RESEARCH METHODOLOGY AND STATISTICS - 1

DEFINITIONS

● “Research is a systematic inquiry to describe, explain, predict, and control the

observed phenomenon. It involves inductive and deductive methods.”- Earl Robert
Babbie ● Kerlinger (1986) - “Scientific research is systematic, controlled, empirical, and
critical investigation of hypothetical propositions about the presumed relations among
natural phenomena”
● John W. Creswell, states that "research is a process of steps used to collect and
analyze information to increase our understanding of a topic or issue"
● According to Kothari (2006) Research is a pursuit of trust with the help of study,
observation, comparison and experiment, the search for knowledge through objective
andsystematic method of finding solutions to a problem.
● Scientific- Systematic- Verifiable. Describe, Understand, Predict and
Explain psychological phenomenon.

OBJECTIVES

1. To gain familiarity with new insights into a phenomenon (i.e., formulative research studies)

2. to accurately portray characteristics of individual, group, or a situation (i.e.,

descriptive research studies)
3. To analyse the frequency with which something occurs (i.e., diagnostic research
studies) 4. To examine the hypothesis of a causal relationship between two variables (i.e.,
hypothesis testing research studies)
5. To develop new theories (ex- Grounded theory approach)

RESEARCH SIGNIFICANCE

● Development of logical habits of thinking and organisation

● Formulation of economic policy
● In solving various problems-medical, industrial, educational
● Seeking explanations to behaviour & decision making
● For the sake of knowledge

RESEARCH APPROACHES
RESEARCH DESIGN

i) Title of the Study

ii) Stating research Problem

iii) Review of Literature

iv) Area and Scope of Study

v) Objectives of the Study

vi) Formulation of Hypothesis

vii) Definition of Concept and Terminology

viii) Determining Tools of Data Collection

ix) Sampling Design

x) Data Analysis

xi) Interpretation of Results - drawing inferences

xii) Limitation: gap in the data, sample

RESEARCH PROBLEM

VARIABLES
● Independent - Dependent
● Mediating - Moderating
● Extraneous (Potential to affect)
● Confound (Does affect)
● Control- Composite

• A variable is any kind of attribute or characteristic that you are trying to measure, manipulate
and control in statistics and research.
Independent variables Dependent variables
Definition A variable that stands alone and A variable that relies on and can
isn't changed by the other be changed by other factors that
variables or factored that are are measured
measured

Example Parenting style Academic achievement

Intervening (mediate) variables Moderating variables

Definition A theoretical variable used to explain Changes the relationship between

a cause or connection between other dependent and independent
study variables variables by strengthening or
weakening the intervening
variable's effect

Example Access to health care: If wealth is the Age: If a study looks at the
independent variable, and a long life relationship between economic
span is a dependent variable, a status (independent variable) and
researcher might hypothesize that how frequently people get physical
access to quality health care is the exams from a doctor (dependent
intervening variable that links wealth variable), age is a moderating
and life span. variable. That relationship might
be weaker in younger individuals
and stronger in older individuals.

Extraneous variables Confounding variables

Definition Factors that affect the dependent Extra variables that the researcher
variable but that the researcher did not did not account for that can
originally consider when designing the disguise another variable's effects
experiment and show false correlations

Example Parental support, prior knowledge of a In a study of whether a particular

foreign language or socioeconomic genre of movie affects how much
status are extraneous variables that candy kids eat, with experiments
could influence a study assessing areheld at 9 a.m., noon and 3 p.m.
whether private tutoring or online Timecould be a confounding
courses are more effective at variable, asthe group in the noon
improving students' Spanish test study might behungrier and
scores. therefore eat more candybecause
lunchtime is typically at noon.

Control variables Composite variables

Definition Characteristics that are constant and Two or more variables combined
do not change during a study to make a more complex variable
 Example In an experiment about plant Overall health is an example of a
development, control variables might composite variable if a researcher
include the amounts of fertilizer and uses other variables, such as
water each plant gets. These amounts genetics, medical care, education,
are always the same so that they do quality of environment and chosen
not affect the plants' growth. behaviors, to determine overall
health in an
experiment.

LEVELS OF MEASUREMENT

Continuous and Discrete

Levels of measurement :

• Nominal • Ordinal • Interval • Ratio

RESEARCH HYPOTHESIS
● Can be
experimental
(IV
manipulated)
and

non-experimental (not manipulated) ● A prediction or a hypothesized

relationship- tested by scientific methods ● Null Hypothesis and Alternative
Hypothesis-A and B are compared

- Ho assumes that both are equal

- Ha assumes that both are not equal

● If Hois accepted, then it implies that Ha is being rejected & vice versa
● Ha is usually the one which a researcher wishes to prove
TYPES OF HYPOTHESIS

● Descriptive Hypotheses - a statement about the existence, size, form, or distribution of

a variable
● Relational Hypotheses- a statement about the relationship between two
variables correlational or explanatory (causal).
● Groups Difference hypotheses- a statement about observed differences/
similarity between two groups
● Directional and Non-directional Hypotheses-whether or not direction is specified in
the statement- 1 & 2 tailed test

CHARACTERISTICS OF HYPOTHESIS

1. Precise and clear

2. Specific and limited in scope
3. Stated in the simplest language
4. Capable of being put to test
5. Derived from the most known facts
6. Able to testing within a stipulated or reasonable period of time

SAMPLING

● Technique of selecting items for the sample from the population or

universe ● Take into consideration

- Type of Population or Universe (include all items): Finite – Infinite

- Source List: sampling frame- all units
- Sampling Unit: selected from sampling frame
- Size of Sample
- Parameters of Interest-proportion, subgroup,
- Budgetary Constraint
- Sampling Procedure- with smallest error

