mcs 7 b (2) variabilebv

7/24/2019 MCS 7 B (2) VARIABILEbv

http://slidepdf.com/reader/full/mcs-7-b-2-variabilebv 1/37

Variable (scientifc research)

Any actor that can:

• take on dierent values is a scientifcvariable

• inuence the outcome oexperimental research

• scientifc experiments measure

quantifable actors: – bservable quantities

– !escribable characteristics or attributes

http://explorable.com/experimental-research.html

http://explorable.com/experimental-research.html



"ypes o variables

• Descriptive variables are those that#hich #ill be reported on$ #ithoutrelatin% them to anythin% inparticular&

• Categorical variables result rom aselection rom cate%ories$ such as'a%ree' and 'disa%ree'&ominal and ordinal variables are

cate%orical&

http://changingminds.org/explanations/research/measurement/types_data.htm

http://changingminds.org/explanations/research/measurement/types_data.htm



"ypes o variables

& !iscrete or *ualitative

Variables&+& ,ontinuous or *uantitative

Variables&



"ypes o variables *ualitatives:(discrete)

• ominal variables di-polichotomic: – dihotomice (poate lua numai o orma din

maxim doua).

– polichotomice (poate lua o orma din maimulte alternative)

Gender: 1. Male2. Female

Marital Status: & /nmarried+& 0arried1& !ivorcee2& 3ido#er




*ualitatives: (discrete)• rdinal variables

• !ummy variables rom quantitativevariables

• 4reerence variables

• 0ultiple response variables




• rdinal variables: are ordered in ameanin%ul sequence$ intervalsbet#een scale points may be uneven&

"he numbers are comparable only interms o relative ma%nitude$ not actualma%nitude:

5 Very lo# or nil+ 5 6o#1 5 0edium2 5 7reat

8 5 Very %reat




• !ummy variables rom quantitativevariables:

A quantitative variable can be

transormed into a cate%orical variable$called a dummy variable by recodin%the values: quantitative variable Agecan be classifed into fve intervals$called dummy variables$#ich are: $+$1$2$8

[Up to 25] 1

9+8$ 2 ; +

92$ 8; 1

98$ <; 2

9Above <; 8




• 4reerence variables: the values are eitherin a decreasin% or increasin% order& =ndicatethe importance o the sources o

inormation by usin% the code 9; or themost important source and 98; or the leastimportant source:

& 6iterature published in the country

+& 6iterature published abroad1& >cientifc abstracts

2& /npublished reports$ material$ etc&

8& !iscussions #ith collea%ues #ithin the research unit

"he interval distance is not the equal&




• 0ultiple response variables are those$#hich can assume more than one value&

"he respondents could score more than

one cate%ory& – A survey questionnaire about the use o

computers in research:Statistical analysis Lab automation

Data base management Modeling and simulation

Scientifc calculations Computer aided design(CAD)

Communication andnetworkin

rap!ics



"ypes o variables ? *uantitatives

*uantitatives – Continuous variables are numeric variables

that can take any value$ such as #ei%ht&

– Discrete variables are numeric variablesthat come rom a limited set o numbers&

"hey may result rom $ ans#erin%

questions such as 'ho# many'$ 'ho# oten'$etc&



"ypes o variables @ *uantitativeshttp:--www.unesco.org-#eb#orld-idams-adv%uide-,hapt1&htm

• "nterval # scale $ariables

•

Continuous %rdinal $ariables• &atio # scale $ariables




Continuous %rdinal $ariables • "he measurements are continuous$ but one

is not certain #hether they are on a linear

scale$ the only trust#orthy inormationbein% the rank order o the observations&Bor example$ i a scale is transormed byan exponential$ lo%arithmic or any other

nonlinear transormation$ it loses itsinterval ? scale property&

• =t is expedient to replace the observationsby their ranks&




"nterval # scale $ariables:• =nterval scale data has order and equal intervals& "hey are measured on a linear scale$ and can take onpositive or ne%ative values& =t is assumed that the

intervals keep the same importance throu%hout thescale& "hey allo# us to rank order the items that aremeasured and also to quantiy andcompare thema%nitudes o dierences bet#een the m&

"he temperature o 2C, is hi%her than 1C,$ and anincrease rom +C, to 2C, is t#ice as much as theincrease rom 1C, to 2C,& ,ounts are interval scalemeasurements$ such as counts o publications orcitations$ years o education$ etc&




&atio # scale $ariables• "hese are continuous positive measurements

on a nonlinear scale (the %ro#th o bacterialpopulation ? equal time intervals multiply the

population by the same ratio)&• 3ith interval data$ one can perorm lo%ical

operations$ add$ and subtract$ but can notmultiply or divide& "o multiply or divide #ehave to use the Delvin scale? has a true Eeropoint& =n social sciences$ the issue o FtrueEeroF rarely arises$ but one should be a#are

o the statistical issues involved&




&atio # scale $ariables• "here are three dierent #ays to

handle the ratio?scaled variables& – >imply as interval scale variables& Go#ever

this procedure should be avoided as it candistort the results&

– As continuous ordinal scale&

–Hy transormin% the data (or example$lo%arithmic transormation) and thentreatin% the results as interval scalevariables&



,/6I7IJIA !A"I6J

• >tudiile evalueaEa reEultate prin observareaunor modifcari-schimbari produse asupra />&

• Aceste schimbari sunt evidentiate de datele

(variabilele) care se modifca la nivelul /> dinpopuatia de baEa&

• JeEultatele se:bserva-0asoara-=nre%istreaEa-AnaliEeaEa

• >e tra% concluEii• =mperativa: defnirea-urmarirea

variabilelor



Variabile

IvalueaEa un enomen-schimbare

• cantitate sau• calitate (atribut)

Trebuie sa e:

bservabile - masurabile? la membrii populatiei de

baEa



Variabile clasifcare

& ,alitative

+& ,antitative

A& !iscreteH& ,ontinue

1& !escriptive -,ate%oriale



"ipuri de variabile

& ,alitative:• A& ominale: permit o clasifcare

calitativa apartinand unor cate%orii ce

nu pot f ierarhiEate$ cuantifcate& 4ot f: – dihotomice (poate lua numai o orma din

maxim doua: masculin - eminin).

