data.from a sample of the data.
-
8/9/2019 Statistic Introduction
5/68
DATADATASINGULARSINGULAR ----------DATUM.DATUM.
PLURALPLURAL ----------------DATA.DATA.
WE MAY DEFINE DATAASWE MAY DEFINE DATAASNUMBERSAND THERE ISNUMBERSAND THERE ISTWO KINDS OF NUMBERSTWO KINDS OF NUMBERS
THAT WE USE INTHAT WE USE INSTATISTICS THE RESULTS OFSTATISTICS THE RESULTS OF: COUNTINGAND: COUNTINGAND
MEASUREMENTSMEASUREMENTS..
-
8/9/2019 Statistic Introduction
6/68
variablevariable
ANYASPECT OF ANANYASPECT OF ANINDIVIDUAL THAT ISINDIVIDUAL THAT IS
MEASURED, LIKEMEASURED, LIKEBLOOD PRESURE,BLOOD PRESURE,
AGE,AGE,
SEX etc.SEX etc.
-
8/9/2019 Statistic Introduction
7/68
Variables divide intoVariables divide into
different typesdifferent types
QUALITATIVEQUALITATIVE
(CATEGORICAL)(CATEGORICAL)
QUANTITATIVEQUANTITATIVE(NUMERICAL)(NUMERICAL)
-
8/9/2019 Statistic Introduction
8/68
ReadRead
not to contradictnot to contradictand confute,and confute,
not to believenot to believe
and take for granted,and take for granted,
not to find talk and discourse,not to find talk and discourse,
but to weigh and consider.but to weigh and consider.
Sir Francis BaconSir Francis Bacon
-
8/9/2019 Statistic Introduction
9/68
WHAT MAKE STATISTICS UNIQUE?WHAT MAKE STATISTICS UNIQUE?
ITSABILITY TOITSABILITY TO
QUANTIFYQUANTIFYUNCERTAINTYUNCERTAINTY,,
TO MAKE ITTO MAKE ITPRECISEPRECISE..
-
8/9/2019 Statistic Introduction
10/68
-
8/9/2019 Statistic Introduction
11/68
A more positive example in statistics isA more positive example in statistics isthethe SALK POLIO VACCINESALK POLIO VACCINE in 1954in 1954
vaccine trials were performed on somevaccine trials were performed on some400,000 children, with strict controls to400,000 children, with strict controls toeliminate biased results.eliminate biased results.
Good statistical analysisGood statistical analysis ofof thetheresults firmly established theresults firmly established the
vaccines effectiveness,vaccines effectiveness,and to dayand to day
POLIO is almost unknown.POLIO is almost unknown.
-
8/9/2019 Statistic Introduction
12/68
To accomplishtheirfeats of mathematicalTo accomplishtheirfeats of mathematicalLEGERDEMAIN (STATISTICIANS)LEGERDEMAIN (STATISTICIANS)
RELY ON THREE RELATED DISCIPLINE:RELY ON THREE RELATED DISCIPLINE:
Data analysisData analysisThe gathering, display, and summaryThe gathering, display, and summary
of dataof data
ProbabilityProbabilityThe law of chanceThe law of chance
Statistical InferenceStatistical InferenceThe science of drawing statisticalThe science of drawing statistical
conclusions from specific data, using aconclusions from specific data, using a
knowledge of probability.knowledge of probability.
-
8/9/2019 Statistic Introduction
13/68
Objectives ofthiscourseObjectives ofthiscourse
At the end of the course the studentsAt the end of the course the studentswill be able towill be able to
1.1. organize dataorganize data2.2. summarize datasummarize data
3.3. reach decision aboutreach decision about
a large body of dataa large body of data
by examiningby examining
only small part of the data.only small part of the data.
-
8/9/2019 Statistic Introduction
14/68
Objective number 1 and 2Objective number 1 and 2
we will discuss inwe will discuss in
DESCRIPTIVE STATISTICS.DESCRIPTIVE STATISTICS.