SAMPLING ERROR AND SYSTEMATIC BIAS

 ● Random sampling error- because the sample is not the whole population- precision of
the sampling plan (measuring)
● Systematic (non-sampling error)bias due to
- Non-response errors
- Response errors (data collection)
- Tabulation errors (data processing)

SAMPLE DESIGNS

● Can be categorised based on representation:

- Non probability sampling (non-random)
- Probability sampling (random)

● Non probability sampling

1. Convenience / availability sampling
2. Judgemental /Purposive sampling
3. Quota Sampling
4. Snowball sampling
 METHODS OF RANDOM SAMPLING

● Simple random sampling- Lottery

● Systematic sampling- fixed intervals
● Stratified Sampling- homogenous
● Cluster Sampling-Area Sampling
● Multi-stage Sampling

ETHICS IN CONDUCTING AND REPORTING RESEARCH

Ethical dilemmas in the three major stages in quantitative research are :

1. Initial stage of subject/participant recruitment
- Obscuring research goals
- Informed consent becomes just a formality
- Voluntary participation

2. Data collection

- Not distort participants response

- Privacy
- Balance benefit & risks

3. Data analysis and production of the report

- Anonymity

● Principles – autonomy, beneficence & justice

Research ethics provides guidelines for the responsible conduct of research. It educates
and monitors scientists conducting research. The following is a general summary of some
ethical principles:

● Honesty: Honestly report data, results, methods and procedures, and publication status.
Do not fabricate, falsify, or misrepresent data.
 ● Objectivity: Strive to avoid bias in experimental design, data analysis, data interpretation,
peer review, personnel decisions, grant writing, expert testimony, and other aspects of
research.
● Integrity: Keep your promises and agreements; act with sincerity; strive for consistency
of thought and action.
● Carefulness: Avoid careless errors and negligence; carefully and critically examine your
own work and the work of your peers. Keep good records of research activities. ● Openness:
Share data, results, ideas, tools, resources. Be open to criticism and new ideas. ● Respect for
Intellectual Property: Honor patents, copyrights, and other forms of intellectual property. Do
not use unpublished data, methods, or results without permission. Give credit where credit is
due. Never plagiarize.
● Confidentiality: Protect confidential communications, such as papers or grants submitted
for publication, personnel records, trade or military secrets, and patient records. ●
Responsible Publication: Publish in order to advance research and scholarship, not to
advance just your own career. Avoid wasteful and duplicative publication. ● Responsible
Mentoring: Help to educate, mentor, and advise students. Promote their welfare and allow
them to make their own decisions.
● Respect for Colleagues: Respect your colleagues and treat them fairly. ● Social
Responsibility: Strive to promote social good and prevent or mitigate social harms through
research, public education, and advocacy.
● Legality: Know and obey relevant laws and institutional and governmental policies. ●
Animal Care: Show proper respect and care for animals when using them in research. Do not
conduct unnecessary or poorly designed animal experiments.
● Human Subjects Protection: When conducting research on human subjects, minimize
harms and risks and maximize benefits; respect human dignity, privacy, and autonomy. ●
Non-Discrimination: Avoid discrimination against colleagues or students on the basis of
sex, race, ethnicity, or other factors that are not related to their scientific competence and
integrity.
● Competence: Maintain and improve your own professional competence and expertise
through lifelong education and learning; take steps to promote competence in science as a
whole.
● Basic Research Misconduct known as the three “cardinal sins” of research conduct,
falsification, fabrication, and plagiarism (FFP) are the primary concerns in avoiding
research misconduct.
● Falsification: Falsification is the changing or omission of research results (data) to
support claims, hypotheses, other data, etc. Falsification can include the manipulation of
research instrumentation, materials, or processes.
● Fabrication is the construction and/or addition of data, observations, or characterizations
that never occurred in the gathering of data or running of experiments. Fabrication is
"filling out" the rest of experiment runs
● Plagiarism is “the appropriation of another person's ideas, processes, results, or words
without giving appropriate credit.
PARADIGMS OF RESEARCH
 ● Descriptive versus Analytical
● Applied (Action) versus Fundamental
● Quantitative versus Qualitative
● Conceptual versus Empirical
● Cross sectional versus longitudinal (cohort studies)
● Laboratory (simulation) research versus field setting research

i) Descriptive versus Analytical

Descriptive research classifies, describes, compares, and measures data. Meanwhile, analytical
research focuses on cause and effect. For example, you may talk about the mean or average trade
deficit in descriptive research. Meanwhile, analytical research measures something different.
Instead, you’d look at why andhow the trade deficit has changed.

ii) Applied (Action) versus Fundamental

Applied Research focuses on practical problem-solving and finding real-world solutions.

Fundamental Research seeks to expand knowledge and understanding, exploring theoretical
concepts and underlying principles.

iii) Conceptual versus Empirical

Conceptual research is about creating an idea after looking at existing data or adding on a theory
after going through available literature. Empirical research involves research based on
observation, experiments, and verifiable evidence.

iv) Cross sectional versus longitudinal

While longitudinal studies repeatedly observe the same participants over a period of time,
cross-sectional studies examine different samples (or a “cross-section”) of the population at one
point in time

v) Laboratory research versus field setting research

Field research entails observation of people in situ or in field setting whereas lab research
observes people in a setting where the variables are controlled at the discretion of the researcher.

