Psychology as an

empirical science: An
introduction to
research methods for
GCSE/GCE students

ATP Conference July 2010

By
Dr George Varvatsoulias
 Learning objectives
1. Exploring research methods in psychology (a general
account)
2. Applying research methods (in the study of memory)
3. Quantitative and Qualitative research methods in
psychology
4. How do we design a study (in detail)
5. Are we as teachers ‘committed’ to using empirical
methodology as scientists?
 Starting with a story...
• In winter people wear • Observation
heavy clothes
• The weather outside is • Rationale
cold and there are
storms
• People wear heavy • Hypothesis
clothes during winter in
order to protect
themselves
• It is so because it is • Prediction
winter

Choose each of the And match it with one of the

above statements above items
Examples of research
designs
Between-participants/Within
participants designs
Types of research
 Making sense of our data
via...

Descriptive statistics Inferential statistics

refer to the central refer to statistical
tendency techniques
Applying research methods in the
study of memory (an example)
 Summary No. 1

• Research methods are a useful tool in psychology

towards studying human behaviour
• Different research types as well as statistical
techniques are employed for different aspects of
research
• In designing a study we pose a research question in
order to seek evidence that will support our
hypothesis
 Quantitative and Qualitative
Research Methods in
Psychology
Specification Content Comment
• Advantages/Disadvantag • Description of each
es of research methods method
related to the scientific • Being aware of at least
nature of psychology two advantages and two
• Experiments weaknesses
• Investigations through • How research methods
correlational analysis relate to the scientific
• Naturalistic observations nature of psychology
• Questionnaires • Distinguishing
• Interviews psychological knowledge
from commonsense
• Ethical guidelines
 Terminology (1)
• Ways in allocating participants to participate to a study: randomly,
opportunity sample (friends, relatives, and anyone who is around that time),
self-selection (via advertising the study, and finding participants who
volunteered themselves – volunteer sample)
• IV (Independent Variable): It is manipulated by the experimenter, i.e. the
experimenter puts down the ‘rules’ to be followed by each participant; each
IV is independent of the value of any other variable (IV)
• DV (Dependent Variable): It is depended on the value provided by the IV
(existence or dismissal of such dependence is questioned in relation to the
hypothesis investigated)
• One-tailed/Two-tailed hypotheses (othewise called: directional/non-directional
hypotheses: One-tailed or directional hypothesis refers to the increase or
decrease of one variable in favour to the other; once there may be an effect
of one variable to the other, and the direction of this effect is unknown we
have a two-tailed or non-directional hypothesis
• Null hypothesis: No significant difference (in our case: no real effects in
measuring the blood pressure of participants)
 Terminology (2)
• Between participants/within participants:
randomly allocating participants to different
groups/no random allocation and all participants
perform on all conditions
• Probability Value (p): Is normally set to p<0.05
(50-50 chance of finding an effect), or p<0.01
(10% chance of finding an effect)
• Effect size (d): The magnitude of the difference
between conditions (the difference between the
means of both or more conditions of a study in
terms of their standard deviations): x₁-x₂
mean SD
 Research methods listing
labels (1)
Research Nature and use Pros & Cons Ethical issues
methods
Lab experiment Manipulation of For: Causal Deception;
the IV to find conclusions; No Informed
effects on DV confounding consent;
under controlled variables; Psychological
conditions Replication harm
Against: Artificial;
Experimenter
and Participant
effects

Field experiment Looking for For: Causal Informed

causal conclusions; consent;
relationships in Ecological Difficulty in
natural places validity; debriefing;
Participants’ Privacy
effects are
 Research methods listing
Research labels
Nature (2)
and use Pros & Cons Ethical issues
methods
Natural Non-manipulation For: Look for Confidentiality
experiment of IV; No random research where
allocation of IV is not
participants; manipulated;
Quasi- Real problem
experimental studies
Against: No
causal
relationships can
be found; Many
confounding
variables

