INTRODUCTION

:
The final outcome of an experiment or research depends to a large
extent on the analysis of comparative values obtained. To the dictum of
Helmholtz that “All science is measurement”, one can also add Sir Henry
ale!s clause that, “All true measurement is essentially comparative”.
Statistics plays an integral role in collection, presentation, analysis and
interpretation of comparative data. "n a broad sense, the term “statistics” has
al#ays been associated #ith studies related to facts and figures e.g. health
statistics, business statistics etc. in the boo$ “Statistical methods in medical
research” statistics has been defined as a discipline concerned #ith the
treatment of numerical data derived from groups of individuals or materials.
DESCRIPTIVE STATISTICS :
Statistics :
"s the science of compiling, classifying and tabulating numerical data
and expressing the results in a mathematical or graphical form.
%r
Statistics is the study of methods and procedures for collecting,
classifying, summarizing and analyzing data and for ma$ing scientific
inferences from such data.
& 'rof. '.(. Su$hatme
Biostatistics :
"s the branch of statistics applied to biological or medical sciences
)biometry*.
"s that branch of statistics concerned #ith mathematical facts and data
relating to biological events.
Variable :
A general term for any feature of the unit, #hich is observed or
measured.
Frequency istribution :
istribution sho#ing the number of observations or fre+uencies at
different values or #ithin certain range of values of the variable.
!ean :
ividing the total of all observations by the number of observations
,g. -alculate the mean of .(T scores /.0, /.1, /.2, 0.1, /.1.
3eometric mean )3.* 4 nth root of the product
3. 5 )x
6
* )x
/
* )x
0
* 77.)x
n
* 5
8hen the variation bet#een the lo#est and the highest value is very
high, geometric mean is advised and preferred.
Harmonic mean )H.* 4 is the reciprocal of the arithmetic mean of the
reciprocal of the observations.
!eian :
"s the middle value, #hich divides the observed values into t#o e+ual
parts, #hen the values are arranged in ascending or descending order.
,g. -alculate the median of .(T scores /.0, /.1, /.2, 0.1, /.1 arrange in
asc order,
/.1, /.1, /.0, /.2, 0.1 5 5 0
rd
value 5 i.e. /.0
x 5
x
6
9 x
/
9 x
0
9 7777. 9 x
n
n
5
∑x
n
x 5
/.0 9 /.1 9 /.2 9 0.1 9 /.1
:
5
6/
:
5 /.;
∑ log x
n
n
H. 5
6

6
n
6
x
6
5
6
∑ 6
x
6
n 9 6
/
: 9 6 9 <
/
!oe :
"s the value of the variable #hich occurs most fre+uently
,g. -alculate the mode of .(T scores /.0, /.1, /.2, 0.1, /.1
.ode 5 /.1
.ode 5 0 x /.0 4 / x /.; 5 /.6
.ode 5 )0 median 4 / mean*
Variance :
• "s the appropriate measure of dispersion for interval or ratio level data
• -omputes ho# far each score is from the man
o This is done by
)x 4 x *
o ,ach score #ill have a deviation from the mean, so to find the
average deviation  #e have to add all the deviations and divide it
by number of scores )=ust li$e calculating mean*
i.e.
but 7.. ∑ )x 4 x * 5 1
i.e. 5 S
/
So to eliminate this zero, s+uare the deviations #hich eliminates the )&* sign.
& "n other #ords it is the average of the s+uared deviations.
Stanar e"iation #Root $ean square e"iation%
• "s defined as the s+uare root of the arithmetic mean of the s+uared
deviations of the individual values from their arithmetic mean
(or small samples S 5 )σ* 5
∑ )x 4 x *
>
∑ )x 4 x *
/
>
∑ )x 4 x *
/
> & 6
∑ )x 4 x *
/
>
(or large samples S 5 )s* 5
8hen there is fre+uency distribution
For s$all sa$&les S 5 )σ* 5
For lar'e sa$&les S 5 )s* 5
?sed of S
• Summarizes the deviations of a large distribution from mean in one
figure used as unit of freedom.
• "ndicates #hether the variation from the mean is by chance or real.
• Helps finding standard error 4 #hich determines #hether the difference
b@n means of t#o samples is by chance or real.
• Helps finding the suitable size of the sample for value conclusions.
Stanar error :
• Standard deviation of mean values
S, 5 5
• ?sed to compare means #ith one another
Coe((icient o( "ariation is a measure used to measure relative variability
i.e.
• Aariation of same character in t#o or more different series. )eg 4 pulse
rate in young ad old person*
• Aariation of t#o different character in one and same series. )eg 4 height
and #eight in same individual*.
-A 5 x 611
∑ ƒ )x 4 x *
/
> & 6
∑ ƒ )x 4 x *
/
>6
Standard deviation
Sample size
S
n
Standard deviation
.ean
Nor$al cur"e an istribution :
• the histogram of the same fre+uency distribution of heights, #ith large
number of observations and small class intervals 4 gives a fre+uency
curve #hich is symmetrical in nature  Nor$al cur"e or )aussian
cur"e.
C*aracteristics o( nor$al cur"e :
• Bell shaped
• Symmetrical
• .ean, .ode and .edian 4 coincide
• Has t#o inflections 4 the central part is convex, #hile at the point of
inflection the curve changes from convexity to concavity.
%n preparing fre+uency distribution #ith small class intervals of the
data collected, #e can observe
6* Some observations are above the mean and others are belo# the mean
/* "f arranged in order, maximum number of fre+uencies is seen in the
middle around the mean and fe#er at the extremes decreasing
smoothly
0* >ormally half the observations lie above and half belo# the mean and
all are symmetrically distributed on each side of mean
An distribution of this nature or shape is called Nor$al or )aussian
istribution
• Arithmetically
.ean ± 6S limits, include <C./2D observations
.ean ± /S limits, include E:.;:D observations
.ean ± 6.E<S limits, include E:D observations
.ean ± 0S limits, includes EE.20D observations
.ean ± /.:CS limits, includes EED observations
Nor$al cur"e an istribution :
Felative or standard normal deviate or variate 4 )G*
"s the deviation from the mean in a normal distribution or curve
G 5 5
& "n standard normal curve the mean is ta$en as zero and S as unity of one
S+e,ness - is the static to measure the asymmetry
• -oefficient of s$e#ness is 1
'ositivity )right* s$e#ed >egativity )left* s$e#ed Bimodal
.urtosis 4 is a measure of height of the distribution curve
• -oefficient of $urtosis is 0
Hepto$urtic )high* 'laty$urtic )flat* .eso$urtic )normal*
TESTS OF SI)NIFICANCE :
'opulation is any finite collection of elements
i.e. 4 individuals, items, observations etc.,
Sample 4 is a part or subset of the population
'arameter 4 is a constant describing a population
Statistic 4 is a +uantity describing a sample, namely a function of observations
Statistic #)ree+% Para$eter #/atin%
.ean x
µ
Standard eviation
σ
Aariance s
/
σ
/
-orrelation coefficient r
ρ
>umber of sub=ects
η
>
%bservation 4 mean
S
x 4 x
S
01POT0ESIS TESTIN) :
0y&ot*esis
"s an assumption about the status of a phenomenon or is a statement
about the parameters or form of population.
>ull hypothesis or hypothesis of no difference 4 States no difference
bet#een statistic of a sample and parameter of population or b@n statistics of
t#o samples.
• This nullifies the claim that the experiment result is different from or
better than one observed already.
• enoted by H
1

