Research Methods

Research methods are how psychologists and scientists come up with and test their theories.
IV= I change
DV= Measure
Control= Stays the same
Extraneous= Effects DV, changes randomly.
To operationalise a variable is to clearly define and measure it in a way that allows researchers to
collect and analyze quantifiable data.

Confounding= changes systematically like extraneous.

Types of study include:

●​ Experiments
●​ Observation
●​ Self-reporting
●​ Case studies
Experimental design + method
These examine how variables affect outcomes. A variable is anything that changes between the two
situations.
-An example of an experiment is Bandura’s Bobo doll, it investigated how role model’s (variable) behaviour
affected how the child played

Experiments can be designed in 2 ways: Independent groups and repeated measures.

Independent groups:
1.​ Participants are divided into two groups
2.​ One group does the experiment with Variable 1, one group does the experiment with variable 2. This
means that none of the participants experience both variables.
3.​ Results are compared

Strength AO3:
No order effects are presented (when participants' responses in the various conditions are affected by
the order of conditions to which they were exposed) because they have not been exposed to any prior
conditions.

Participants are less likely to guess the aim of the experiment because they have only experienced one
variable, reducing demand characteristics.
Weakness AO3: Produces confounding variables (one that influences both Variable 1 and Variable 2)
because it is impossible to make sure that the two groups of people are exactly the same. As a result, it
is hard for the IV to be the only difference between the groups (aka participant variable, where the
people differ from each other). As a result, the internal validity (the extent to which the observed results
represent the truth in the population we are studying) is ruined.

(Participant variables are always extraneous variables but they're not always confounding variables. Some
participant variables just don't matter much. For example, in a memory experiment it probably doesn't
matter what sign of the Zodiac each participant is; it's not like Leos or Capricorns have better
memories. You don't need participants to be exactly identical, just similar in important respects.)

Also inconvenient to carry out because a large sample size is needed.

Solution: Randomly allocate participants so that everyone has an equal chance to be in each condition.

Repeated Measures
1.​ Participants are not divided into groups
2.​ All participants do the experiment with variable 1, then with variable 2
3.​ Results are compared

All participants do the experiment with variable 1, then with variable 2

Results are compared
Strength AO3: Uses fewer people because you’re not splitting them into groups and therefore, the only
difference is the IV because you are testing the same people twice aka fewer participant variables and
more convenient
Weakness AO3: Order effects present because the participants experience both variables.

Individuals may behave differently for the second variable due to boredom or fatigue (formally known as
fatigue effects) or because they feel they need to show demand characteristics.
-Related to above, if you use the same method of testing the variable each time the individuals will become
familiar with it and as a result the individuals will respond differently. E.g giving students a maths test
before and after break to examine the effect of relaxation on memory, the students may perform
better in the 2nd trial due to familiarity with the questions. (this is called practice effects).
Solution: Split participants into two. One group does the first variable and then the second, the second
group does the second variable and then the first.

Matched Pairs Design (another form of independent groups).

1.​ Participants are selected and matched with others having similar characteristics, this could be
through a test based on the variable e.g a maths test.
2.​ Provides two relevantly similar groups, controlling for differences that might skew results. For
example, you could group people together because based on their maths skills
3.​ Experiment conducted as normal independent groups design
To get around fatigue and practice effects, you could use randomisation (determining which condition a
participant experiences by random chance (tossing a coin, drawing a card) or counter-balancing (splitting
the group into sub-groups: one sub-group does the experimental condition first, then the control; the
other sub-group does it the other way round.)

No solution as this is a solution for the first two.