Objective number 3Objective number 3
we will discusswe will discuss
INFERENCIALSTATISTICSINFERENCIALSTATISTICS
-
8/9/2019 Statistic Introduction
15/68
What is biostatisticsWhat is biostatistics
-
8/9/2019 Statistic Introduction
16/68
-
8/9/2019 Statistic Introduction
17/68
TheThe statisticsstatistics partpart
involvesinvolves
CCollectionollection,,
OrganizationOrganization,, AAnalysisnalysis,,
andand
interpretationinterpretation
of numerical dataof numerical data
-
8/9/2019 Statistic Introduction
18/68
BiostatisticsBiostatisticsis the applicationis the application
ofofstatisticsstatistics
to a wide range of topics into a wide range of topics inbiologybiology. It has particular. It has particularapplications toapplications to medicinemedicine
andand
toto agricultureagriculture..
-
8/9/2019 Statistic Introduction
19/68
StatisticsStatistics calculationscalculations
are an importantare an important
part of data analysis,part of data analysis,
butbut interpreting datainterpreting dataalsoalso
requiresrequiresa great deal of judgmenta great deal of judgment
-
8/9/2019 Statistic Introduction
20/68
UnderstandingUnderstanding
the statistical calculationsthe statistical calculations
is only a small partis only a small part
ofof1.1. evaluating clinicalevaluating clinical
2.2. pharmaceuticalpharmaceutical
3.3. biological research.biological research.
-
8/9/2019 Statistic Introduction
21/68
WhyWhyis it hardis it hard
to learnto learn
statistics?statistics?
-
8/9/2019 Statistic Introduction
22/68
Theterminology is deceptiveTheterminology is deceptive
You have to understand :You have to understand :
1.1. SignificantSignificant
2.2. ErrorError3.3. HypothesisHypothesis
4.4. Null hypothesisNull hypothesis5.5. Confidence intervalConfidence interval
-
8/9/2019 Statistic Introduction
23/68
Level of significantLevel of significant
p valuep value
PopulationPopulation
SampleSample
Paired and unpairedPaired and unpaired
samplessamples
-
8/9/2019 Statistic Introduction
24/68
RememberRememberThe phraseThe phrase
statistically significantstatistically significant
isisseductiveseductive
andand
isis
oftenoften misinterpretedmisinterpreted..
-
8/9/2019 Statistic Introduction
25/68
StatisticsStatistics
isisat the interfaceat the interface
ofofmathematicsmathematics
andandsciencescience
-
8/9/2019 Statistic Introduction
26/68
HoweverHoweverYou can learnYou can learn
toto useuse
statistical teststatistical test
and interpretand interpretthe resultsthe results
eveneven
if you dont fully understandif you dont fully understand
how they work.how they work.
-
8/9/2019 Statistic Introduction
27/68
Inductiv
e reasoningInductiv
e reasoning
-
8/9/2019 Statistic Introduction
28/68
often describedoften describedasas
"going from the specific"going from the specificto the general."to the general."
ANDANDIt is based onIt is based on
-
8/9/2019 Statistic Introduction
29/68
observingobserving
specific instancesspecific instancesof a certain qualityof a certain quality
in individual membersin individual members
of a group ofof a group of
1.1. peoplepeople
2.2. animals oranimals or
3.3. eventsevents
-
8/9/2019 Statistic Introduction
30/68
notingnoting
the individual membersthe individual members
in whichin which
a certain qualitya certain quality
occursoccurs
belong to a certain groupbelong to a certain group
-
8/9/2019 Statistic Introduction
31/68
generalizinggeneralizing
to the conclusionto the conclusion
thatthat
other membersother members
of that groupof that group
have the same quality.have the same quality.
-
8/9/2019 Statistic Introduction
32/68
For exampleFor example
if you were to go to a cat show,if you were to go to a cat show,you would see many breedsyou would see many breeds
of cats with tails.of cats with tails.
After walking up and down,After walking up and down,
you might begin to noticeyou might begin to notice
a pattern,a pattern,and your reasoningand your reasoning
might go something like this:might go something like this:
-
8/9/2019 Statistic Introduction
33/68
Siamese cats have tails.Siamese cats have tails.
Persian cats have tails.Persian cats have tails.
Himalayan cats have tails.Himalayan cats have tails.
Russian Blues have tails.Russian Blues have tails.American Tabbies have tails.American Tabbies have tails.