RESEARCH METHODOLOGY AND STATISTICS - 2

OBSERVATION

Intentional and planned systematic viewing- usually of nonverbal behaviour

 - Procedure of observation:
● Using mechanical devices- scopes, camera, videotapes, sensors
● Checklist & Schedules
● Time Sampling-frequency of behaviour during the time intervals that are
systematically spaced
● Event Sampling- records all instances of a particular event & ignore other in a
time period
● Specimen Sampling- record description of the subject’s entire scheme of behaviour for
a period

- Types:
● Non-participant: Naturalistic Observing of events- Artificial probes or manipulation
might destroy the character of the event being studied- To avoid intrusion ●
Participant Observation: participant can interface with the subjects-Undisguised &
Disguised Participant Observation (observer disguise to the extent of being accepted)
● Hawthorne effect- observation bias

SURVEY- QUESTIONNAIRES

● A notable feature- usually only one opportunity to collect data from each
informant Therefore, questions needs to be clear, comprehensive and effective
● Close-ended Questions - MCQ’s- straight-forward, quick to answer- responses are
easily turned into quantitative data. Types of Closed Questions are:

- Checklist: Presents a range of possible answers to tick as many as may apply

- Ranking Questions: placing items in rank-order according preferences or
frequency
- Graded Response Questions: degree of magnitude

● Open-ended questions: Inviting a more detailed, and personal response

INTERVIEWS
 ● They involve social interaction- researchers need training- different types of
questions generate different types of data
● Interview schedule- a set of prepared questions designed to be asked exactly as worded-
astandardised format
● Recording of interviews- data written up as a transcript- written account of
interview ● Special care when interviewing vulnerable groups
● Language appropriate to the vocabulary of the group- social background/
respondents' age / educational level / social class /ethnicity etc.
● Types of Interviews: Structured & Unstructured

STRUCTURED INTERVIEW

● Formal interview- questions are set / standardized order and the interviewer will
not deviate from the interview schedule- based on worded, closed-ended- questions.
● Strengths
- easy to replicate as a fixed set of closed questions are used- high reliability. - fairly
quick to conduct- many interviews can take place within a short amount of time- a large
sample can be obtained- generalized to a large population.
● Limitations

- Not flexible- no impromptu questions

- Answers lack detail- intentions of a person cannot be elicited (why?)

UNSTRUCTURED INTERVIEW

● Discovery interviews- ‘guided conservation’ – also called informal interviews

● Interview schedule not used- if used -flexible- will contain open ended
questions ● Strengths
- more flexible- questions adapted and changed depending on the respondents’ answers. -
Generate qualitative data- talk in depth, use own words- develop a real sense of a person -
Increased validity- relevant questions- ask for clarification & steer the direction of
interview
● Limitations
- Time consuming to conduct and analyse the qualitative data
- Employing and training interviewers is expensive- certain skills may be needed -include
the ability to establish rapport & knowing when to probe

GROUP INTERVIEW
 ● Many respondents are interviewed together – ‘focus group’
● A researcher must be highly skilled- ability to establish rapport and knowing when to
probe- make sure the group interact with each other and do not drift off topic ●
Strengths- Same as unstructured interview
● Limitations
- Keeping confidentiality and respecting participant privacy difficult- the researcher cannot
guarantee that other participants will keep information private
- less reliable- use open questions and answers may deviate- difficult to repeat the question -
May lack validity- social desirability bias- conform to peer pressure and give false answers
THE INTERVIEWER EFFECT

● Interview is a social interaction- appearance or behaviour of the interviewer may

influence the answers of the respondent- can bias the results– make them invalid. ●
Gender, ethnicity, body language, age, and social status of the interviewer ● For
example, if a female researcher was investigating sexism amongst males- male
participants may lie & pretend they are not sexist to impress
● Recall effect-certain aspects are recalled more by the respondent

FIELD RESEARCH

● Social scientists observe, interact and understand people while they are in a
natural environment- how they react to situations around them
● Expensive and timely, however amount and diversity of the data invaluable-
original ● Techniques- Direct observation, document analysis, informal interviews,
surveys ● Data analysis is based mostly on correlation; No manipulation
● Data collected is specific only to the one purpose or setting- not generalizable ●
Method- clearly stating the problem > defining area of study > a hypothesis, or a theory
isset forth> decision on how data to be classified and scaled> Observations are classified
-what to look for and what to disregard> Observations are also scaled- a way to rank the
importance or significance > data analysed and processed > resolve the problem or
accept/reject the hypothesis

CROSS-CULTURAL STUDIES

● Comparative research focuses on systematically comparing culture to another culture

incidence, distributions, and causes of phenomenon
● Cultures change over time, so focus on particular time frames
 ● Uniqueness and similarity between cultures are always present,
Simultaneously.
● To identify what may be universal and variable about human cultures, as well as to
discover reasons why the variation exists
● Cross-cultural psychologists generally collect primary data and is limited often only to
two-cultures (the investigator’s own and one other)
PHENOMENOLOGY

● Study of lived experience

● What is the essence of the phenomenon as it is experienced by the
participant? ● Diverse subjective perspectives of the participants
● Main data source is in-depth conversations with participants who experienced
the phenomenon being studied
● Researcher eliminate any prior assumptions and personal biases
(BRACKETING)- empathize with the participant’s situation
● Eg: experience of living with a person with Alzheimer's disease
● Two types: descriptive & interpretive phenomenology