Laboratory experiments
An experiment conducted in an artificial, controlled environment
●​ E.g. Bandura’s Bobo the doll experiment or Asch’s conformity experiments
It tends to have high internal validity because the researchers can control all variables and therefore the
main difference between the control group and the experimental group is the independent variable being
studied

Strengths of Lab:
-Controlled Environment which minimises the risk of other variables skewing results, making causal
effects of variables clearer.
-Casual Clarity : Any changes in outcome must result from a change in the variable due to the tightly
controlled environment

Weakness of Lab:
-Reduce ecological validity: Results might not translate to real-life scenarios due to artificial prison
environment
-Demand characteristics: Participants may alter their behaviour because they know they are part of an
experiment, behaving in ways they think are expected rather than naturally
Field Experiment
An experiment carried out in a natural, real-world environment. Can include confederates. Participants are
not aware that they are being observed in an attempt to generate more realistic behaviour or responses
which can be generalised to real-world settings.
●​ E.g. Bickman’s study of the effects of uniforms on obedience

Strengths of Field
-More naturalistic than a field experiment because participants are not aware they are being studied and
therefore demand characteristics are not present also ecologically valid
-The independent variable is controlled

Weaknesses of Field
-Ethical considerations as participants are not aware that they are being studied and these types of
experiments are likely an invasion of privacy and lack of consent.
-Loss of control of extraneous variables because precise replication in each individual is not possible.

Natural experiments
Where the variable changes naturally and the researcher seizes the opportunity to study the effects,
and so the IV occurs naturally but temporarily e.g studying the effect a change in drug laws (variable)
has on addiction. The IV is not brought about by the researcher and it would have happened even if the
researcher was not there e.g studying reactions to earthquakes.

A notable example of a natural experiment is the study conducted by Charlton et al. (2000), which
investigated the impact of television. Before 1995, the residents of St. Helena, a small island in the Atlantic,
had no access to television. However, with its introduction, researchers had the opportunity to explore
how exposure to Western programming might affect their behavior. In this study, the independent variable
(IV) was the arrival of television, an event that researchers did not control and which would have been
challenging to manipulate. The dependent variable (DV) consisted of measures of pro-social and anti-social
behaviors, which were evaluated through questionnaires, observations, and psychological assessments.

Strengths of Natural:
-Provides opportunities for research that would have otherwise been unethical or unpractical to carry
out.
-High external validity as you are dealing with realistic scenarios

Weaknesses of Natural
-Naturally occurring events may be rare and so these experiments are hard to replicate, therefore hard
to generalise findings due to little evidence.
-Very difficult to place participants into groups due to confounding and extraneous variables being a
problem
Quasi
IV is not created by researcher but is rather naturally occurring and permanent e.g gender. Compares
two variables that can’t be changed.

Strengths of Quasi:
-Controlled conditions so highly replicable, leading to high internal validity

Weaknesses of Quasi:
-Very difficult to place participants into groups due to confounding and extraneous variables being a
problem

Observational method
The researcher will observe the participants interacting with each other or the variables depending on
the nature of the experiment, essentially examining the behaviour of the participants. This type of method
only looks at the dependent variable and therefore this is more of a technique rather than a method,
often used in conjunction with other methods.

This type of method only uses behavioural categories to examine which groups exert which behaviour the
most often, certain behaviours are prioritised e.g fear, anger, etc. Numerical or qualitative ratings can be
used to measure each behavioural category.
-Researchers will often try their best not to interfere with the behaviour being studied so that the
results are valid and realistic. Oftentimes, researchers will decide when and where to record behavioural
data through time and event sampling. Event sampling is counting how many times an individual carries out a
behaviour, whilst time sampling is counting what the participant does at regular intervals e.g making note
of their behaviour each time a minute passes.
AO3:
Inter-observer reliability is required as different researchers may note down behaviour in different
ways. Therefore, ratings must be clearly defined and explained aka behaviour which consists of a ‘3’
rating must be explained why it is a ‘3’ and not a ‘4’ or ‘2’ This avoids subjectiveness of the rating, which is
often an issue with this method.

Observational design
Naturalistic observation
Watching and recording behaviour where it usually takes place eg setting up cameras in an office or
school to observe how people interact in those environments

Strength AO3:
High ecological validity as the behaviour is in its natural environment without other variables.
High external validity as the experiment is done in a natural environment.

Weakness AO3:
Participants may feel like they are being watched, and so this technique may lack ecological validity if this
occurs.
Can be difficult to replicate
Uncontrolled confounding and extraneous variables present.

Controlled observation
Watching and recording behaviour in a controlled and artificial environment like a laboratory.

