Research Methods in Psychology Typically, these are always written ‘There

Research methods in psychology are will be a difference ….’

systematic procedures used to observe, All research has an alternative hypothesis
describe, predict, and explain behavior (either a one-tailed or two-tailed) and a
and mental processes. They include corresponding null hypothesis.
experiments, surveys, case studies, and Once the research is conducted and
naturalistic observations, ensuring data results are found, psychologists must
collection is objective and reliable to accept one hypothesis and reject the
understand and explain psychological other.
phenomena.
Hypotheses So, if a difference is found, the
Hypotheses are statements about the Psychologist would accept the alternative
prediction of the results, that can be hypothesis and reject the null. The
verified or disproved by some opposite applies if no difference is found.
investigation.
Sampling techniques
There are four types of hypotheses:
Sampling is the process of selecting a
1. Null Hypotheses (H0) – these predict
representative group from the population
that no difference will be found in the
under study.
results between the conditions.
Typically, these are written ‘There will
A sample is the participants you select
be no difference…’
from a target population (the group you
2. Alternative Hypotheses (Ha or H1)–
are interested in) to make generalizations
these predict that there will be a
about.
significant difference in the results
between the two conditions. This is
Representative means the extent to which
also known as the experimental
a sample mirrors a researcher’s target
hypothesis.
population and reflects its characteristics.
3. One-tailed (directional) hypotheses –
these state the specific direction the
Generalizability means the extent to
researcher expects the results to move
which their findings can be applied to the
in, e.g. higher, lower, more, less. In a
larger population of which their sample
correlation study, the predicted
was a part.
direction of the correlation can be
either positive or negative.
Types of Sampling
4. Two-tailed (non-directional)
1. Volunteer sample: where participants
hypotheses – these state that a
pick themselves through newspaper
difference will be found between the
adverts, noticeboards or online.
conditions of the independent variable
2. Opportunity sampling: also known as
but does not state the direction of a
convenience sampling, uses people
difference or relationship.
who are available at the time the study
is carried out and willing to take part.
It is based on convenience.
3. Random sampling: when every person that variables are in a form that can be
in the target population has an equal easily tested.
chance of being selected. An example
of random sampling would be picking For instance, we can’t really measure
names out of a hat. ‘happiness’, but we can measure how
4. Systematic sampling: when a system is many times a person smiles within a two-
used to select participants. Picking hour period.
every Nth person from all possible
participants. N = the number of people By operationalizing variables, we make it
in the research population / the easy for someone else to replicate our
number of people needed for the research. Remember, this is important
sample. because we can check if our findings are
5. Stratified sampling: when you identify reliable.
the subgroups and select participants
in proportion to their occurrences. Extraneous variables are all variables
6. Snowball sampling: when researchers which are not independent variable but
find a few participants, and then ask could affect the results of the experiment.
them to find participants themselves
and so on. It can be a natural characteristic of the
7. Quota sampling: when researchers participant, such as intelligence levels,
will be told to ensure the sample fits gender, or age for example, or it could be a
certain quotas, for example they might situational feature of the environment
be told to find 90 participants, with 30 such as lighting or noise.
of them being unemployed.
Demand characteristics are a type of
Variables extraneous variable that occurs if the
Experiments always have an independent participants work out the aims of the
and dependent variable. research study, they may begin to behave
in a certain way.
The independent variable is the one the
experimenter manipulates (the thing that For example, in Milgram’s research, critics
changes between the conditions the argued that participants worked out that
participants are placed into). It is the shocks were not real and they
assumed to have a direct effect on the administered them as they thought this
dependent variable. was what was required of them.

The dependent variable is the thing being Extraneous variables must be controlled
measured, or the results of the so that they do not affect (confound) the
experiment. results.

Operationalization of variables means Randomly allocating participants to their

making them measurable/quantifiable. conditions or using a matched pairs
We must use operationalization to ensure
experimental design can help to reduce the second occasion because of tiredness
participant variables. or boredom.