-
8/9/2019 Statistic Introduction
34/68
Aftera while, you would probablyAftera while, you would probably
cometo theconclusioncometo theconclusion --
All catsAll cats
have tailshave tails
-
8/9/2019 Statistic Introduction
35/68
The problem with thisThe problem with this
conclusion isconclusion isthat it isn't true.that it isn't true.
TheThe manxmanxis one breed of catis one breed of cat
that has no tail.that has no tail.
-
8/9/2019 Statistic Introduction
36/68
So the conclusionSo the conclusion
ofof
an inductive argumentan inductive argumentcan be shown to be wrongcan be shown to be wrong
if only one instanceif only one instance
does not fitdoes not fit
the general pattern.the general pattern.
-
8/9/2019 Statistic Introduction
37/68
TheTheManxManx isaisabreedbreed ofcats withanaturallyofcats withanaturally
occurringoccurring mutationmutation oftheofthespinespine..
-
8/9/2019 Statistic Introduction
38/68
ForthisreasonForthisreasonthe result ofthe result of
anan inductive argumentinductive argument
is never consideredis never considered
to be TRUE or FALSEto be TRUE or FALSE
insteadinstead
we refer to the conclusions reachedwe refer to the conclusions reachedthroughthrough
inductive reasoninginductive reasoning
asas
MORE ORLESSRELIABLE.MORE ORLESSRELIABLE.
-
8/9/2019 Statistic Introduction
39/68
Before proceeding,thinkaboutthisBefore proceeding,thinkaboutthis
question:question:
What would make theWhat would make theconclusion from an inductiveconclusion from an inductive
argument MORE reliable?argument MORE reliable?
-
8/9/2019 Statistic Introduction
40/68
Ininductivereasoning,Ininductivereasoning,
the morethe more specificspecific instancesinstances
you observe,you observe,
the more reliablethe more reliable
your conclusion.your conclusion.
-
8/9/2019 Statistic Introduction
41/68
BecauseBecause
the conclusion fromthe conclusion from
an inductive argumentan inductive argument
cannot be consideredcannot be considered
true or falsetrue or falsethe conclusion must be qualified.the conclusion must be qualified.
-
8/9/2019 Statistic Introduction
42/68
TheconclusionTheconclusion
from your observations about catsfrom your observations about catsand their tails might be qualifiedand their tails might be qualified
in one of the following ways:in one of the following ways:
-
8/9/2019 Statistic Introduction
43/68
-
8/9/2019 Statistic Introduction
44/68
A conclusionA conclusion
that isthat is
more or lessmore or less
reliablereliableThe more specific instancesThe more specific instances
observedobserved
the more reliable the conclusion.the more reliable the conclusion.
-
8/9/2019 Statistic Introduction
45/68
-
8/9/2019 Statistic Introduction
46/68
Inductive Reasoningallows you to learnallows you to learn
something newsomething new
about the world.about the world.
Deductive Reasoningallows you to apply whatallows you to apply what
you have learned.you have learned.
-
8/9/2019 Statistic Introduction
47/68
-
8/9/2019 Statistic Introduction
48/68
The first premiseThe first premise
("All men are mortal.")("All men are mortal.")
is the resultis the result
ofof
inductive reasoning.inductive reasoning.
-
8/9/2019 Statistic Introduction
49/68
The second premiseThe second premiseidentifiesidentifies
a specific membera specific member
of that group (Abdullah).of that group (Abdullah).
-
8/9/2019 Statistic Introduction
50/68
-
8/9/2019 Statistic Introduction
51/68
In order for the conclusionIn order for the conclusion
ofof
a deductive reasoning processa deductive reasoning process
to be true,to be true,
all of its premisesall of its premises
must be true.must be true.
..
-
8/9/2019 Statistic Introduction
52/68
Most peopleMost people
would agreewould agree
thatthat
the first premise giventhe first premise given
is true.is true.
-
8/9/2019 Statistic Introduction
53/68
Most peopleMost people
would also agreewould also agree
thatthat
the second premisethe second premise
given is truegiven is true
-
8/9/2019 Statistic Introduction
54/68
IF both of the premises are true,IF both of the premises are true,
thenthenthe conclusion must also be truethe conclusion must also be true
if and only if it follows necessarilyif and only if it follows necessarily
fromfromthe informationthe information
given in the premises.given in the premises.