Strength AO3:
Easy to replicate
More control over extraneous and confounding variables due to ability to manipulate every aspect of the
experiment and environment.
Researcher can focus on one particular perspective of the behaviour rather than multiple because you
can isolate a variable through the lab settings
Weakness AO3:
Low ecological validity and practical applications.
Demand characteristics present as the participants are likely to figure out both the aim and the idea of
the experiment.
More likely to observe unnatural behaviour as the experiment itself is taking place in an unnatural setting.

Overt
Participants are aware that they are taking part in a study.
Strength AO3:
Highly ethical as consent is given prior to study.
Weakness AO3:
Less valid as demand characteristics are often present.
May record unnatural behaviour as the study takes place in an unnatural environment

Covert
Participants are not aware that they are taking part in a study.
Strength AO3:
Natural behaviour recorded as less demand characteristics are present, high internal validity of results.
Participants are less likely to try and guess the aim of the study or to make sense of the situation they
are in (participant reactivity).
Weakness AO3:
Ethical issues as informed consent is not given, additionally may be breaching privacy

Participant
Researcher is part of the group being observed for the study.
Strength AO3
Can be more insightful for the researcher
Weakness AO3
Researchers may identify too strongly with the participants and therefore lose track of their objective
which is to observe behaviour.
Participants may still find out that they are being watched

Non-participants
Researcher observes from a distance and so is not a part of the group being observed.
Strength AO3
Researchers are less likely to identify with the participants and therefore may obtain more valid results.
Weakness AO3
Researcher bias may be present as the results rely on the researcher, e.g stereotyping.
Researchers may lose valuable insight as they are observing from the external environment rather than
being involved in the study.

Self report techniques

In a self report, participants provide information about themselves. There are two techniques for this,
interview and questionnaire.
A questionnaire is a standardized list of questions which participants answer at the end of the study e.g
Hazan and Shaver. Open questionnaires are where responses are not fixed e.g yes or no, but rather the
questions are open to more answers. Closed questionnaires have closed yes or no questions.
AO3:
+​ They are quantifiable e.g when using closed questions you can quantify these to get direct data.
+​ Because questions are standard, they can be easily replicated for other studies. This also allows
for peer-review in order to strengthen certainty of findings.
-​ May be biased in some cases where the questionnaires are handed out randomly. As a result,
individuals with more free time may answer them and so the data collected may not be generalizable.
-​ Participants may lie for taboo questions
-​ Individuals may interpret each question differently and so a lack of understanding if individuals are
not briefed.
-​ Cannot get enough detail compared to interviews

Interviews are where participants are asked questions in person e.g Bowlby’s thieving children study.
Structured interviews are where the interviewer has a set of questions which are pre-planned and are
asked in the same order. Unstructured interviews are where the interviewer has a discussion with the
participants in regards to the study. There are also semi-structured interviews which are a mix between
the two.
AO3:
+​ More detail than questionnaire, mostly unstructured ones
+​ Structured interviews are easily replicated because the questions are all the same.
-​ Lack of quantifiable data so the information given is left to the interviewer to condense so hard
to compare (even with unstructured despite the fact they give more data).
-​ Interviewer’s appearance or tone may affect the answers of the participants.

Case studies
These are detailed investigations into an individual, group of people or event. Many researchers use
interviews, questionnaires and observational techniques because these methods provide qualitative data and
a case history of the information is made.

Most case studies focus on unusual individuals vs typical individuals. Most studies take place over a longer
period of time (longitudinal) where researchers check in with the subjects at specific times (time
sampling).
+​ Provides detailed qualitative data because they don’t focus on only one point of time
+​ Allows for investigation into things that are unethical as you are essentially waiting for them to
happen rather than making them take place e.g removing a toddler’s left hemisphere compared to waiting
until it is medically necessary.
-​ May not be replicated as easily and so lack validity in this aspect
-​ Researcher bias likely present due to small sample size so researchers often need to place their
own subjective interpretations as they don’t have enough evidence to draw an accurate conclusion
 Scientific processes
Aim: What the investigation concerns and why it's being investigated.

Hypothesis: A prediction of what will happen. An experimental hypothesis is a prediction that changing IV will
have a certain effect on DV. A null hypothesis is a prediction that IV will have no effect on DV.

