PRESENTERS:

KASTHURI SUPPIAH

CHITRADEVI ALUMALAI
 DEFINITION & CHARACTERISTICS
 Survey is a method of collecting information by asking questions.

 contain predetermined questions that is given to a sample

 common types: telephone,face to face,postal pencil and paper, web-based and

email forms questionnaire

 usually used in descriptive studies , comparative studies or to explain

relationship between two variables

eg: What is the average length of time teachers

use to prepare lessons? (descriptive studies)

The relationship between school factors and pupils'

well- being. (relationship between variables)
 SURVEY RESEARCH DESIGN:
CROSS-SECTIONAL SURVEYS

 collected from selected individuals at a single point in time

on a population.
 single, stand alone study
 effective to for providing snapshot of the current
behaviours, attitudes, and beliefs in a population.
 provides data quickly
 SURVEY RESEARCH DESIGN:
LONGITUDINAL SURVEYS
 data collected at two or more times.
 useful for studying the dynamics of a topic or issue
over time.
 requires extended commitment by the researcher and
the participants
 difficult in keeping track on sample members over
time and maintaining sample's wilingness to
participate in the study
 trend study, panel study, cohort study
STEPS IN DESIGNING A
SURVEY STUDY
1. Define research objectives
 important to determine the research design
eg: if we want to look at how teachers' view of
effective pedagogy changed over time, we
would have to do a longitudinal study,
surveying the teachers over a number of years.

• also important to decide on the methods to be used

2. Formulate hypotheses
 not necessary to develop and test in survey design

 specific predictions are often made in the form of hypothesis in

survey research: the relationship between self-concept and
achievement will be mediated by emotional intelligence

 experimental study : the relationship between self-concept and

achievement will increase as the child grows older

 Therefore, whether one wants to test specific hypotheses or

conduct a more descriptive study depends on research the
question.
3. Define what information you
need
 decide what information is needed to answer these
research questions.
 what questions to ask
 how long to make the survey
4. Decide what your population is
 the population is the group you want to generalise
your findings to
 important to determine the sample
 if the population size is small, there is no need to
give the questionnaire to the sample.- known as
cencus
 simple
random
sampling

stratified
volunteer
sampling random
sampling

sampling
method

field poll quota

sampling sampling

convenience
sampling
5. Decide how to sample from the
population
 we need to have an unbiased sample of the population
 probability sampling method- simple random
sampling
 sometimes, it is not suitable because too small sample
for a very large population
 so, stratified random sampling would be suitable
 the field poll sampling is used by dialling random
digits for a telephone survey
 volunteer sampling is also used in this research design.
 quota sampling- We may have a population with 10
per cent of Orang Asli pupils, so we will sample untill
we reach our quota, 10 per cent of 1000 Orang Asli
pupils or 100 Orang Asli pupils.
 convenience sampling- the most common and
preferred by the researchers for easy access.
 but has problems when the samples may not be the
representative of the population.
 Therefore, wherever possible, it is advisable to use
probability sampling method to avoid biasness.
6. Design your research instruments
 written questionnaire
 phone questionnairre
 online survey
 face to face
 postal pencil and paper

 important process because once data are collected we

will not be able to rectify any problems with the
instruments
 7. Collect the data
DATA COLLECTION ADVANTAGES DISADVANTAGES
METHODS

Pencil and paper • familiarity to users • low responses rate and

questionnaires time consuming
telephone interviews • direct input into the • intrusive
computer system • little time to think
• easy to adapt questions before answering
based on previous
responses
face to face interviews • allow interviewer to • intrusive
reach sample size • non-cooperation
targets
online and email • directly analysed • technical issues
questionnaires • save data input time • no interviewer to
and costs clarify
8. Analyse the data
Univariate Analysis- variables one at a time

 range
 average
 median
 mode
 variance
 standard deviation
 histograms and normal distribution
Bivariate Analysis- the relationship between two
variables

 cross tabulation or contingency table

 chi square- extension of a cross-tabulation thats gives
more information about the relationship between the
variables.
 DIFFERENT TYPES OF SCALES
 Dichotomous Scales

 A dichotomous scale is a two-point scale which presents options

that are absolutely opposite each other. This type of response
scale does not give the respondent an opportunity to be neutral
on his answer in a question.

 Examples:

 Yes- No
 True - False
 Fair - Unfair
 Agree – Disagree
 Rating Scales
 Three-point, five-point, and seven-point scales are all included in the umbrella
term “rating scale”. A rating scale provides more than two options, in which the
respondent can answer in neutrality over a question being asked.