DESCRIPTIVE PHENOMENOLOGY

● Put forth by Edmund Husserl

● Phenomenology is to let a thing show itself to us
● Describes human experience- not looking deep into the meaning
● May involve maintaining reflective journal for Bracketing
● Phenomenological reduction- pure ego- bracket non-essence- epoche (suspension
of judgements)
● Insists on the careful portrayal of ordinary conscious experience of everyday life-
a depiction of “things” as people experience them
● Experience of hearing, seeing, believing, feeling, remembering, deciding

INTERPRETIVE PHENOMENOLOGY

● Based on philosophy of Heidegger -Heideggerian hermeneutics

● Emphasis on interpreting and understanding experience, not just describing it
● Bracketing does not occur- to understand the deeper meaning (deep structures
symbolic meaning) we also bring our own experiences
● Main source- in depth interviews, also use diaries, photo journals other arts
● Sample size 10 or fewer
● Probes: “can you describe a typical day” or “can you describe that particular
incident in more detail?”
● Units of significance or Meaning units experiences a transition in meaning. Meaning
units separated by slashes in the text
 ● Giorgi's method of analysis: Interviews are recorded and transcribed > reads the whole
transcripts to get a sense of the whole > read again to establish “units of significance” or

● Interpreting recorded data about a social phen

that phenomenon (Glaser and Strauss, 1967)
● Decision making and its processes- Basic socia
of how people resolve the concern
● Emergent Coding techniques – classifying and ca
set of codes (concepts), categories (constructs),
● 20 to 30 people are used since theory making is i
● Constant comparisons with the data obtained to
categories (commonalities & variations)
● Theories “grounded in” (or based on) participan
and social interactions
● To ensure that the theory is based solely on obser
researcherssuspend any pre-existing theoretical
analysis, and let the data dictate the formulation
● Strauss and Corbin (1998) describe three coding
Opencoding, Axial coding, Selective coding

CODING-ANALYSIS
“meaning units” > Transform the
meaning units-to transferable
language and then to
psychological meanings ● OPEN CODING- a tool for initial analysis of the
(intentionality)> using the reading text several times & dividing into segm
transformed meaning units as the data that summarize an idea done at this stage
basis for explaining the ● AXIAL CODING- next step is organizing and co
psychological structure of the (Axes)- The tentative labels are grouped based
phenomenon> General structure. (20-30) to get core themes (3 -5)- requires induc
● Colaizzi’s method (returning to for causal relationships
validate)- Van kaam method ● SELECTIVE CODING- finding the core categor
(intersubjective agreement using categories- Create a scheme or map that connec
expert judges) them to core category- story lining- overall expl

GROUNDED THEORY

MEMO & CONCEPT MAPPING

 ● Iterative process- moving forward and back to data and analysis
● Memos are theorized write-ups of ideas about concepts and their theoretically coded
relationships as they evolve during ground theory analysis. It is a tool to keep track
of and refine ideas that develop during the analysis. They help to discover patterns
and relationships between categories
● Concept mapping is a graphical representation of concepts and relationships
between those concepts (e.g., using boxes and arrows). These are later readjusted to
best fit the observed data.
● After a grounded theory is generated, one has to check how the findings confirm
or challenge the literature- Discriminant sampling
● Researchers recruit new participants- same questions-if they experience similar
processes – theory is validated- If the theory contradicts with observed evidence, the
coding processmay be repeated to reconcile such contradictions or unexplained
variations

Codes Categories Theme

Medication side effects Health related factors Struggling with a life-long

Absence of disease symptoms disease

Family problems Socio economic factors

Poor financial status

Lack of financial support

Fear of long-term effects Psychological factors

of medication

A need to protect the child Motivation to resume

treatment
ETHNOGRAPHY

● Derived largely from the field of anthropology- studying a phenomenon within the
context of its culture
● Researcher must be deeply immersed in the social culture over an extended period of
time (usually 8 months to 2 years) and should engage, observe, and record the daily life of
the studied culture and its social participants within their natural setting ● The primary
mode of data collection is participant observation
 ● Advantages
- sensitiveness to the context- natural setting
- the rich and detailed understanding it generates
- Minimal respondent bias
● Disadvantages:
- an extremely time and resource-intensive approach
- findings are specific to a given
culture and less generalizable to
other cultures ● The classic example:
Jane Goodall’s study of primate
behaviours- lived with chimpanzees
at Gombe National Park in Tanzania,
how chimpanzees seek food and
shelter, socialize communication
patterns, their mating behaviors
NARRATIVES

● Focus on the lives of individuals as told through t

it is narrated
● In most cases one will be creating an aggregate o
eachbearing on the others- explore the significan
● No one standard set of procedures- weaknesses -
● Data analysis involves a ● Data sources: field notes; journal records; intervi
“sense-making” approach- observations; storytelling; letter writing; autobio
de-contextualised and school and class plans, newsletters, and other te
recontextualised- pictures; audio and video recordings
trans-situational ● Dedicated research software help the researcher t
● Researcher must take extensive the data
field notes, and narrate her ● Main characteristics- narrative research is a colla
experience in descriptive detail ● The research subject is regarded as a collaborator
so that readers may experience informant guided by the agenda of the researche
the same culture as the ● Stories of experience are shaped through discussi
researcher- emic (within) – etic ● Data is arranged in chronological order (restory)
(outside) mixand narrate- past future and present
● Researcher has two roles: rely on ● Analysis specifies turning points, specific tension
her unique knowledge and and categories that emerge
engagement to generate insights ● Biographical, auto-ethnography, life history or or
(theory), and convince the
scientific community of the
trans-situational nature of the
studied phenomenon