Sampling: A part of the population that is representative of the entire group.

-Random sampling is where participants are selected from the target population at random and so each
member has a chance of being picked. This type of sampling is good because it is unbiased and
representative because it is likely that the sample will be representative of the population as a whole.
However this type of sampling is impractical because it's often difficult to include all members of a
target population for selection e.g doing a study on women or a certain racial group
-Systematic sampling involves selecting participants at preset intervals e.g selecting every 50th person
from a list or other intervals. This type of sampling is unbiased and also representative however in a
more logical way as it is independent of any particular characteristic (apart from that of the individual
order in the list). However some characteristics may be present at certain intervals and so bias is not
entirely eradicated.
-Stratified sampling is where you divide the population into groups, working out how much of the
population has a certain characteristic within the group and you arrange random sampling to select a
certain number of characteristics from each group. (For example, let’s say 20% of the population is aged
0-18, and 50% of the population is aged 19-65, and 30% of the population is aged >65. A stratified sample
of 100 participants would randomly select 20x 0-18 year olds, 50x 19-65 year olds, and 30x people over
65). This type of sampling is designed to be unbiased and representative of everyone however it requires
knowledge of the population breakdown e.g how many people have each characteristic. If population
breakdown is not known, then the sample will be biased.
-Opportunity sampling is where you approach participants and ask them to take part in the study e.g
asking them in the street. Volunteer sampling is participants offering to take part in the study. Both
opportunities are quick and easy to carry out, and with opportunity sampling it may be the only option
for natural experiments. However, both options are unrepresentative.

Variables: You alter the independent variable to see the effect on the dependent one. Extraneous variables
are ones which are not the DV but affect the IV before the experiment. Confounding variables are
extraneous variables which are not controlled in the experiment.

Operationalisation of variables: Where you clearly define and specify the variables in detail e.g through
measurements. If variables are not operationalised, the study cannot be replicated and the findings may
not be valid.

Pilot studies are practice runs of proposed research projects before they happen. The purpose of
these studies is to identify problems or areas for improvement, it may also show whether the results
are going to be statistically significant (e.g whether IV is affecting DV properly or its just chance).
Study reporting

Psychological reports contain:

1)​ Title which is a short description of the research
2)​ Abstract which summarises the hypothesis, methods, results and conclusions
3)​ Introduction which is a broad overview of the context (aka funnel technique) before focusing on
why the study was carried out, its aims and the hypothesis of the researcher.
4)​ The method is a detailed description of how the study will be carried out. The reason for the
method is so that the research can be evaluated and replicated. It may include subsections like study
design (observational or experimental method, independent groups or repeated measures, identification and
operationalisation of variables), participants (description of target population, sampling method and how
many participants took part), equipment used, standardised procedure (step-by-step instructions),
controls (how extraneous variables were controlled).
5)​ Results is a presentation of the key findings of the research, aka a written summary of raw
data. The raw data is often included in the appendices.
6)​ Discussion is what the results mean and how they relate to the hypothesis e.g do they prove it or
not. any issues with how results were generated, how the results fit with other research, and
suggestions for future research
7)​ Conclusion: Key findings
8)​ References - Which things prove your writing, written throughout the research.
9)​ Appendices - Supporting materials which are too long to include in study e.g raw data, complete list
of questions from questionnaires.

Peer-review is where you assess the research before publishing it in a journal, the purpose of this is to
prevent false ideas and bad research.
The researchers submit their paper to the journal they want it to be published in, and the editor of that
journal sends the paper to expert reviewers (i.e. psychologists who are experts in that area – the
researchers’ ‘peers’) who evaluate the paper’s scientific validity. The reviewers may accept the paper as
it is, accept it with a few changes, reject it and suggest revisions and resubmission at a later date, or
reject it completely.
There are several different methods of peer review:

Types of peer-review:
Open-blind - Researcher and reviewer know each other.
Single-blind - Researchers don’t know reviewers. prevents the researchers from being able to influence
the reviewer. This is the most common form of peer review.
Double-blind- Neither researcher nor reviewer know each other. This additionally prevents the reviewer’s
bias towards the researcher from influencing their decision whether to accept their paper or not.