– polichotomice (poate lua o orma din mai

multe alternative:casatorit-necasatorit-divortat-vaduv)

• H& rdinale: iau valori care pot f aseEateintr?o insiruire lo%ica a acestor valori



"ipuri de variabile

+& ,antitative

• A& ,ontinue pot f masurate cuacuratete$ iau valori succesivecontinuu ("A$ %reutatea)&

• H& !iscrete ? pot lua doar anumitevalori fxe& =ntre aceste valori nuexista alte valori intermediare&(masc&- em&)



"ipuri de variabile

1& !ependente ? independente

• Variabilele dependente (V!) depind deuna sau mai multe variabile independente

(V=)? statistica poate arata doar corelarea$nu cauEalitatea&

• V= pot f modifcate de cercetator$ (in caEulstudiilor experimentale$ nu la celeobservationale)

• V= si V! ? sunt le%ate intr?un anumitcontext$ nu pot f trecute de la un studiu la

altul&



>cale de masurare avariabilelor

• ominala

• rdinala• !e

intervale

• !erapoarte

Biecare scala

are toatecaracteristicile

scaleianterioare

plus Î ncăceva




ominala

• >et de cate%orii (distincte) care

pot varia doar calitativ (nucantitativ)&

• u diera ierarhic&

•>e pot atasa numere acestorcate%orii&

!". #uloareaoc$ilor




rdinala

=dem nominala K :

• ,ate%oriile sunt ierarhiEate lo%ic&• u exista dierente e%ale intre

treptele ierarhiei cate%oriilor&

!". stadiile cancerului




!e intervale=dem ordinala K :

• !ierentele dintre intervale sunt

e%ale• %u e"ista o valoare de re&erinta

ca L M

• Ix&: scala '# N 2?8 O, 5 N +?1O, $

dar < O, N 1 O, x +




!e rapoarte• =dem intervale K :

• e"ista o valoare de reerinta ca L M

• Ix&: scala '( . L M5 ?+P1 O,N 2?8 OD 5 N +?1 OD $

iar < OD 5 1 OD x +



,hestionare

• !e: ansamblu de intrebari scrise$administrate de operatori cedetermina raspunsuri ce suntinre%istrate in scris&

• =nstrument de baEa in cule%ereadatelor&



,hestionare clasifcare

=n unctie de:

•

0odul de administrare• "ipul de inormatii cerute

• "ipul de intrebari olosite



,hestionare clasifcare

Modul deadministrare

)*irect + indirect,

• 4rin posta?=

• "eleonice

• =nterviu ? !



,hestionare clasifcareTipul de in&ormatii

cerute

• !ate

• ,unostinte

• Atitudini - opinii

• ,omportament

• Acordul inormat

• =normatii desprerude



,hestionare clasifcareTipul de intrebari

&olosite

>tructurate:

• !ihotomice:da-nu• Ale%eri multiple

• =erarhiEate pe scala

devalori:inconort-durere

• Jaspuns numeric:varsta

,u raspuns:

•

inchis?avantaQe: – "intite

– /niorme

– /sor de administrat

– /sor de codifcat – "imp scurt de

analiEa

• !eschis? cer timp

RR Vocabularul



,odifcarea intrebarilor

"rebuie sa fe:

• !"$austiva @cate%oriile sa cuprindatoate raspunsurile$ sa fe codatetoate cate%oriile

• -pecica @ cate%oriile sa se excludareciproc

• 4ot f olosite intrebari in cascada @se raspunde la o intrebare care deapt nu exista in chestionar ca atare&



>tructura chestionarelor

=ntrebarile se clasifca dupa unctie:

& =ntroductive @ de contact

+& !e trecere @ tampon

1& Biltru

2& Hiurcate

8& !e ceS<& !e control

P& !e identifcare



>tructura chestionarelor?7allup

• >tandard pentru chestionare de opinie:

intrebari alternand cu DȊ

• =ntrebari fltru? -Ȋ R.M. cunoaste problemaS

• =ntrebari deschise? atitudunea ata deproblema

• >istem de intrebari ( -Ȋ la problema discutata)

•

=ntrebari D @ motivatia opiniei• =ntrebari - intensitatea opinieiȊ

!/JA"A interviului ≈ T h

.M& 5 raspuns



Validitatea intrebarilor

• >tructurala ? prin repetareormulata dierit

• 4rin repetare la un micnumar de repondentialesi randomiEat

U (intrebari reeritoare lalucruri concrete$ nu deopinie)



Verifcarea validitatii chestionarelor

• 4rin anchete @

conruntarea cu datecunoscute

• ,ontrolul incrucisat alintrebarilor



Borma - aspectulchestionarului

• Istetic-discret

• ecesare: – 4retestarea chestionarului

– >pecialiEarea operatorului

– Jepetarea chestionarului

– Verifcarea codifcarii

mcs 7 b (2) variabilebv

Documents