Situational variables are controlled by Counterbalancing is the best way of

using standardized procedures, ensuring preventing order effects from disrupting
every participant in a given condition is the findings of an experiment, and
treated in the same way involves ensuring that each condition is
equally likely to be used first and second
Experimental Design by the participants.
Experimental design refers to how If we wish to compare two groups with
participants are allocated to each respect to a given independent variable, it
condition of the independent variable, is essential to make sure that the two
such as a control or experimental group. groups do not differ in any other
important way.
1. Independent design (between-groups
design): each participant is selected Experimental Methods
for only one group. With the All experimental methods involve an IV
independent design, the most (independent variable) and DV
common way of deciding which (dependent variable).
participants go into which group is by
means of randomization. 1. Lab Experiments are conducted in a
2. Matched participants design: each well-controlled environment, not
participant is selected for only one necessarily a laboratory, and therefore
group, but the participants in the two accurate and objective measurements
groups are matched for some relevant are possible.
factor or factors (e.g. ability; sex; age).
3. Repeated measures design (within The researcher decides where the
groups): each participant appears in experiment will take place, at what time,
both groups, so that there are with which participants, in what
4. exactly the same participants in each circumstances, using a standardized
group. procedure.

The main problem with the repeated 2. Field experiments are conducted in
measures design is that there may well be the everyday (natural) environment of
order effects. Their experiences during the participants. The experimenter
the experiment may change the still manipulates the IV, but in a real-
participants in various ways. life setting. It may be possible to
control extraneous variables, though
They may perform better when they such control is more difficult than in a
appear in the second group because they lab experiment.
have gained useful information about the
experiment or about the task. On the
other hand, they may perform less well on
3. Natural experiments are when a Correlational studies typically involve
naturally occurring IV is investigated obtaining two different measures from a
that isn’t deliberately manipulated, it group of participants, and then assessing
exists anyway. Participants are not the degree of association between the
randomly allocated, and the natural measures.
event may only occur rarely.
The predictor variable can be seen as
Case Study occurring before the outcome variable in
Case studies are in-depth investigations of some sense. It is called the predictor
a person, group, event, or community. It variable, because it forms the basis for
uses information from a range of sources, predicting the value of the outcome
such as from the person concerned and variable.
also from their family and friends.
Relationships between variables can be
Many techniques may be used such as displayed on a graph or as a numerical
interviews, psychological tests, score called a correlation coefficient.
observations and experiments. Case
studies are generally longitudinal: in Types of correlation. Scatter plot.
other words, they follow the individual or Positive negative and no correlation
group over an extended period of time.
a. If an increase in one variable tends to
Case studies are widely used in be associated with an increase in the
psychology and among the best-known other, then this is known as a positive
ones carried out were by Sigmund Freud. correlation.
He conducted very detailed investigations b. If an increase in one variable tends to
into the private lives of his patients in an be associated with a decrease in the
attempt to both understand and help other, then this is known as a negative
them overcome their illnesses. correlation.
c. A zero correlation occurs when there
Case studies provide rich qualitative data is no relationship between variables.
and have high levels of ecological validity. d. After looking at the scatter graph, if
However, it is difficult to generalize from we want to be sure that a significant
individual cases as each one has unique relationship does exist between the
characteristics. two variables, a statistical test of
correlation can be conducted, such as
Correlational Studies Spearman’s rho.
Correlation means association; it is a
measure of the extent to which two The test will give us a score, called a
variables are related. One of the variables correlation coefficient. This is a value
can be regarded as the predictor variable between 0 and 1, and the closer to 1 the
with the other one as the outcome score is, the stronger the relationship
variable. between the variables. This value can be
both positive e.g. 0.63, or negative -0.63.
 and the researcher deliberately adopts an
Types of correlation. Strong, weak, and informal approach to break down social
perfect positive correlation, strong, weak, barriers.
and perfect negative correlation, no
correlation. Graphs or charts ... There are no set questions, and the
participant can raise whatever topics
A correlation between variables, however, he/she feels are relevant and ask them in
does not automatically mean that the their own way. Questions are posed about
change in one variable is the cause of the participants’ answers to the subject
change in the values of the other variable.
Unstructured interviews are most useful
A correlation only shows if there is a in qualitative research to analyze
relationship between variables. attitudes and values.
Correlation does not always prove
causation, as a third variable may be Though they rarely provide a valid basis
involved. for generalization, their main advantage is
that they enable the researcher to probe
Interview Methods social actors’ subjective point of view.
Interviews are commonly divided into
two types: structured and unstructured. Questionnaire Method

Structured interviews are formal. The Questionnaires can be thought of as a

interview situation is standardized as far kind of written interview. They can be
as possible. Structured interviews are carried out face to face, by telephone, or
formal, like job interviews. post.

A fixed, predetermined set of questions is The choice of questions is important

put to every participant in the same order because of the need to avoid bias or
and in the same way. ambiguity in the questions, ‘leading’ the
respondent or causing offense.
Responses are recorded on a
questionnaire, and the researcher presets 1. Open questions are designed to
the order and wording of questions, and encourage a full, meaningful answer
sometimes the range of alternative using the subject’s own knowledge
answers. and feelings. They provide insights
into feelings, opinions, and
The interviewer stays within their role understanding. Example: “How do you
and maintains social distance from the feel about that situation?”
interviewee. 2. Closed questions can be answered
with a simple “yes” or “no” or specific
Unstructured interviews are informal, information, limiting the depth of
like casual conversations. A general response. They are useful for
conversation normally precedes them, gathering specific facts or confirming
 details. Example: “Do you feel anxious participants’ behavior is from a
in crowds?” distance
3. Postal questionnaires seem to offer
the opportunity of getting around the Pilot Study
problem of interview bias by reducing A pilot study is a small-scale preliminary
the personal involvement of the study conducted in order to evaluate the
researcher. feasibility of the key steps in a future, full-
4. Its other practical advantages are that scale project.
it is cheaper than face-to-face
interviews and can be used to contact A pilot study is an initial run-through of
many respondents scattered over a the procedures to be used in an
wide area relatively quickly. investigation; it involves selecting a few
people and trying out the study on them.
Observations It is possible to save time, and in some
There are different types of observation cases, money, by identifying any flaws in
methods: the procedures designed by the
researcher.
1. Covert observation is where the
researcher doesn’t tell the participants A pilot study can help the researcher spot
they are being observed until after the any ambiguities (i.e. unusual things) or
study is complete. There could be confusion in the information given to
ethical problems or deception and participants or problems with the task
consent with this particular devised.
observation method.
2. Overt observation is where a Sometimes the task is too hard, and the
researcher tells the participants they researcher may get a floor effect, because
are being observed and what they are none of the participants can score at all or
being observed for. can complete the task – all performances
3. Controlled: behavior is observed are low.
under controlled laboratory
conditions (e.g., Bandura’s Bobo doll The opposite effect is a ceiling effect,
study). when the task is so easy that all achieve
4. Natural: Here, spontaneous behavior virtually full marks or top performances
is recorded in a natural setting. and are “hitting the ceiling”.
5. Participant: Here, the observer has
direct contact with the group of Research Design
people they are observing. The In cross-sectional research, a researcher
researcher becomes a member of the compares multiple segments of the
group they are researching. population at the same time
6. Non-participant (aka “fly on the wall):
The researcher does not have direct Sometimes, we want to see how people
contact with the people being change over time, as in studies of human
observed. The observation of development and
lifespan. studies, leading to a more objective
Longitudinal research is a research appraisal of the evidence.
design in which data-gathering is
administered repeatedly over an extended This is done by looking through various
period of time. databases, and then decisions are made
about what studies are to be
In cohort studies, the participants must included/excluded.
share a common factor or characteristic
such as age, demographic, or occupation. Strengths: Increases the conclusions’
A cohort study is a type of longitudinal validity as they’re based on a wider range.
study in which researchers monitor and Weaknesses: Research designs in studies
observe a chosen population over an can vary, so they are not truly comparable.
extended period. Peer Review
A researcher submits an article to a
Mixed methods means using more than journal. The choice of the journal may be
one research method to improve the determined by the journal’s audience or
study’s validity. prestige.