AO3 of peer-review:
-​ Bias as much academic research takes place with a small group of people so they all know each
other. Many academics are funded by organisations which prefer certain ideas over others, so even if
the researcher is acceptable it may be rejected.
-​ Peer-review does not prevent fraudulent or bad research from being published
-​ Prevents progress of new ideas as reviewers are often older individuals so they conform to old
ideas
-​ Reviewers may steal good ideas for themselves in single blind and double blind as the researcher is
anonymous
-​ Can take a long time to carry out

Study evaluation
Ethical issues question whether the study is morally right or not. The British Psychological Association has
set out a set of guidelines for the ethics of studies.
-​ Valid consent is where participants are told the study’s aims, the data being collected and any risks
associated with the study. Participants have the right to withdraw consent at any time. Researchers are
allowed to use initiatives (e.g money) to encourage participation but they can’t be influential enough that
they would compromise the participant’s freedom of choice. A researcher must also consider a
participant’s ability to consent e.g their age, mental ability.
-​ No deception is where researchers must avoid misleading participants, some degree of deception
may be needed to reduce risk of demand characteristics. Ways to limit this are: prior (general) consent
is where participants are aware they will be deceived without knowing the nature of the stud which may
affect behaviour as they try to figure out point of study, retrospective consent is where participants
are informed they were deceived after the study took place and then asked for consent but if they
don’t consent its useless, presumptive consent is where you ask people not involved in the study if they
would be willing to take part and then assume if these people want to then so must the others.
-​ Confidentiality is where personal data obtained about participants should not be disclosed unless the
participants agreed in advance, even so any data published cannot be used to identify the participant.
-​ Debriefing is where the researchers explain relevant details of the study to participants once the
data gathering is complete. If a study harmed the individual e.g inducing a negative mood it is ethical for the
debrief to address this harm so the participant doesn’t leave in a worse state than they came in

Studies must also be reliable, this is where the results can be consistently replicated under the same
circumstances. If the results are inconsistent it is unreliable, but if they are reliable it doesn’t make them
valid. Ways to assess reliability:
-​ Test-retest is when you give the same test to the same person on two different occasions, if
results are the same or similar this shows reliability.
-​ Inter-observer reliability is testing the reliability of observational studies, it is where you assess the
rating of observers in order to make sure they are all alike. It can be assessed mathematically by looking
for correlation within scores and improved by setting clearly defined behavioural categories.

The results of a study are valid if they accurately measure what they are supposed to. Ways to assess
validity include:
-​ Concurrent validity is where the test results correlate with that of a similar test. E.g assessing
intelligence levels with your own test, then using an IQ test to show concurrent validity.
-​ Face is where the study looks like it measures what it's supposed to measure at face value. E.g
measuring intelligence by asking their favourite meal would not show face validity, but using an IQ test
would.
-​ Ecological validity is the extent to which a study’s findings apply outside of the environment it was
conducted in.
-​ Temporal validity is where the study’s results stay true over time

Ways to control extraneous variables:

Demand characteristics (extraneous variables which take place when the demands of a study cause
behaviour to change in a way it wouldn’t outside the study) reduce ecological validity.
For example, if a participant guesses the purpose of an experiment they are taking part in, they may try
to please the researcher by behaving in the ‘right’ way rather than the way they would naturally.
Alternatively, the participant might rebel against the study and deliberately try to sabotage it (e.g. by
deliberately giving wrong answers).
Researchers can control this by using single-blind trials. For example, a drug trial could give half the
participants the actual drug and the other half a placebo but not tell participants which treatment they
received. This way, both groups will have equal demand characteristics and so any differences between
them should be down to the drug itself.

Investigator effects are where the characteristics of the researcher affect participant behaviour,
which also reduces ecological validity. One way to reduce this is using double-blind methods as the
researcher is unable to give any clues, conscious or not, to participants that would affect behaviour.
Many characteristics – e.g. the researcher’s age, gender, accent, what they’re wearing – could potentially
influence the participant’s responses. For example, in an interview about sex, females may feel less
comfortable answering questions asked by a male interviewer and thus give different answers than if
they were asked by a female. The researcher’s biases may also come across in their body language or
tone of voice, affecting the participant’s responses.