 Examples:

 1. Three-point Scales
 Good - Fair – Poor
 Agree – Undecided - Disagree
 Extremely- Moderately - Not at all
 Too much - About right - Too little
 2. Five-point Scales (e.g. Likert Scale)
 Strongly Agree – Agree – Undecided / Neutral - Disagree - Strongly
Disagree
 Always – Often – Sometimes – Seldom – Never
 Extremely – Very - Moderately – Slightly - Not at all
 Excellent - Above Average – Average - Below Average - Very Poor

 3. Seven-point Scales
 Exceptional – Excellent – Very Good – Good – Fair – Poor – Very
Poor
 Very satisfied - Moderately satisfied - Slightly satisfied – Neutral -
Slightly dissatisfied - Moderately Dissatisfied- Very dissatisfied
  Semantic Differential Scales
• Marked Semantic Differential Scale
 Please answer based on your opinion regarding the
product:
very slightly neither slightly very

Inexpensive ( ) ( ) ( ) ( ) ( ) expensive

Effective ( ) ( ) ( ) ( ) ( ) ineffective

Useful ( ) ( ) ( ) ( ) ( ) useless

Reliable ( ) ( ) ( ) ( ) ( ) unreliable
• Unmarked Semantic Differential Scale
 The central line serves as the neutral point:

 Inexpensive ____________|_____________ Expensive

 Effective _______________|________________ Ineffective

 Useful _________________|__________________ Useless

 Reliable ________________|________________ Unreliable

 FACTORS TO CONSIDER WHEN
DEVELOPING QUESTIONNAIRES
 keep it brief

 keep your questions clear and simple

 questions related to respondents' data such as age, gender and social

status should be put at the end of the questionnaire

 include a 'don't know' option in rating scale to give respondents who

don't have an answer to make this choice

 avoid double negatives- if you don't disagree

 ask only one question in one item

 take into account cultural differences

 ADVANTAGES
 Relatively easy to administer

 Can be developed in less time

 Cost-effective

 Can be administered remotely via online, mobile devices, mail,

email, kiosk, or telephone.

 reduce or prevent geographical dependence

 Capable of collecting data from a large number of respondents

 Numerous questions can be asked about a subject, giving
extensive flexibility in data analysis.

 With survey software, advanced statistical techniques can be

utilised to analyse survey data to determine validity,
reliability, and statistical significance, including the ability to
analyze multiple variables.

 A broad range of data can be collected (e.g., attitudes,

opinions, beliefs, values, behavior, factual).

 Standardized surveys are relatively free from several types of

errors
 DISADVANTAGES
 Respondents may not feel encouraged to provide accurate,
honest answers

 Respondents may not be fully aware of their reasons for any

given answer because of lack of memory on the subject, or
even boredom.

 Surveys with closed-ended questions may have a lower

validity rate than other question types.
 Data errors due to question non-responses may exist. The
number of respondents who choose to respond to a survey
question may be different from those who chose not to respond,
thus creating bias.

 Survey question answer options could lead to unclear data

because certain answer options may be interpreted differently
by respondents.

 Customized surveys can run the risk of containing certain types

of errors.
EXAMPLES OF SURVEY RESEARCH

 The relationship between school factors and pupils'

well being.
 The study of relationship between health and stress
among teachers.
 Definition and Characteristics of
Correlational research
 According to Creswell, correlational research designs are
used by investigators to describe and measure the degree of
relationship between two or more variables .

 Two variables are measured, without manipulation of

any variables, to determine whether there is a relationship.

 Referred as a form of descriptive research – describes

relationship between variables. Describes the degree to
which two or more quantitative variables are related.
 Sample size of at least 30 participants.

 One group of subjects measured on two variables.

 Variables must be quantifiable and usually represent at

least ordinal scale of measure.

 Correlations does not imply causations. Maybe there is

a third variable causes the change.
 No correlation
-No relationship between
two variables
Positive Ex : The relationship Negative
Correlation between IQ level and Correlation
body weight.
-Variable A -When variable A
increases, so does increases, variable B
variable B will decrease
-High with high, low -High with low , low
with low with high
-Ex: The relationship -Ex: The relationship
between hours of between low
study and economical
examination scores conditions and
Degrees of domestic violence
correlations
 CORRELATION COEFFICIENT
 Once measure two variables ,then compute a statistical
measure called correlation coefficient.

 A correlation coefficient is decimal number between –

1.00 and +1.00.It describes the strength and direction of a
relationship.

 The correlation coefficient of -1 indicates a weak/negative

relationship. A +1 indicates a strong/positive relationship
while a 0 indicates no relationship at all.
A scatterplot is used to represent a
correlation between two variables
 EXPLANATION DESIGN
 Help explain important human behaviours (Explanatory
Design)
 Is done when there is already a hypothesis as to why
something is happening.
 Questions and tests are designed to support that
hypotheses, and prove if it is correct or not.
 Correlate two/more variables
 Unrelated variables(when correlation below .20)
dropped from further consideration
 Collect data at one point in time
 Analyse all participants as a single group
 Obtain at least 2 scores for each individual in the
group (one for each variable)
 Report the use of the correlation statistical test in
the data analysis
 Make interpretations/draw conclusions from the
statistical test results.
 Most researchers most probably trying to gain some
ideas about cause and effect.
 However it does not establish cause and effect
 Ex :Smoking causes lung cancer (correlational
evidence of the relationship between frequency of
smoking and incidence of lung cancer)
 PREDICTION DESIGN
 Predict likely outcomes
 Predict a score on one variable if a score on the other variable is
known
 Predictor Variable -variable that is used to make the prediction
 Criterion(outcome) Variable- variable about which the
prediction is made
 Typically include the word 'prediction' in the title (might also be
in the purpose statement/research questions).
 Measure the predictor variable(s) at one point in time and the
criterion variable at a later point in time.
 Forecast future performance.
 Ex : A person with a high GPA in high school would be likely
have a high GPA in college.
 High school grade( prediction variable)
 College grade ( criterion variable)
 STEPS IN CORRELATIONAL RESEARCH
1 • Problem selection

2
• Sample

3
• Instruments

4
• Design and Procedures

5
• Data collection

6
• Data analysis and Interpretation
PROBLEM SELECTION

 Researcher should have some reason for thinking certain variables

may be related

 Three major types of problems :

 a. Is variable X related to variable Y?

 b. How well does variable P predict variable C?
 c. What are the relationship among a large number of variables and
what predictions can be made?

SAMPLE

 30 or more participants -important to increase the validity of the

research.

 If the variables correlated have low reliabilities and validities, a

bigger sample is necessary.
INSTRUMENTS

 Involve some types of instruments -tests, quizzes,

questionnaire and sometimes observation.
 Must give reliable scores

DESIGN AND PROCEDURE

 Complex part of a correlational study is determining how
to effectively measure each variable.

 Scores for two variables are obtained for each member of a

selected sample and paired scores are correlated.

 validity and reliability

DATA COLLECTION
 Data on both variables will usually be collected in a short
time.
 Instruments used are administered in a single session or
two sessions

DATA ANALYSIS AND INTERPRETATION

 When variables correlated, a correlation coefficient is

produced.
 Threats to Internal Validity
in Correlational Research
The following must be controlled to reduce threats to internal validity :

 Subject characteristics - Individuals or groups have two or more

characteristics; might be a cause of variation in the other two variables.

 Location - Location is different for different subject. One location may

be more comfortable compared to others.

 Instrument decay - care must be taken to ensure the observers don’t

become tired, bored or inattentive.

 Testing - Experience of responding to the first instrument may

influence subject responses to the second instrument.
Examples of correlation research
 A Correlational Study of the Relationships Among
Student Performance and Teacher Perceptions.

 Features and Correlations between Teaching Behaviors

and Learning Activities.

 The Relationship between Intelligence and self

esteem.

 The Relationship between Hours of Study and

Examination Scores.
SURVEY RESEARCH & CORRELATIONAL RESEARCH

 Test the same hypothesis

 Examine relationship of responses to one question in a

survey to another or of score base on one set of survey
questions to a score based on another set.

 Questionnaire of the variables are prepared, then the

relationship determined by calculating correlation
coefficients.
 EXERCISE
1. A researcher records a correlation of r=+ 0. 2
a. Identify the direction of the correlation.
b. Identify the strength of this correlation.

2.A researcher finds that students who attend fewer

classes get poorer grades.
a. Is this a positive or negative correlation?
b. Why might we find a relationship between attendance
and grades?
3. What is a cross-sectional design?

a) A comparison of two or more variables longitudinally

b) A design that is devised when the researcher is in a bad
mood.
c) The collection of data from more than one case at one
moment in time.
d) Research into one particular section of society, e.g. the
middle classes.

4. What are the differences between univariate analysis

and bivariate analysis.
 VIDEOS
 Cross sectional and longitudinal study
 https://www.youtube.com/watch?v=TA03H5VdURM

 Research strategy: Survey

 https://www.youtube.com/watch?v=GDTydIJfcng

 Correlational research
 https://www.youtube.com/watch?v=UJvQ6GBCw2c
REFERENCES
 Fraenkel, J., R., & Wallen, N., E., (1990). How to
design and evaluate research in education. New
York.
 E.Slavin, R. (2007). Educational Research. Pearson
Education.
 Gay, L., E.Mills, G., & Airasian, P. (1976). Educational
Research. Pearson International Edition.
 Jhonson, B., & Christensen, L. (n.d.). Educational Research:
Quantitative, Qualitative, and Mixed Approaches.
 Muijs, D. (2004). Doing Quantitative Research in
Education:with SPSS. Thousand Oaks, CA: SAGE
Publications.
 https://www.researchconnections.org/childcare/data
methods/survey.jsp

 https://explorable.com/survey-response-scales

 https://saylordotorg.github.io/text_principles-of-
sociological-inquiry-qualitative-and-quantitative-
methods/s11-survey-research-a-quantitative.html
 The End

Thank You 