wiihout ony preiest scores, it is impossible to determine if ony chonge within the group itself
hos tcken
. ONE GROUP PRETEST POSTTEST STUDY
lnclusion of o pretesi to determine boseline scores
we con now of leost stote whether c chonqe in ihe outcome or dependent vorioble is token ploce
whcl we connot soy is if this chonge would hove occuned even without the opplicotion of the ireotment.
. THE STATIC GROUP COMPARISON STUDY
Attempts to moke up for the lock of o control group but folls short in relotion to showing if o chonge
hos occurred.
Does not include pre tesling ond therefore ony difference between the two groups priir
to ihe study ore unknown. 1l,r
PTcTURE SA RtcHT)
True Experimenlol Design
They employ both o control group ond o meons 1o meosure the
chonge thot occurs in born groups
only reseorch method thot ccn odequctely meosure the couse ond effect relotionship
. Positest Equivolent Groups Studv
Rondomizotion ond the comporison of both o conlrol ond on expenmentol group PrSs^t [4ui!d]nt Grorpi
ore utilized in ihis E)t.
iype of study. {}:D,
while ihis is opproochlng the besi method, it folls short in its lock of o pretest
meosure. Fr$esl Poin * Fsivatsni GmlF
n s.( c1
Rondomizolion does well to mix subjects but it does not completely ec,(cl
ossure us thoi this mix is truly
creoling on equivolency belween the two groups.
. Pretest-Posttest Equivolent Groups Study
It is the most effective method but olso the most difficurt to perform
Quosi-experimeni
"Quosi" meons /lkeness or resembltng, so quosi-experiments
shore chorocteristics of true experiments which seer
interventions or ireotments.
The key difference in this empiricol opprooch is the lock of rondom
ossignment.
Another unique element ofien involved in this experimentotion metbod is use
intorrr rnlar]
of time series onolysis: interrupted ond non-
Design of Quosi-experimenl
l. ldentify the voriobles.
2. The quosi-independent vcrioble will be the x_voricble.
3 Grouping meons two or more groups such os o treotmeni group ond c control group.
4. The predicted outcome is the dependent vorioble which is the y-vorioble.
5' In o time series onolysis, the dependent vorioble is observed over time
for cny chonges
thot moy toke ptoce.
6' Once the voriobles hove been identified ond defined, o procedure should then be implemented
ond group
differences should be exomined.
Advontoges
o They ore iypicclly eosier to set up thon true experimentol designs, it tokes r ruch
less effort to study ond compore
subiects or groups of subjects thol ore olreody noturolly orgonized thon lo hove io
(t tnta-ic
conduct rondom ossiqnment of
o Utiliiing quosi-experimentol designs minimizes threots to externol volidity
o Since quosi-experiments ore noturol experiments, findings in one moy be opplied to
other subjecls ond setiings,
ollowing for some generclizotions to be mode oboui pooulotion.
o lt is efficient in longitudinol reseorch thot involves longer time periods which ccn be
followed up in different
environments.
Disodvonfoges
o The control ollowed ihrough ihe monipulotion of the quosi-independeni voricble con leod io unnoturol
circumstonces, olthough the dongers of ortificiolily ore considerobly less relotive to lrue experimenls
o lt moy ollow studies to be more fecsible, but this olso poses mony chollenges for the invesiigotor in terms of internol
volidity.
o Becouse rondomizotion is obsent, some knowledge obout ihe doto con be opproximoted, but conclusions of
cousol relotionships ore difficult to determine
Focloriol Design
o Foctoriol Designs ore designs used extensively in educoiionol reseorch, becouse essenliclly Fffi.
they ore the designs when two or more independeni voriobles ore included in the design. t',,ffi
-,*x,I.rT.r:ip.
o The term foctorrol is used to indicoie thoi oll possible combinotions of the foctors ore
constoereo.
o The bosic construciion of o fociorioi design is thot oll levels of eoch independeni vorioble ore
token in combinotion with the levels of the other independent voriobles.
o The design requires o minimum of two independent voriobles, with ot leosl two levels of eoch vcricble.
O li involves two or more independent vcriobles, colled focfors, in o single design.
o The number of different groups involved in o foctoriol design increoses very ropidly with the increose of the number
of independent voriobles ond number of levels.
Advonioges
o Foctoriol design provides the economy of o single design rother thon seporcte designs for eoch of the independeni
voriobles, ond it ollows the reseorcher to invesiigote lhe interoctions between the voriobles.
O The fociorlol design, os well os simplifying the process ond moking reseorch cheoper, ollows mcny levels of onolysis.
Disodvonloge
O The moin disodvcntoge is ihe difficulty of experimenting with more thon two foctors, or monv levels. lt hos to be
plcnned meticulously, os on error in one of the levels, or will jeopordize o greot omount of work
STATISTICAT TOOLS FOR DATA ANATYSIS
Stotistics- Mothemoticol science perloining to the collection, onolysis, interpretotion or explonotion, ond presentofion of
Stotisticol methods con be used to summorize or descrlbe o collection of doto; ihis is cclied Descripfive Stofisllcs
thct ore then used to drow inferences crhor ri tha nror-asc or populotion being studied; this is colled lnferenfiot
UHfi ,
t/- /
'::
"//
Non-parametric - refers to (a function on a sample) whose interpretation does not depend on the population fitting any
oarameterized distributions.
ff you get it wrong you risk using on incorrect statistical procedure or you may use o less powerful procedure.
Parametric Non-parametric
variables, without making any other assumptions about the particular nature of the relationship between the variables ./l
COLLECTION of DATA (pp 97-104) ...Psychologicaltest...Anecdotal
INSTRUMENTS FOR DATA COIIECTION Records...Autobiographies...Checklist, Score cards,
1. QUESTIONNAIRE (pp 105-1271 Rating scales, etc
2. fNTERVfEW (pp 729-1,37) CHARACHTERIS'T,G OF SATISFACTORY CASE STDY-Continuity
3. OBSERVATION (pp 139-154) 9 Completeness of data Validity of data Confidential
4. TESTS (pp 155-160) recording Scientific synthesis
SAMPIING (pp 163-177) TYPES
NATUREOFDESCRIPTIVE V ExploratoryCaseStudy, ExplanatoryCaseStudy,
V lt describes and interprets "what is". Descriptive Case Study
I lt is concerned with conditions of relationships that ADVANTAGES
exist,' practices that prevail; beliefs, processes that V Case Study provides a great amount of description
are going on; effects that are being felt, trends that and detail.
are developing. V Case Study present opportunities that researchers
PURPOSE could not otherwise have.
V Determines and describes the way things are DISADVANTAGES
V Compares how sub-groups view issues and topics V diffi€ult to generalize
BASIC STEPS V there is considerable room in case studies for
V Statement ofthe problem "researcher bias" to creep in
V ldentification of information needed to solve the SURVEY METHOD
problem V a study used to measure existing phenomenon
V Selection or development of instruments for without inquiring why it exists
gathering the information V a fact-finding study with adequate and accurate
V ldentification of target population and determination interpretation
of sampling procedure V is used to collect demographic data about people
V Deslgn of procedure for information colfection PURPOSE
g Collection of information V To describe the characteristics, behaviors or opinions
V Analysis of information of a particular population
V Generalizations and/or predictions NATURE
DISADVANTAGES . Information is collected from a group of people.
1. What caused the prevailing conditions is not emphasized. . The information is collected through asking
2. Low response rates are common questions.
3. There are difficulties in interpreting the findings without the TYPES
data representing non-respondents'views 1. Cross-sectional survey
4.Unc|ear/ambiguousitems.Researcherneedstodeve|op>
recording forms that collect the data objectively and reliably 2. Longitudinal survey
5.Becausetherearenovariab|esmanipu|ated,thereisno>
way to statistically analyze the results. Many scientists regard time.
this type of study as very unreliable and 'unscientific'. STEPS
6.Theresultsofobservationa|studiesarenotrepeatab|e>
which changes over time, and so there can be no replication 1. Population lssues
of the experiment and reviewing of the results. 2. Sampling lssues
CASE STUDY 3. Question lssues
V is an extensive and intensive investigation of a unit 4. Content lssues
represented, whether the unit is an individual, a 5. Bias lssues
family, a social group, an institution, or a community 5. Administrative lssues
V PURPOSE- to identify causal factors to some 7. Planning and survey design
abnormality or deficiency and to find and recommend 8. Data collection
a solution, a treatment, or developmental procedures 9. Data analysis
METHOD- has been employed to describe a plan of organizing 10. Reporting
and presenting instructional materials in law, medicine, social ADVANTAGES
work, and even in education, psychology, and sociology, where 1. lt is not labor intensive.
as a rule, the case materials used are the product of case study 2. Methodology is not complex.
investigation 3. Small sample size can be used.
NEEDED INFORMATION- ldentifying data... ldentification and 4. lt reveals what is typical, average or normal.
statement of the problem...Health and development 5. lts results may be used for prediction.
history...Family history...Educational history...Social 6. lt makes possible for the formulation of generalizations.
History....Economic history...Psychological history 7. lt reveals problems.
STEPS lN ORGANIZTNG AND CONDUCTING- Problem A case 8. The instruments for gathering the data are easy to
study should address a problem determine.
g lmplementation-Theirnplementationdiscusseshow DISADVANTAGES
the problem was addressed 1. Retrieval problem of the questionnaire fielded.
V Result- Result discusses the outcomes of how the 2. Perceptions are not always reliable.
addressed
problem was 3. Questionnaire may not be able to address the
V INSTRUMENTS- problem due to inadequacy of information.
Interview...Observation...Questionnaire
Survey Case study
a. Usually largepopulation a. Usually one person, small community
b. Limited variables in the life of the group b. All the variables in the life are included
surveyeo c. Finding the causes is always a part
c. Cause-effect relationship are not given d. Representativeness is not important
emphasis e. From birth or origin or even the future
d. Representativenessisimportant
e. Present conditions
r Satisfaction measurements
r Product features desired
r Medical errors
r Patient outcomes
r Viewer/readership nterests
i