Alternate *y&ot*esis - any hypothesis alternate to null hypothesis, #hich
is to be tested
• enoted by H
6

Note : the alternate hypothesis is accepted #hen null hypothesis is re=ected
Ty&e I an ty&e II errors
H
1
accept H
6
accept
H
1
is true >o error Type 6 error
H
6
is true Type "" error >o error
• Type " error 5 α
• Type "" error 5 β
8hen primary concern of the test is to see #hether the null
hypothesis can be re=ected such test is called Test of significance.
The probability of committing type " error is called 2P3 "alue.
The p&value is the chance that the presence of difference is concluded
#hen actually there is none.
Type " error is important and it is fixed in advance at a lo# level, such
upper limit of tolerance of the chance of type " error is called Hevel of
Significance )α*
Thus α is the maximum tolerable probability of type " error
DIFFERENCE B4N /EVE/ OF SI)NIFICANCE AND P5VA/UE
/e"el o( si'ni(icance
6* .aximum tolerable chance of type " error
/* α is fixed in advance
P5"alue :
6* Actual probability of type " error
/* -alculated on basis of data follo#ing procedures
The '&value can be more than α or less than α depending on data,
8hen '&value is less than α result is statistically significant.
• The level of significance is usually fixed at :D )1.1:* or 6D )1.16* or
1.6D )1.116* or 1.:D )1.11:*
• .aximum desirable is :D level
• 8hen '&value is b@n
1.1:&1.16 5 statistically significant
I than 1.16 5 highly statistically significant
lo#er than 1.116 or 1.11: 5 very highly significant
TESTS OF SI)NIFICANCE :
• Are mathematical methods by #hich the probability )'* or relative
fre+uency of an observed difference, occurring the chance is found
• Ste&s 6 &roceure o( test o( si'ni(icance 4
6* State null hypothesis
/* State alternate hypothesis
0* Selection of the appropriate test to be utilized and
calculation of test criterion based on type of test.
;* (ixation of level of significance
:* Select the table and compare the calculated value #ith
the critical value of the table
<* "f calculated value is J table value, H
1
is re=ected
2* "f calculated value is I table value, H
1
is re=ected
C* ra# conclusions
STEPS IN STATISTICA/ STUD1 :
The chronology of steps in involved in a statistical study are as
follo#s K
Selection o( sa$&le si7e :
%ften, the primary problem encountered by a student of research is
the number of samples or sample size to be selected.
-riteria for selection of sample size are as belo# K
• A sample size of /:&01 in each group is ade+uate if there is one variable
or one parameter in the study.
• "n invivo studies #here there is less availability of samples, a slight
decrease in sample size may be acceptable.
• Harger sample size #ill be needed if
o Harger variation is expected
o Fare characteristic is present
TESTS IN TEST OF SI)NIFICANCE
'arametric
)normal distribution L >ormal culture*
>on&parametric
)not follo# normal distribution*
Muantitative data
Student Nt! test
'aired
?npaired
G test )for large samples*
%ne #ay A>%AA
T#o #ay A>%AA
Mualitative data
G 4 prop test
χ
/
test
Mualitative
)+uantitative converted to +ualitative*
.ann 8hitney ? test
8ilcoxon ran$ test
Orus$al #allis test
(riedmann test
o .ore variable are present
o .ore precision re+uired
o .ore reliability re+uired
Selection o( test :
The tests employed to complete a study can be classified as K
A. (or comparison of mean )average of observations* of
different samples
B. (or comparison of proportion )percentage* of different
samples.
-. -orrelation tests
. Fegression tests
A8 For co$&arison o( $ean #a"era'e o( obser"ations% o( i((erent
sa$&les :
T#o types of tests are available, namely parametric and non&parametric