THEMATIC ANALYSIS
 ● TA was first developed by Gerald Holton, a physicist and historian of science, in the
1970s- currently the most common form of analysis method used in Qualitative research
● Pinpointing, examining and recording patterns or themes within the data ● Themes are
patterns across data sets that are important to the description of a phenomenon and are
associated to a specific research question
● The theme is the meaningful ‘essence’ of the data
● It is done in 6 phases to establish meaningful patterns:
1. Familiarization with the data (reading & re-reading & notetaking)
2. Generating initial codes (meaningful parts, cyclical process- going back &
forth) 3. Searching for themes among the codes (meaning identifying
phrases/sentences) 4. Reviewing themes (search data that supports/refutes
proposed theory- collapse or condense themes,)
5. Defining and naming themes (engaging & descriptive theme
names) 6. Producing the final report (important final themes
discussed)

CASE STUDY

● Case studies are in-depth investigations of a single person, group, event or community to
describe a phenomenon
● Data source-observations & interviews, diaries, personal journals, medical reports,
documents- case history in clinics
● The information is mainly biographical- past & current life
● Can be analysed by any method- grounded theory/ phenomenology
● Provides rich qualitative information- for further research
● Time consuming, non-generalizable, researcher bias, non-replicable
 RESEARCH METHODOLOGY AND STATISTICS - 3

LEVELS OF EVIDENCE

NON- EXPERIMENTAL DESIGNS

● IV not manipulated
● Subjects not randomised
● No control group
● Provides Level 4 evidence
● Also called as Observational studies
● Descriptive research- designed to provide a snapshot of the current state of
affairs ● Correlational research- designed to discover relationships among
variables and to predict future events from present knowledge

TYPICAL DESCRIPTIVE
 ● Cross sectional survey- single variable research- survey
● IV and DV are not used because causality or assessing relationships between
variables are not an aim
● To get a clearer description of the phenomenon
● Protection against bias
- conceptual & operational definitions of variables
- Sample selection techniques
- Large sample size

COMPARATIVE DESCRIPTIVE

● To describe variables and to examine difference in variables between two groups

that occur naturally in a setting
● Comparative descriptive research study the same variable in two populations
 DESCRIPTIVE CORRELATIONAL

● Examines association between two or more variables in a single group

● Do the variables co-vary?
● To quantify the strength and direction of
Relationship
● They are potential foundation for experimental studies
● Useful when the variables cannot be manipulated
● Eg: is there a relationship between coffee consumption and exam performance
PREDICTIVE CORRELATIONAL DESIGN

● To predict the value of one variable from the value of another variable ●
Trying to find causal phenomenon (initial step) and thus the variables are called
independent and dependant variables
● Statistics used in such a research are regression analysis and multiple variate
statistics ● Example: to determine if smoking could predict the incidence of lung
cancer

EXPERIMENTAL RESEARCH

● One or more IV are manipulated (treatment) -randomly assigned to different

treatment levels- effect of the treatments on outcomes (dependent variables) are
observed ● The unique strength- internal validity (causality)-ability to link cause and
effect treatment & controlling for the effect of extraneous variable
● Best suited for explanatory research (rather than for descriptive or exploratory
research), where the goal of the study is to examine cause-effect relationships
● When there are limited and well-defined set of independent variables that can either
be manipulated or controlled
● Loftus & Palmer Car crash experiment- verb used- smash/collide/bump/hit/contact
MENTAL RESEARCH DESIGNS

● “IV” stands for independent variable (also called the “treatment”)

● “DV” for dependent variable
● “EV” for extraneous variable
● In experimental research we manipulate the IV and observe any resulting change in the
DV. Because we are manipulating it experimentally, the IV will probably assume
only a very few values, maybe as few as two
● The DV may be categorical or may be continuous
● The EVs are variables other than the IV which may affect the DV
● To be able to detect the effect of the IV upon the DV, we must be able to control the EVs

CAMPBELL AND STANLEY NOTATION

 ● X- exposure to a treatment/experiment variable
● O- process of observation or measurement
● Xs & Os in a single row are given to a single participant
● Left to right- temporal sequence
● Two rows vertically arranged are simultaneous treatments
● R indicates random assignment to separate treatments
● Dashed separation line- non randomised groups; No dashed lines- randomised groups

PRE- EXPERIMENTAL DESIGNS

● Pre-experimental designs neither have random assignment nor true

treatment manipulation
● Holfing’s hospital study on obedience (1966)- 20mg instead of 10mg. 21 out of 22
nurses complied. People are very unwilling to question supposed 'authority', even when
they might have good reason to.

PRE - EXPERIMENTAL

● Do not provide a control or equivalent of a control group

● Have some basic qualities of experimental designs
● Often conducted before a true experimental research
ONE SHOT CASE STUDY

● The researchers simply find some group of subjects who have experienced an event X
and then measure them on some criterion variable
● The researcher then tries to relate X to O
● No variable is manipulated- No pretested data
● No control/comparison group
● Eg: conducting research to determine the effect of reading stories to children at night
impact language skills in children
● Implicit comparison group using Norms on the measuring instrument
● Researcher doesn’t know whether subjects already differed from the “norms” prior
to experiencing the X
ONE GROUP PRE TEST POST TEST

● There is no control group

● A single group undergoes –
1. A pre-experimental observation
2. The experimental treatment
3. Post experimental observation
● Threats to internal validity due to: History, maturation, testing instrumentation ● Eg:
pretesting ADHD children, giving cognitive training and post testing their attention level

STATIC GROUP COMPARISON

● No Pre-tests
● Two groups
● The experimental group is exposed to the
treatment and then evaluated ● Other group
(un matched) evaluated without any treatment
● Eg: effect of praising on learning mathematics in school going children

QUASI-EXPERIMENTAL
● Quasi means partial or half or resembling
● Due to ethical issues and other constraints- Ex-smoking in pregnant women -
Treatment manipulation done (by selection) , but random assignment not done -
Mostly conducted in field settings and not laboratory
- Constraints in controlling variables by researcher
● Internal validity & external validity- threats
- History threat- covid19 condition
- Maturation threat- age not controlled- in children
- Test effect threat- social desirability biases- demand characteristics- practise effect -
Selection bias- groups are not comparable
- Regression to the mean- score closer to middle in the second time -
Attrition- incomplete data due to leaving of participants (dropout)

QUASI-EXPERIMENTAL DESIGNS
● Randomness is important!- eliminate any possible biases and balances the groups
● When randomness is neither possible nor practical
● Disadvantage: do not
control for all confounding
variables so they can’t
completely rule out
alternative explanations for
the observed results
● Possible solution:
Matching instead of
random assignment-
individual matching;
ex-post-facto
matching; aggregate
matching (using
statistics)

EX-POST-FACTO RESEARCH
● After-the-fact Research/Causal
Comparative Research
● Investigation starts after the fact has
occurred
● Researcher tries to
 find out the causes behind its occurrence by going backwards in history
● Manipulation is not possible
● E.g. Earthquake hit area, Cause of delinquency, etc.

NRCG-PRETEST-POST TEST DESIGN

● Nonrandomized Control Group Pretest- Posttest Design
● It is better than static group comparison but not upto
pretest-posttest control group
● Have pre-tests and post tests
● Two groups with non-random participant assignment
● No guarantee for group similarity
● Reasonably conclude the cause and effect relationship
SIMPLE TIME SERIES DESIGN
● Measuring the DV on several occasions in one single group- series of observations
● Then introducing the intervention and then observing again
● If the second series of observations have considerable change, we conclude was the
introduced factor made the effect
● The first set of observations is referred to as baseline
● Weakness: history- confounding variable- event in the outside world happens accidentally
at the same time
● Single group interrupted time series

Eg:

CONTROL GROUP TIME SERIES DESIGN

● A variation of design simple time series
● but with higher validity by adding control group
● Solves the history problem
● If an outside event is the cause of any change, it should influence both groups
 EQUIVALENT TIME SAMPLE DESIGN
● Periodic introduction of different treatments followed by measurements over a time
● Eg: Effect of student integrity in taking exams when an invigilator is present or not
present
● Vulnerable to multiple treatment interference
TRUE EXPERIMENTS
● Gold standard – most rigorous of all designs
● Main 3 characteristics
1. Random selection and assignment of subjects to groups
2. Full control of the researcher on variables and subject
3. Laboratory setting
● Blinded study- Placebo study & double blind study
RESEARCH METHODOLOGY AND STATISTICS - 4

PRETEST POSTTEST CONTROL GROUP

● Both groups are carefully selected through randomization

● Pre-test, subjected to treatment and then post test
● We are sure whether the change happened after the treatment
● Eliminate other possible explanations involved- confounding
variables ● draw conclusions about cause and effect relationships

POSTTEST ONLY CONTROL GROUP

● Pretesting might influence the results of an experimental manipulation- practice

effect pre-test sensitization effect
● Randomly assigned two equivalent groups given treatment and observed afterwards
● Similar to Static group comparison except that the groups are equivalent in all
aspects except for treatment and randomly assigned
SOLOMON FOUR GROUP DESIGN

● An extension of Pretest Posttest control group design

● Proposed by Solomon in 1949
● Compare maturation change
● Eliminates the pretest influence
● Enhances the external validity
● It has a larger sample
● Drawback: – Needs more time and energy from the researcher

INDEPENDENT AND CORRELATED SAMPLES DESIGN

● Independent samples design: cases are assigned to groups in a way that should not
create any correlation between the scores in any one group and the scores in any other
group. Also called between subjects designs (randomly selected)
● Correlated samples design cases are assigned to groups in a way that should produce
a positive correlation between the scores in any one group and the scores in any other

RANDOMIZED BLOCK DESIGNS

● When you have a blocking variable (positively correlated with the DV)
● Eg: age/ gender as blocking variable
● Variability within blocks is less than the variability between blocks
● You match the subjects up in blocks of two, such that within each block the subjects are
nearly identical on the blocking variable(s)
● Then you randomly assign, within each block, one case to Treatment 1, another to
Treatment 2 (control)
● Design could be described as matched pairs
 REPEATED MEASURES DESIGN
● Often interchanged with the term 'within subjects’ design
● A subtype of the within subjects design
● Uses the same subjects with every treatment of the research including the control
● Disadvantages: practice effects; boredom and fatigue; dropout
● Crossover design (counterbalancing)- ensuring that all of the subjects receive all of the
treatments- presence
of carryover effect
● The within or
intrasubject variability
in will be smaller than
the between or
intersubject variability
used for the
comparison of
treatments in parallel
groups design

LATIN SQUARES
● The crossover design is a type of Latin square
● In a Latin square the number of treatments equals the number of patients
● Another important factor- order of treatment
 ● Order of treatments included in a balanced way
● The net result is an N X N array (where N is the number of treatments or patients) ● This
is most easily shown pictorially with N letters such that a given letter appears only once in
a given row or column
● A-CT
● B-BT
● C-CBT
● D- Pharma

SMALL N & SINGLE-SUBJECT DESIGNS

● Single- case design is a
Small N design with only
one single participant ● To establish the existence of cause effect
relationships
● Different from case studies- not descriptive, experimental
● Typically involve a series of observations made over a time
1. Baseline (ABA) Designs- withdrawal design
2. Multiple-Baseline Designs- Across subject, behaviour, situation
3. Changing-Criterion Designs- treatment incrementally increased
4. Discrete-Trials Designs- To study invariant phenomenon- psychophysics studies- no
baseline present
BASELINE (ABA) DESIGNS - WITHDRAWAL DESIGN

MULTIPLE - BASELINE DESIGNS

CHANGING-CRITERION DESIGNS

CHARACTERISTICS
 ● One or only a few subjects tested ("single-subject designs")
● Subjects not put into groups, but run as individuals
● Experiments are long, allowing performance to stabilize over time
● Data from subjects usually not combined but considered
separately ● Data analysed visually with minimal use of inferential
statistics
● Used in educational, clinical and animal research studies

FACTORIAL DESIGN
● Have more than one independent variable
● See the combined effects of two or more IVs on a single DV
● Effect of expertise & suggestion on false memory (Mazzoni, Loftus, Seitz &
Lynn: 1999)- Participants never been bullied
- Asked to report a recurring dream to
(1) another participant or (2) Psychiatrist
- In experiment condition (a) confederates (1 & 2) say “ this types of dreams are as a result
of being bullied as a child”
- In control condition (b) confederates ( 1 & 2 ) say “this type of dreams are a result of
feeling powerless as a child”
- One week later: As a separate study they were asked if they were bullied as a child

DISTRIBUTION OF A VARIABLE

● Frequency
distribution: a
table/ polygon of
how many times
each score occurs
 ● Histogram is a graph plotting values of observations on horizontal axis- each bar show
how many times each value occurred in the data set
● Normal curve
- In an ideal world data would be distributed symmetrically around the centre of all scores -
If we drew a vertical line through the centre of the distribution then it should look the same
on both sides
- This distribution is known as a normal distribution and is characterized by the
bell-shaped curve
DEVIATION FROM NORMAL

● Skewness- lack of symmetry

- Positive / right (>+1)
- Symmetrical (between +1 & -1)
- Negative / left (< -1)
● Kurtosis- heaviness of tails- measure of outliers
- Leptokurtic (Kurtosis > 3)- heavy tail- many outlier
- Mesokurtic (kurtosis =3; excess kurtosis= 0)
- Platykurtic (Kurtosis < 3)- light tail- less outlier
MEASURES OF CENTRAL TENDENCY

● To describe the typical, average and centre of a distribution of

scores ● Mean, Median, Mode
- Mode is the most frequently occurring score in a distribution
- Median is the midpoint of a distribution of scores
- Mean is the average of a distribution of scores

MODE & MEDIAN

● Mode is the tallest bar-the score that occurs the most is the mode
● Median is the middle score when scores are ranked in order of magnitude -
arrange these scores into ascending order
- (n + 1)/2
- unaffected by extreme scores
MEAN
 ● Add up all of the scores and then divide by the total number of
scores ● Used only with interval or ratio data
● It uses every score
● Disadvantage- it can be influenced by extreme scores
● Mode and median ignore most of the scores in a data set
● Z score is a numerical measurement that describes a value's relationship to the
mean. Z-score = 0, score is identical to the mean score

DISPERSION

It is the spread of the scores

1. Range
 ● Range of scores: take the largest score and subtra
affected dramatically by extreme scores
● Quartiles are the three values that split the sorted
calculate the median, which is also called the se
into two equal parts.
● Interquartile range- cut off the top and bottom 25
of the middle 50% of scores (Q3- Q1) (half lost

PERCENTILE RANKS

● Quantiles are values that split a data set into equa

that split the data into four equal parts
● Noniles (split the data into 9 equal parts)
● Percentiles (split the data into 100 equal pa
percentageof scores in the distribution that ar
number lower andconverting to a percentage of

2. Inter-Quartile range
3. Percentile ranks
4. Variance
5. Standard deviation
6. Coefficient of variation
RANGE & INTER-QUARTILE RANGE
VARIANCE
 ● Deviance-the difference between each score and the mean- some deviations
positive some negative
● Sum of squared deviances (SS)- add up the square of deviance for each data point in
the distribution- total dispersion score
● Variance- average this total dispersion. Why N-1?

DEGREES OF FREEDOM

● Ball in cup- data required for understanding the distribution- mean takes in all the
scores ● If mean is already in the calculation- we only require the N-1 data points ● Mean
= Total / N
● Total = Mean * N

1 2 3 4 5 Mean

8 5 15 14 ? 10

● Sample variance = Population variance when calculated using N-1=

Unbiased ● Sample variance not equal to population variance when used N for
calculation

STANDARD DEVIATION

● Variance gives the squared units

 ● Standard deviation is the square root of variance
● SD of a distribution is the average distance between the scores and the mean
● Computing: Finding mean > finding difference between each score and the mean>
squaring each difference> finding the mean of these squared differences> and finally
finding the square root of that mean.
COEFFICIENT OF VARIATION

● Also known as Coefficient of dispersion

● SD can be understood only in the context of Mean of the data
● CV is unit-less
● Used to compare variations of distributions with different units
● Difficult to calculate if the scores are both positive and negative- impossible if Mean=0

RESEARCH METHODOLOGY AND STATISTICS - 5

NORMAL DISTRIBUTION
 ● Gaussian distribution- Carl Friedrich Gauss
● Mean=median=mode= 0 ; SD=1
● The values of skew and excess kurtosis are 0
● Half of the population is less than the mean and half is greater than the mean. ●
Empirical Rule allows you to determine the proportion of values that fall within certain
distances from the mean
● ≈68% of the data falls within 1 SD
● ≈95% of the data falls within 2 SD
● ≈99.7% of the data falls within 3 SD

EMPIRICAL RULE OF NORMAL CURVE

● We can use Z scores to understand where a specific observation falls relative to the
entire distribution
● Can place observations drawn from different normal distributions (diff mean and SD)
on a standard scale and compare them
● This process is called standardization
● To standardize your data- convert the raw scores into Z-scores
 EXAMPLE - COMPARE APPLES TO ORANGES

● Compare their weights- we have 110 gms apple and 100 gms orange

Apples Oranges

Mean weight grams 100 140

Standard Deviation 15 25

● Apple = (110-100) / 15 = 0.667

● Orange = (100-140) / 25 = -1.6
● Apple weighs more than the average apple
● Orange has –ve Z-score- below the mean
SAMPLING DISTRIBUTION OF MEAN

● If repeated random samples of a given size n are taken from a population- for a
quantitative variable, where the population mean is μ (mu) and the population
standard deviation is σ (sigma) then the mean of all sample means (x-bars) is equal to
population mean μ (mu)

STANDARD ERROR OF MEAN (SEM)

 ● The variability of the sampling distribution is the standard error of the mean
(SEM) ● SEM assess how far your sample mean is likely to fall from the population
mean, it evaluates how closely your sample estimates the population

DESCRIPTIVE & INFERENTIAL STATISTICS

● Descriptive statistics summarize the characteristics of a data set

- Eg: Measures of central Tendency, Dispersion, Correlation
● Inferential statistics allow you to test a hypothesis or assess whether your data
is generalizable to the broader population
- Eg: Test statistics

PROCEDURE OF HYPOTHESIS TESTING

Making a Formal Statement > Selecting a Significance Level > Deciding the Distribution to
Use > Determining the suitable test to use > Selection of a Random Sample > Collecting
required data > Calculate the probability of the sample > Reject or Accept Null hypothesis

CONFIDENCE LEVEL
 ● Confidence interval denotes the values within which population parameter is
expected to lie. Consists of the upper and lower bounds of the estimate you expect to
find at a given level of confidence
● Confidence level- Probability that a parameter will fall between a set of values- If
95%, 1/20 times parameter is missed.

LEVEL OF SIGNIFICANCE & P VALUE

● Ho & Ha
● The significance level, also denoted as alpha or α, is the probability of rejecting the
null hypothesis when it is true- Predefined (α may be kept as 0.1, 0.05, or 0.01) ● P
value or calculated probability is the estimated probability of rejecting the null
hypothesis (Ho) of a research question when that hypothesis is true
DISTRIBUTIONS

TEST STATISTICS AND DISTRIBUTIONS

● Test statistic- a number calculated from a statistical test of a hypothesis (central value,
variation, sample size, and no. of predictor variables)
● Discrete distributions:
- Uniform (outcomes are equally likely)
- Binomial (success/ failure in n trials)
- Bernoulli (success/ failure in one trial)
- Geometric ( no. of failures before success)
- Poisson (probability of n events)
 ● Z test-normal distribution; t-test- t distr.; ANOVA- F distr.
● Symmetrical distributions (t and z distributions)
● Asymmetrical (F and chi-square distributions)
TYPE I AND TYPE II ERRORS

● Whether there is an effect in our population two possibilities:

1. There is, in reality, an effect in the population- Ha
2. There is, in reality, no effect in the population- Ho
● We check the test statistic probability and decide which is more probable ● Type I
error alpha (0.05 or 5%)-(false-positive) occurs if an investigator rejects a null
hypothesis that is actually true in the population
● Type II error beta (0.2 or 20%) -(false-negative) occurs if the investigator fails to reject
a null hypothesis that is actually false in the population.

POWER ANALYSIS
● Ability of a test to find an effect is known as its statistical power - probability of
finding an effect when it exists.
● Power of test= 1 − β (Type II error)
● We typically aim to achieve a power of 0.8 (1 - 0.2)
● Power can be calculated and reported for a completed experiment to comment on
the confidence one might have in the conclusions drawn from the results of the study

EFFECT SIZE
 ● A significance does not always tell about the strength of the effect
● Bigger effects will be easier to spot and be significant
● It is a standardized measure of the magnitude of observed effect.
● Most common measures of effect size are
- Cohen’s d- d = 0.2 (small), 0.5 (medium) and 0.8 (large)
- Pearson’s correlation coefficient r = 0.1 (small), 0.3 (medium), 0.5 (large)

CORRELATIONAL ANALYSIS
● Relationship between predictor variable and outcome variable
● We measure what naturally goes on in the world without directly interfering with
it ● r is the coefficient of correlation
● Types- Pearson correlation, Spearman correlation, partial, multiple, special
PRODUCT MOMENT CORRELATION
 ● Also known as Pearson r
● Two continuous variables, or one continuous variable and a categorical variable-
two categories
● It can vary from –1 (a perfect negative relationship) through 0 (no relationship) to +1
(a perfect positive relationship)
● > +0.5 or <-0.5 is strong correlation
● >+0.3 or <-0.3 is moderate correlation
● Between -0.3 to +0.3 weak correlation
● 0 means there is no relationship at all

PLOTTING PREDICTOR VARIABLE AND OUTCOME VARIABLE