Participant variables are differences between participants. To control this, use random allocation.
For example, in an experiment on the effect of caffeine on reaction times, participants would be
randomly allocated into either the caffeine group or the non-caffeine group. A non-random allocation
method, such as allocating caffeine to men and placebo to women, could mean variables in the allocation
method (in this case gender) skew the results. When participants are randomly allocated, any extraneous
variables (e.g. gender in this case) will be allocated evenly between each group and so not skew the
results of one group more than the other.

Situational variables are effects caused by the environment in which the experiment is taking place. To
control this, use standardisation.
For example, all the tests of caffeine on reaction times would be conducted in the same room, at the
same time of day, using the same equipment, and so on to prevent these features of the environment
from skewing the results.

Task order only takes place in repeated measures, and to control this researchers can use
counterbalancing in order to be in charge of order of completion of tasks. For example, half of
participants would do task A followed by task B, and the other half would do task B followed by task A.

Implications of Psychological Research for the Economy

Features of science:
Observations made from a neutral perspective without bias rather than the scientist’s subjective
viewpoint.
A tape measure provides a more objective measurement of something compared to a researcher’s
guess. Similarly, a set of scales is a more objective way of determining which of two objects is heavier
than a researcher lifting each up and giving their opinion

Replicability is where scientific procedures and experiments are repeated to double check the findings
are valid.

Falsifiability is where a hypothesis is tested and there must be some possible observation to prove it
false.
E.g. The hypothesis that “water boils at 100°c” could be falsified by an experiment where you heated water
to 999°c and it didn’t boil. In contrast, “everything doubles in size every 10 seconds” could not be falsified
by any experiment because whatever equipment you used to measure everything would also double in size.
Freud’s psychodynamic theories are often criticised for being unfalsifiable: There’s not really any
observations that could disprove them because every possible behaviour (e.g. crying or not crying) could
be explained as the result of some unconscious thought process.
Philosopher Thomas Kuhn argues that science is not as unbiased and objective as it seems. Instead, the
majority of scientists just accept the existing scientific theories (i.e. the existing paradigm) as true and
then find data that supports these theories while ignoring/rejecting data that refutes them.
Rarely, though, minority voices are able to successfully challenge the existing paradigm and replace it with
a new one. When this happens it is a paradigm shift. An example of a paradigm shift in science is that
from Newtonian gravity to Einstein’s theory of general relativity. Paradigm shifts are essentially social
cryptoamnesia.

Data handling and analysis

Quantitative data is numerical. It is good because it allows for mathematical and scientific comparisons,
however is less detailed and so may miss out important details.

Qualitative data is non-numerical. It cannot be compared as easily but is more detailed and allows for
important facts to be included.

Content analysis is a way to analyse qualitative data due to its difficulty to be analysed. Here, a sample of
qualitative data is collected before relevant coding units are identified and operationalised. Data is then
analysed according to the coding units in order to produce quantitative metrics, and statistical analysis is
carried out after.
-For example,a bunch of unstructured interviews on the topic of childhood are collected, coding units
could be discussion of traumatic events, happy memories, births and deaths. Researchers listen to the
unstructured interviews and count how often these events are mentioned, statistical analysis carried out.

Primary data is the original data collected for the study. Secondary data is data from another study
previously conducted.

A meta-analysis is a study of studies. It involves searching for several smaller studies within a certain
topic before using statistics to identify similarities and trends for a larger study. A good meta-analysis is
often more reliable than a regular study as it is based on a data set, and any issues with one singular
study will be balanced out by the others.

Descriptive statistics
Mean, median and mode are measures of central tendency. This refers to ways of reducing large data
sets into averages.

For the mean, add all the numbers together and divide by the total quantity of numbers.
-​ Good because it uses all data in a set and is accurate as the number is precise. Bad because it
can be skewed by one or two high or low abnormal scores.
For the median, arrange the numbers from smallest to biggest and find the number in the middle. If two
numbers are the median, add them and divide by two to find the midpoint
-​ Good because won’t be skewed by freak scores. Bad because it isn’t representative of central
tendency and is less accurate than the mean.
Mode is found by finding the most common number in a set.
-​ Good because won’t be skewed by freak scores, and makes more sense when aiming to present
central tendency in data sets. Weak because it doesn’t use all the data and more than one mode may be
present.