Reliability The journal selects two or more

Reliability is a measure of consistency, if a appropriate experts (psychologists
particular measurement is repeated and working in a similar field) to peer review
the same result is obtained then it is the article without payment. The peer
described as being reliable. reviewers assess: the methods and
designs used, originality of the findings,
Test-retest reliability: assessing the same the validity of the original research
person on two different occasions which findings and its content, structure and
shows the extent to which the test language.
produces the same answers.
Inter-observer reliability: the extent to Feedback from the reviewer determines
which there is an agreement between two whether the article is accepted. The
or more observers. article may be: Accepted as it is, accepted
with revisions, sent back to the author to
Meta-Analysis revise and re-submit or rejected without
Meta-analysis is a statistical procedure the possibility of submission.
used to combine and synthesize findings
from multiple independent studies to The editor makes the final decision
estimate the average effect size for a whether to accept or reject the research
particular research question. report based on the reviewers comments/
recommendations.
Meta-analysis goes beyond traditional
narrative reviews by using statistical Peer review is important because it
methods to integrate the results of several prevent faulty data from entering the
public domain, it provides a way of
checking the validity of findings and the 3. Primary data is first-hand data
quality of the methodology and is used to collected for the purpose of the
assess the research rating of university investigation.
departments. 4. Secondary data is information that has
been collected by someone other than
Peer reviews may be an ideal, whereas in the person who is conducting the
practice there are lots of problems. For research e.g. taken from journals,
example, it slows publication down and books or articles.
may prevent unusual, new work being
published. Some reviewers might use it as Validity
an opportunity to prevent competing Validity means how well a piece of
researchers from publishing work. research actually measures what it sets
out to, or how well it reflects the reality it
Some people doubt whether peer review claims to represent.
can really prevent the publication of
fraudulent research. Validity is whether the observed effect is
genuine and represents what is actually
The advent of the internet means that a out there in the world.
lot of research and academic comment is
being published without official peer 1. Concurrent validity is the extent to
reviews than before, though systems are which a psychological measure relates
evolving on the internet where everyone to an existing similar measure and
really has a chance to offer their opinions obtains close results. For example, a
and police the quality of research. new intelligence test compared to an
established test.
Types of Data 2. Face validity: does the test measure
1. Quantitative data is numerical data e.g. what it’s supposed to measure ‘on the
reaction time or number of mistakes. face of it’. This is done by ‘eyeballing’
It represents how much or how long, the measuring or by passing it to an
how many there are of something. A expert to check.
tally of behavioral categories and 3. Ecological validity is the extent to
closed questions in a questionnaire which findings from a research study
collect quantitative data. can be generalized to other settings /
2. Qualitative data is virtually any type of real life.
information that can be observed and 4. Temporal validity is the extent to
recorded that is not numerical in which findings from a research study
nature and can be in the form of can be generalized to other historical
written or verbal communication. times.
Open questions in questionnaires and
accounts from observational studies Features of Science
collect qualitative data. a. Paradigm – A set of shared
assumptions and agreed methods
within a scientific discipline.
b. Paradigm shift – The result of the A type I error is when the null hypothesis
scientific revolution: a significant is rejected when it should have been
change in the dominant unifying accepted (happens when a lenient
theory within a scientific discipline. significance level is used, an error of
c. Objectivity – When all sources of optimism).
personal bias are minimised so not to
distort or influence the research A type II error is when the null hypothesis
process. is accepted when it should have been
d. Empirical method – Scientific rejected (happens when a stringent
approaches that are based on the significance level is used, an error of
gathering of evidence through direct pessimism).
observation and experience.
e. Replicability – The extent to which Ethical Issues
scientific procedures and findings can Informed consent is when participants
be repeated by other researchers. are able to make an informed judgment
f. Falsifiability – The principle that a about whether to take part. It causes them
theory cannot be considered scientific to guess the aims of the study and change
unless it admits the possibility of their behavior.
being proved untrue. To deal with it, we can gain
presumptive consent or ask them to
Statistical Testing formally indicate their agreement to
A significant result is one where there is a participate but it may invalidate the
low probability that chance factors were purpose of the study and it is not
responsible for any observed difference, guaranteed that the participants
correlation, or association in the variables would understand.
tested. Deception should only be used when it is
approved by an ethics committee, as it
If our test is significant, we can reject our involves deliberately misleading or
null hypothesis and accept our alternative withholding information. Participants
hypothesis. should be fully debriefed after the study
but debriefing can’t turn the clock back.
If our test is not significant, we can accept
our null hypothesis and reject our All participants should be informed at
alternative hypothesis. A null hypothesis the beginning that they have the right
is a statement of no effect. to withdraw if they ever feel distressed
or uncomfortable.
In Psychology, we use p < 0.05 (as it
strikes a balance between making a type I It causes bias as the ones that stayed are
and II error) but p < 0.01 is used in tests obedient and some may not withdraw as
that could cause harm like introducing a they may have been given incentives or
new drug. feel like they’re spoiling the study.
Researchers can offer the right to
withdraw data after participation.
Participants should all have protection
from harm. The researcher should avoid
risks greater than those experienced in
everyday life and they should stop the
study if any harm is suspected. However,
the harm may not be apparent at the time
of the study.

Confidentiality concerns the

communication of personal information.

The researchers should not record any

names but use numbers or false names
though it may not be possible as it is
sometimes possible to work out who the
researchers were.