-
8/9/2019 Statistic Introduction
55/68
SupposeSuppose
we return to our observationswe return to our observations
about catsabout cats
and their tailsand their tails
and the conclusionand the conclusionwe arrived atwe arrived at
by usingby using
inductive reasoning:inductive reasoning:
-
8/9/2019 Statistic Introduction
56/68
Most cats haveMost cats have
tails.tails.
Remember that thisRemember that this
conclusion had to beconclusion had to bequalified.qualified.
Ra is a cat.Ra is a cat. I could show you pictures.I could show you pictures.
Ra has a tail.Ra has a tail. Therefore,Therefore,
this conclusionthis conclusion
follow necessarily?follow necessarily?
-
8/9/2019 Statistic Introduction
57/68
A trick question:A trick question:
When it rains, the streets get wet.When it rains, the streets get wet.
The streets are wet.The streets are wet.
Therefore,Therefore,
it has been raining.it has been raining.
Is the conclusion valid?Is the conclusion valid?
Why or why not?Why or why not?
-
8/9/2019 Statistic Introduction
58/68
Theconclusionto thattrickTheconclusionto thattrick
questionis NOT valid:questionis NOT valid:
When it rains, the streets getWhen it rains, the streets getwet.wet.
The streets are wet.The streets are wet.
Therefore, it has been raining.Therefore, it has been raining.
-
8/9/2019 Statistic Introduction
59/68
The conclusion of a deductiveThe conclusion of a deductive
argument must follow necessarilyargument must follow necessarilyfrom the argument's premises;from the argument's premises;however, the first premise in thishowever, the first premise in thisargument is a conditional statement.argument is a conditional statement.
(It gives a condition under which(It gives a condition under whichsomething becomes true.) Thesomething becomes true.) Thestatement can be restated like this:statement can be restated like this:
-
8/9/2019 Statistic Introduction
60/68
IF it rains,IF it rains,
THENTHEN
the streets get wet.the streets get wet.
-
8/9/2019 Statistic Introduction
61/68
The problem withtheconclusionThe problem withtheconclusion
from theaboveargumentfrom theaboveargument
isthatthereare otherconditionsisthatthereare otherconditionswhich may causethestreetsto get wet:which may causethestreetsto get wet:
-
8/9/2019 Statistic Introduction
62/68
snow, sleet or other forms ofsnow, sleet or other forms ofprecipitation besides rainprecipitation besides rain
fire hydrants being openedfire hydrants being opened
people washing their cars in the streetspeople washing their cars in the streets
floodingflooding
-
8/9/2019 Statistic Introduction
63/68
-
8/9/2019 Statistic Introduction
64/68
Giving the second part of aGiving the second part of aconditional statementconditional statement
(the THEN part),(the THEN part),results inresults in
an invalid argument.an invalid argument.
-
8/9/2019 Statistic Introduction
65/68
Anotherway of looking atthisAnotherway of looking atthis
problem isto thinkinterms ofproblem isto thinkinterms of
our firstargumentour firstargument(about Abdullah):(about Abdullah):
-
8/9/2019 Statistic Introduction
66/68
Step in ProcessStep in Process SocratesSocratesArgumentArgument
Rain ArgumentRain Argument
Identify a sharedIdentify a sharedquality of the set:quality of the set:
All men areAll men aremortalmortal
All rainy daysAll rainy dayscause the streetscause the streetsto get wet.to get wet.
Identify a memberIdentify a memberof the set:of the set:
Abdullah is a manAbdullah is a man It is raining today.It is raining today.
Valid conclusion:Valid conclusion: Abdullah isAbdullah ismortal.mortal.
The streets areThe streets arewet.wet.
-
8/9/2019 Statistic Introduction
67/68
All men are mortal.All men are mortal.
Abdullah is mortal.Abdullah is mortal.
Therefore,Therefore,Abdullah is a man.Abdullah is a man.
-
8/9/2019 Statistic Introduction
68/68
Thisconclusion doesnot followThisconclusion doesnot follow necessarilynecessarily
from the premisesfrom the premises
becausebecause
Abdullahcould beAbdullahcould be
acat,acat,
a mule,a mule,a dog,a dog,
orany other living thing.orany other living thing.