Range and standard deviation are measures of dispersion. They quantify the extent to which scores in a
data set vary.
Range is found by subtracting the smallest value from the largest.
-​ Good because it is easy to carry out and accounts for the freak scores. Bad because it can be
skewed by freak scores (The difference between the biggest and smallest numbers can be skewed by
a single anomalous result or error, which may give an exaggerated impression of the data distribution
compared to standard deviation) and does not account for how common scores are.
Standard deviation is a measure of how much numbers in a data set deviate from the mean. To find the
standard deviation: find mean, subtract mean from each number in the set, square these numbers and
then add them together. Divide the result by the amount of numbers (for a sample, do n-1), square root it.
This is your standard deviation.
-​ Good because it is less skewed by freak scores as you measure average difference from mean.
Bad because it takes longer to calculate than the range.
Percentage change is finding the difference between the original and new number, dividing the difference
by the original and multiplying by 100.

Normal distribution is where the majority of the scores are on or near the mean. It is also symmetrical
(equal number of scores above and below). In normal distribution, scores become rarer the more they
deviate from the mean.
When on a histogram, the data will be like this:

Skewed distribution is where it isn’t symmetrical, and the scores are not
distributed on either side. This is caused by outliers which throw off the
mean. A positive skew is where an incredibly high score makes the mean
much higher than the rest of the scores and so most scores are below
the median. Negative skew is where an incredibly low score makes the mean
lower than most scores so most are above the mean.

Correlation is how closely related two or more things are. You can find this
from a scattergram. If R is +1 it means that two things are positively
correlated, when one goes up so does the other by the same amount (positive correlation). If R = -1 then
they are negatively correlated, when one goes up the other goes down by the same amount. If R=0 then
the two are not at all correlated, any changes are independent of each other.

Presentation of data
Tables can present raw data in data tables or summarised results in results tables.

Scattergrams illustrate two variables at various data points.

Bar charts illustrate discrete (distinct and countable with no values in between) data categories for
comparison. The x-axis will list the independent (category) and the y the dependent (results between
categories).

Histograms illustrate continuous or interval data. The x axis shows the continuous data whilst the y axis
shows the frequency of scores in each category. No gaps between bars.

Line graphs (sometimes called frequency polygons) illustrate continuous data but possibly for more data
categories.

Pie charts illustrate how commonly different things occur relative to each other, and show a visual
representation of percentages.

Inferential testing
The point of inferential testing is to see whether data is statistically significant i.e if the change is due to
a known variable or if it occurs randomly.
-​ Probability is represented by p. A lower p shows more statistical significance, but you can
never get a value of p, there’s usually a threshold of <0.05, p is lower than 0.05. P
measures how random it is, so if p=0.06 then it is 60% due to random chance, not
statistically significant.
-​ Insignificant is due to null hypothesis. Significance is due to alternate hypotheses.

A type 1 error is a false positive. It is where the researchers conclude an effect is due to a real
variable when it is not.

-The null hypothesis states that there will be no effect of the independent variable on the dependent one.

A type 2 error is a false negative, it is where researchers conclude that an effect is due to a random
variable when it is not.

One-tailed means you’ll have either positive or negative. Two-tailed is where you only hypothesise there will
be an effect.
Related experimental design is where the same participants are used in multiple conditions (repeated
measures). Unrelated is where different participants are used for different conditions (i.e. independent
groups).

Types of data

Nominal: involves categories. For example types of sport, types of behaviour.

Ordinal: Ordered ranks of things but not necessarily precise because you’ve chosen these ranks. For
example, ranking mood scale from 1-10
Interval: Uses standardised measurements. For example, kg, time, temperature.

When testing differences, there are different tests for related and unrelated experiment designs. There
are also tests for correlation. Correlation is the relationship between variables, difference is the
comparison between variables.

Sign test: