8/9/2019 Choosing Statistical Procedures for Research Problems
1/17
STATISTICALPROCEDURES FOR
RESEARCH PROBLEMS
8/9/2019 Choosing Statistical Procedures for Research Problems
2/17
Each statistical procedure has a setassumptions for its appropriate app
In selecting among procedures, anyresearcher must therefore considernumber of factors such as the follow
General and Specic Purpose of Res ype of !uestions"#ypothesis General ype of Statistic
8/9/2019 Choosing Statistical Procedures for Research Problems
3/17
Research $uestions can be di%ided intobroad types:
&I''ERE()E◦ Is * di+erent from -
◦ Is there a di+erence between * and -
◦ *re there di+erences among *, and )-
*SS)I*I(*/◦ Is * associated or related with -
◦ Is there a correlation between * ad -
◦ 0hat is the relationship between * and -
8/9/2019 Choosing Statistical Procedures for Research Problems
4/17
&ES)RIPI1E◦ 0hat is the mean %alue of *-
◦
0hat are the usual %alues in *-
8/9/2019 Choosing Statistical Procedures for Research Problems
5/17
GE(ER*/
P2RPSE
E4P/RE RE/*I(S#IP E0EE(
1*RI*/ES
&
SPE)I'I)
P2RPSE
)8P*RE
GR2PS
'I(& SRE(G#
'
*SS)I*I(S,
RE/*E 1*RI*/E
S28
6PE '!2ESI("
#6P#ESIS
&I''ERE()E *SS)I*I(*/ &
GE(ER*/ 6PE '
S*ISI)
&I''ERE()E
I('ERE(I*/
S*ISI)S
*SS)I*I(*/
I('ERE(I*/
S*ISI)S
&
S
LEVEL OF DIFFERENCE ASSOCIATIONAL D
8/9/2019 Choosing Statistical Procedures for Research Problems
6/17
LEVEL OF
MEASUREMEN
T
DIFFERENCE
INFERENTIAL
STATISTICS
ASSOCIATIONAL
INFERENTIAL
STATISTICS
D
S
Nominal
• Chi-Squar
;test for
proportion
Phi )oe
8/9/2019 Choosing Statistical Procedures for Research Problems
7/17
Di.rn! r$ar!h qu$#ion$/these $uestions, we compare score
the dependent %ariable7 of two or mdi+erent groups, each of which iscomposed of indi%iduals with one of%alues or le%els on the independent
%ariable. his type of $uestion attemto demonstrate that groups are notsame on the dependent %ariable.
8/9/2019 Choosing Statistical Procedures for Research Problems
8/17
A$$o!ia#ional r$ar!h qu$#ion$/these $uestions, we associate or relate
or more %ariables. his approach usuain%ol%es an attempt to see how two ormore %ariables co;%ary 5e.g., higher %aon one %ariable correspond to higher o
lower %alues on another %ariable for thsame persons7 or how one or more%ariables enables one to predict anoth%ariable
8/9/2019 Choosing Statistical Procedures for Research Problems
9/17
D$!ri0#i) r$ar!h qu$#ion$/ hese are not answered with infere
statistics. hey merely describe orsummarie data, without trying togeneralie to a larger population ofindi%iduals
8/9/2019 Choosing Statistical Procedures for Research Problems
10/17
* )aria1l is dened as a characteristic of the partisituation for a gi%en study that has di+erent %alues study and must be able to %ary or ha%e di+erent %al
Gender: 8ale or 'emale *ge: has a large number of %alues ype of reatment"Inter%ention
In $uantitati%e research, %ariables are dened opera
and are commonly di%ided into in'0n'n# )aria5acti%e or attribute7, '0n'n# )aria1l$, and 2)aria1l$.
8/9/2019 Choosing Statistical Procedures for Research Problems
11/17
two types of independent%ariables
*)I1E *RI2E
8/9/2019 Choosing Statistical Procedures for Research Problems
12/17
*lso >nown as manipulated indepen%ariable.
*n acti%e independent %ariable is a%ariable such as wor>shop, newcurriculum, or other inter%ention, atleast one le%el of which is gi%en to a
group of participants, within in aspecied period of time during the s
A!#i) in'0n'n# )aria1l$
8/9/2019 Choosing Statistical Procedures for Research Problems
13/17
*lso >nown as measure independent %a%ariable that cannot be manipulated, ye
ma?or focus of the study. he %alues of the attribute independent
are pre;e@isting attribute of the personsongoing en%ironment that are not syste
changed during the study. E@amples are education, gender, age, e
group, I! and self;esteem. Studies with attribute independent %ariables are calle
20rimn#al studies.
A##ri1u# in'0n'n# )aria1l$
8/9/2019 Choosing Statistical Procedures for Research Problems
14/17
he dependent %ariable is assumed to measureassess the e+ect of the independent %ariable.
It is thought of as the presumed outcome or cr &ependent %ariables are often test scores, ratin
$uestionnaires, readings from instruments or mof physical performance.
/i>e independent %ariables, dependent %ariable
ha%e at least two %aluesA most dependent %ariamany %alues, %arying from low to high so they easy to diagram as the independent %ariable.
D0n'n# Varia1l
8/9/2019 Choosing Statistical Procedures for Research Problems
15/17
E@traneous %ariables 5also >nown as nuisancco%ariates in some designs7 that are not of i
a particular study but could inBuence the de%ariable. En%ironmental factors 5e.g. temperature or
distractions7, time of day and characteristicse@perimenter, teacher or therapist are some
e@traneous %ariables that need to be contro
E2#ranou$ Varia1l
8/9/2019 Choosing Statistical Procedures for Research Problems
16/17
P*R*8ERI) ES
Independent 1ariableis between;sub?ects
Independent 1ariablehas C categories
Independent groups t-test
Independent 1ariable
has D or morecategories
ne;way between;sub?ects analysis of
%ariance 5';test7
Independent 1ariableis within;sub?ects
Independent 1ariablehas C categories
)orrelated groups
t-test
Independe
has D ocateg
ne;way measures
%aria
5';t
8/9/2019 Choosing Statistical Procedures for Research Problems
17/17
((P*R*8ERI) ES
Independent1ariable is between;
sub?ects
Independent1ariable has C
categories
0ilco@on ran> sumtest " 8ann;0hitney
2 test
or )hi;S$uare test
Independent1ariable has D ormore categories
rus>al;0allis
#;test or
)hi;s$uare test
Independent1ariable is within;
sub?ects
Independent1ariable has C
categories
0ilco@on
signed;ran> test
Indepen1ariable hmore cate
'riedman a%ariance b
Top Related