Para$etric tests Non &ara$etric tests
• ,mployed if the distribution of
the population from #hich the
samples are dra#n is $no#n. )i.e.
normally distributed #ith less
variation*.
• ,mployed if distribution is
un$no#n )large variation present*
• "n the computation of parametric
tests the arithmetic processes of
addition, division and
multiplication are used.
• ata are changed from
measurements or scores to ran$s
or even to signs.
• ?sed if ade+uate sample size is
present.
• ?sed if ade+uate sample size is
not present
i% Ine&enent t5test
,mployed to compare mean of
t#o groups using one variable.
,g. -omparison of bond strength
of amalgam and composite.
i% !ann 9*itney U test
"t is e+uivalent to independent t&
test.
ii% Paire t5test
Use (or ,it*in 'rou& co$&arison
at different time intervals. ,g.
>umber of microbes in root canal
before and after antibiotic therapy.
ii% 9ilo:an si'n ran+ test
It is equi"alent to &aire t5test8
iii% ANOVA #Analysis o( "ariance%
.ean of any number of groups using
one variable is determined by this
test. ,g. Sealing capacity of different
endodontic sealers.
iii% .rus+al 9allis 05test
"t is e+uivalent to A>%AA.
A8 For co$&arison o( &ro&ortion #&ercenta'e% o( i((erent sa$&les :
The test employed are as follo#s K
i% C*i5square test :
"t chec$s the proportion bet#een any number of groups using one
variable. "t is used if ade+uate sample size is available.
,g. The effect of ampicillin, sulphonamides and tetracycline in a
certain percentage of people.
ii% Fis*er;s test :
This test is similar to -hi&s+uare test if less sample size is available.
iii% !c5Ne$ar test :
"t compares proportion of one variable #ithin a group at different
time intervals.
E'8 Pro&ortion o( &eo&le *a"in' sensiti"ity be(ore an a(ter usin'
esensiti7in' &aste8
B8 Correlation tests :
"t is used to find if t#o variables co&vary #ith each other or are independent.
,g. The susceptibility rate of organisms in root canal follo#ing
increase in antibiotic dosage.
C8 Re'ression tests :
"t describes the dependence of one variable on another independent variable.
,g. ,ffect of bonding agent on strength of composite.
ANA/1SIS OF RESU/T :
• >on&parametric tests li$e .ann 8hitney ?&test, 8iloxan sign ran$ test,
Orus$al 8allis H&test are less sensitive than parametric tests li$e
"ndependent t&test, A>%AA as they use random ran$ing instead of
original values.
• 'robability )p* value indicates level of significance )sensitivity* of a test.
& < =8==>  0i'*ly si'ni(icant
The probability that the difference bet#een t#o groups
occurring by chance is less than 6 in 6111.
& < =8=>  !oerately si'ni(icant
The probability of the difference occurring by chance is
less tan 6 in 611.
& < =8=?  /ess si'ni(icant
The probability of the difference occurring by chance is
less than : in 611.
& < =8=?  Not si'ni(icant
The probability of difference occurring by chance is very
high.
CONC/USION :
This seminar attempts at explaining importance of statistics as an
essential protocol for any research program. Statistics is the greatest leveler.
"t covers up for all the variations that can creep into the results thereby
providing a foolproof system for proper interpretation of data. Hence it
#ould be appropriate to term it as “The Aital Statistics”.
VITA/ STATISTICS
Outline
• INTRODUCTION
• DESCRIPTIVE STATISTICS
• TESTS OF SI)NIFICANCE
• STEPS IN STATISTICA/ STUD1
o SE/ECTION OF SA!P/E SI@E
o SE/ECTION OF TEST
o ANA/1SIS OF RESU/T
• CONC/USION
CO//E)E OF DENTA/ SCIENCES
DEPART!ENT OF CONSERVATIVE DENTISTR1 AND
ENDODONTICS
S,.">AF
ON
VITAL STATISTICS

H
1
accept H
6
accept
H
1
is true >o error Type 6 error
H
6
is true Type "" error >o error

PRESENTED B1 :
Dr8 Si*es,aran V8