Correlational Co-variables are For: Variables Misunderstanding

analysis studied for cannot be of findings
positive, manipulated;
negative, or zero Correlation does
association not infer
 Research methods listing
labels (3)
Research Nature and use Pros & Cons Ethical issues
methods
Naturalistic Variables are free For: Can study Informed
observations to vary behaviour where consent;
cannot Difficulty in
manipulate debriefing;
variables; Privacy
Ecological
validity
Against: No
control of
confounding
variables;
Observer bias,
low observer
variability

Questionnaires Set of questions For: Collection of Confidentiality;

huge amount of Privacy
 Research methods listing
labels (4)
Research Nature and use Pros & Cons Ethical issues
methods
Interviews Pre- For: Rich data; Confidentiality;
determination of Telephone Privacy
questions, or interviews
created in Against: Social
response to desirability bias;
answers Skilled personnel
is required
 Design & Implementation
(1)
• Assumed study: A psychologist chose to
investigate the kinds of tasks which created
stress for teachers. He asked for teachers
from several schools to volunteer to take
part in the study, and selected 20 male
teachers. Each volunteer was asked to
perform two tasks: Presentation of a
prepared speech to a group of 200 students;
marking a set of books full of mistakes. After
each task the researcher recorded the blood
pressure of each participant
 Design & Implementation
(2)
• Research method: Field experiment (Class environment reactions
are tested, both at the speech as well as at the marking tasks)
• Aim of the study: Investigating kinds of tasks to see if they create
stress to teachers
• Hypothesis: Teaching to a large number of pupils as well as marking
their homework may have an effect in teachers’ blood pressure. The
hypothesis will be non-directional
• Design: Quasi-experimental, because we don’t randomly allocate
participants to conditions; we have only one gender group of
participants; we cannot manipulate each participant in terms of
each task performance
• IV/DV: Teaching before a large group & Marking homework – 2 IVs;
Blood pressure recordings (score noted for each participant) - DV
• Within participants study: All participants will take part in both
conditions
• Co-variables in the study: Teaching before a large group & Marking
homework
 Design & Implementation
(3)
• Pros & cons of the design employed: Participants seem to not having been informed about
the purpose of the study, and then to be asked to take place; no reference to how the
experimenter would deal with the data collected
• Confounding/extraneous variables: No reference to the age of each participant, so individual
differences in performance to be discussed; anxiety proneness of teachers prior to teaching
was not taken into account, so to possibly influence their performance during the study as
well as their blood pressure scores after both tasks; speaking to public as a stressful life
event was not discussed; female participants could also be employed
• Ethical issues: Lack of further procedural information to participants before the study to be
taken place, such as anonymity, debriefing, etc
• Internal validity of the study: No reference about the subject of the talk to pupils, or if each
participant could choose a subject of his own likeness; if so, whether different topics affect
teachers’ performance
• External validity: It isn’t known if speaking to the public out of a class environment would
have the same results; external validity is not clear in many types of the experimental
designs
• Ecological validity: A variety of reasons could fall in replicating the study with different
teachers in different classes; probably the study is not ecologically valid, i.e. it cannot rely
on the collection of the same data when applied to different environments
 Design & Implementation
(4)
• Discussion: The study could be replicated by asking all participants to perform on
the same tasks via two different conditions for each of the IVs: easy topic/difficult
topic, and good homework/bad homework conditions
• Selection of participants: Volunteer sample (self-selection)
• Pilot study prior the study to be conducted: Half of volunteer teachers prepared a
speech and delivered it to a small audience of students; they also marked the
scripts of both exceptional and non-exceptional students. In this way, confounding
variables such as the ones named above could be avoided
• Statistical procedure possibly to fail: By what means of an instrument the talk and
marking of homework would be measured? Would that be a quantitative approach
or qualitative? In which way data will be collected in terms of the performance of
teachers, so these to be discussed in relation to their blood pressure scores after
both tasks had been performed? Also, would increased of decreased blood
pressure be associated to both tasks, or to one of them, or otherwise (looking for
stressful life events other than the participation to these tasks)?
 Summary No. 2
• Different research methods are used for
different types of studies
• By clearly stating the aims, the hypothesis, and
the design of a study we can carry on with the
other issues associated to the study
• Ethical issues should be presented before
embarking on a study
• The ways to selecting participants can help us
to identify possible confounding variables
related to the execution of the experiment
 How do we design a study
(in detail-1)
• A research project into eating
problems sought to find out about
adolescent and pre-adolescent
dieting practices. The study involved
the use of a questionnaire, including
both quantitative and qualitative
questions. The questionnaire was
piloted with a small group of
university students
 How do we design a study
(in detail-2)
• Aspects to be considered in the design of this study
1.Research method
2.Aims
3.IV/DV
4.Hypothesis
5.1-tailed or 2-tailed hypothesis
6.Selection of participants
7.Quantitative (numbers)/Qualitative (interviews)
8.Possible confounding variables not only to the type of
study, but also to the use of a questionnaire, and the
selection of participants
 Studies using correlation
• Positive correlation: When both variables increase in a study
• Negative correlation: When one variable increases and the
other decreases
• Co-efficients in correlation: A figure (number)
demonstrating the degree to which both variables are
related (positive correlation +1/negative correlation -1). The
closer the correlation to a perfect relationship (+1 or -1),
the stronger the association between variables
• Stating correlations:
1. Zero (0 or closer to zero)
2. Weak (±0.1-±0.3)
3. Moderate (±0.4-±0.6)
4. Strong (±0.7-±0.9)
5. Perfect (+1/−1)
 Example of a correlational
analysis
“Larger class sizes are positively
correlated to exam success”
1. Try to criticise this research, and think
of the arguments you might give to
support the claim that smaller class
sizes are actually of greater benefit to
schoolchildren
2. Consider any extraneous or
confounding variables that might have
affected the above claim
 Advantages of correlational
designs
- Predictions more easily plausible: If there
has been found a correlation, then we can
make predictions regarding the effect of one
variable to the other
- Allows quantification of relationships:
Correlations show the strength of a
relationship between two variables
- Manipulation is not possible: No requirement
of manipulating the behaviour of
participants, so correlations are ethically-
prone to collecting data and analysing them
 Disadvantages of correlational
designs
- Quantification problems: Although, correlations seem to appear quite low, they
can be meaningful or significant if the number of scores recorded is quite high.
Contrary to that, we may have a large figure in correlation, and the results not to
be significantly meaningful
- Cause and effect: We do not infer causation from correlation. Effects are
practically inexistent since the relationship between variables cannot justify (be
sure of) the direction of the association, for no IV or DV can be selected
- Extraneous associations: There may be other variables, except the ones studied;
i.e. blowing wind and people’s character traits in a normal winter day, however
what about summer temperatures?
- Only works for linear relationships: Straight-line relationships are tested whether
in a negative or positive correlation. If we have a curvilinear correlation it means
we have aero significance in our results (correlation coefficient/Pearson’s r)
 An example for discussion
• Possible correlations:
1. Lower-income parents have more children
2. Older people make poorer eyewitnesses
3. Sales of running shoes have increased at the
same time as sales of personal computers
D.Think up of possible extraneous variables for
the above assumed correlations
E. Consider hypothetical correlations that might
account for a link between those variables
F. By changing one variable with a confounding
variable (modifying our hypothesis) can we
work out a more viable correlation?
 Observational techniques
1. Participant observation: By
becoming, the experimenter a
‘participant’ of his/her participants’
observation actual response to the
study
2. Non-participant observation: By
being the researcher involved in the
observation of his/her participants’
behaviour from a distance
 Advantages of observational
techniques
- High external validity: Observation takes place
in the natural environment; participants tend to
behave naturally, and results can be easily
generalised to other settings
- Practical method: Social context instead of
cooperation with participants: natural
environments where deliberate manipulations
would be unethical or impractical
- Fewer demand characteristics: Participants do
not know they are observed, so with their
responses they cannot satisfy the experimenter
Disadvantages of observational
techniques
- Cause and effect: With observations we do not have
manipulations. Little control over extraneous variables
- Observer bias: The question about reliable observation because of
the danger the results coming out of observations to be fabricated,
it is cross-checked with another’s experimenter procedure of the
same study (inter/between reliability). Also, by checking all stages
of the study and by comparing them with one another from
inception to completion, we look for intra/within reliability
- Replication: Lack of proper control participants (proper, in terms of
repeating the same study again) leads to a stable problem of
reliability and validity of all observations (all of us do not possess
the same ‘senses’, or ‘sensitivity abilities’ when observing
situations)
- Ethics: Issues of informed consent and invasion of privacy arise
when participants are unaware of being observed
- Practical problems: It is difficult for experimenters to remain
unobserved, so to avoid demand characteristics, or to categorise
the observed behaviours accurately
 Practical learning: Classroom
observation
• You are a government inspector of psychology classes.
You decide to use a non-participant observation
technique:
1.Outline exactly how you observe a psychology lesson
2.Consider the number of times students contribute to
the lesson; how frequently they take notes, their level
of engagement (behavioural categories). How would
you record the categories?
3.Observe a psychology lesson and write up your
experiences
4.Was the lesson accurately taught to students?
 Questionnaires
1. Closed (fixed) questions or
statements: Usually ask for ticking
boxes. Easy for quantification, but
are restriction-prone
2. Open questions/statements: Allowing
participants to answer in their own
words. They are difficult to analyse,
but allow freedom of expression
 Advantages of
questionnaires

• Quick and clear

• Large samples
• Quantitative and qualitative analysis
• Replication – by the use of
standardised questions
 Shortcomings of
questionnaires
1. Misunderstandings: Misinterpretation of
questions by the participants
2. Biased samples: Needs time to be spent;
asked from people who are literate, or
samples are chosen to satisfy the findings
expectation of the experimenter (his/her
hypothesis to be supported)
3. Low response rate: Not all questionnaires are
returned back
4. Superficial issues: Sensitive issues arise that
need detailed understanding
5. Social desirability: Answers being presented
in a socially positive light
 Interviews
1. Structured (formal) interviews: Identical
questions addressed to participants in
the presence of the experimenter
2. Unstructured or informal interviews: Less
controlled and involve an informal
discussion as a particular topic.
Questions are posed in relation to
participants’ answers to the subject
 Advantages of interviews
• Complex issues: Complicated or
sensitive issues are more easily to be
dealt with in a face-to-face interview
• Ease misunderstandings: Interviewers
have the chance to clarify any question
of participants during interview
• Data analysis: Both quantitatively and
qualitatively: Variety and flexibility to
the collection of data
• Replication: The more structured an
interview the more easier to replicate
Disadvantages of interviews
• Interviewer or interviewee effects: Biased questions so
to elicit ‘fabricated’ answers. Interviewers may bias
the respondents’ answers, but interviewees may
develop demand characteristics, and social desirability
issues
• Interview training: structured interviews do not require
formal training. However, skills are needed for an
unstructured interview to be conducted
• Ethical issues: If participants do not know the true
purpose of the interview, ethical issues such as
deception and respect to privacy may arise
 Think of a topic both of a
quantitative and qualitative
research
• Examples:
1.Investigating people’s views on
euthanasia
2.Questioning people’s exposure to
stressful life events, and how they
have experienced their impact
 Summary No. 3
• Most common aspects that arise in both
quantitative and qualitative research
are:
1.Experimenter and Participant bias
2.Social desirability issues
3.Structured and unstructured interviews
4.Participant observation and demand
characteristics
Reading a story and preparing
it for research
• A research project on memory aimed to find
out about individuals’ earliest memories. In
order to do this the research team conducted
unstructured interviews with people aged 10
to 50 years. The interviewers started with a
set of standard questions, but then adapted
their questions in response to the answers
given by the interviewees
 Getting ‘committed’ as
teachers in research
methodology
• Research methodology is difficult
also for us teachers...
• Assuming hypothetical studies while
in class with the students
• Outlining the combination between
theory and research practice (theory
as a tool, research as the means
towards maintaining findings)
TEΛΟΣ
(The End)