Research Methods Flashcards
cause and effect
psychologists use the experimental method to identify the “effect” one variable has on another variable, in other words does one variable ‘‘cause’’ the other variable to change
being able to conduct research that establishes cause and effect is a key feature of scientific research
In a true experiment, there must be …
a control condition and an experimental condition and participants must be randomly assigned to these conditions
this is so that the researcher can make fair comparisons between the two groups
IV DV controlled
a researcher manipulates the independent variable in order to test its effect on the dependent variable
everything else is kept the same (controlled) between the two conditions
if there is a significant difference in the results of the two groups …
we can conclude that the independent variable caused the change in the dependent variable (cause and effect)
IV
independent variables
a variable that is manipulated by the researcher - or changes naturally
DV
dependent variable
the variable that is measured by the researcher. any effect on the DV should be caused by changes in the IV
in order to test the effect of the IV …
we need different experimental conditions
control condition
provides a baseline measure of behaviour without experimental treatment
experimental condition
the one in which there has been researcher manipulation. this is the condition in which the researcher is particularly keen to see if a difference in behaviour has occurred
operationalism
clearly defining variables so they can be measured
aim
general statement of what the researcher intends to investigate, essentially the purpose of the study
hypothesis
a testable statement predicting the outcome of the study which is made at the start of the study
what are the two different experimental hypothesis
directional (one tailed)
non-directional (two tailed)
directional hypothesis
makes it clear what sort of difference or relationship is between the 2 conditions mainly using words like ‘less’ ‘lower’ ‘more’ ‘higher’
what are the two different kinds of hypothesis and explain them
null hypothesis - this predicts that there will be no difference or relationship
alternate/experimental hypotheses - these predict a difference or relationship and can be directional or non-directional
non-directional hypothesis
used when there has been no previous research to suggest what direction the research will go in or the previous findings have been contradictory
‘will be a difference’ or ‘there will be a relationship’ between the 2 conditions but the there is no direction of the outcome
types of experiments
laboratory experiments
quasi experiments
field experiments
natural experiments
lab experiments
conducted in highly controlled environments
researcher manipulates the IV and records the effects of the DV
participants in a lab experiment can be randomly allocated to conditions
a lab experiment is therefore considered a ‘true’ experiment
strengths of lab experiments
high control over extraneous variables meaning cause and effect can be established
replication is possible due to the high level of control. This also means results can be checked for reliability
limitations of lab experiments
participants are often aware of being tested - possible demand characteristics
artificial environment means it may lack generalisability
investigator effect may occur (unless it is a double blind experiment - this is when both the participants and the researcher conducting the experiment does not know the aims of the investigation)
field experiments
researcher still manipulates the IV and records the effects on the DV but the experiment is conducted in a real life setting
strengths of field experiment
high ecological validity due to being conducted in a real-life setting
behaviour is likely to be more valid & authentic (less demand characteristics)
limitations of field experiments
there is less control over extraneous variables (these will be different depending on the experiment)
it is difficult to replicate them completely because they tend to be less controlled
possible ethical issues if participants are unaware they’re being studied
standardisation
this is using exactly the same procedure for all participants, such as the same environment, instructions and experience
Confounding variable
any variable, other than the IV, that has not been controlled does affect the DV. Therefore we cannot be sure of the true reason for the changes to the DV / differences found
natural experiments
the researcher takes advantage of a naturally occurring IV and the effect it has on the DV
the experimenter has no manipulated the IV directly; the IV would vary naturally whether or not the researcher was interested
the researcher cannot randomly allocate participants to conditions and/or has no control over the IV
this is not a ‘true’ experiment
field experiment: it is the IV that is …
natural, not necessarily the setting
strengths of natural experiment
provides opportunities for research that may not be otherwise conducted due to practical/ethical reasons e.g. does smoking when pregnant lead to behavioural problems in infants
they have high external validity because they involve the study of real-life
limitations of natural experiments
a naturally occurring event may happen, rarely limiting generalisation to other situations
participants may not be randomly, allocated to conditions
Quasi experiments
IV is not something that varies at all - it is simply a difference between people that exists
the researcher records the effects of this ‘quasi-IV’ on the DV
the researcher cannot randomly allocate participants to conditions and/or has no control over the IV
This is not a ‘true’ experiment
example, experiments where the IV is a variable such as age, gender
strengths of quasi experiment
carried out under controlled conditions & share the strengths of a lab experiment
limitations of quasi experiments
participants are aware of being tested - possible demand characteristics
participants cannot be randomly allocated and therefore they may be confounding variables
type of experiment > environment conducted in > independent variable
lab > controlled > controlled
field > natural > controlled
quasi > controlled > naturally occurring
natural > natural > naturally occurring
extraneous variables
any variable other than the IV may have an effect on the DV if it is not controlled
these variables can come from the participant (e.g. age, intelligence), the experimental situation (e.g. noise levels, temperature) or the experimenter (e.g. personality, appearance or conduct of the researcher)
demand characteristics
participants are not ‘passive’ in experiments and they may work out what is going on and change their behaviour to please the experimenter or even act negatively
demand characteristics occur when a participant may receive a ‘cue’ from a researcher or the situation so the participant changes their behaviour as a result
investigator effects
any effects of the investigator’s behaviour (conscious or unconscious) on the research outcome (DV) , this may include everything from the design of the study, to the selection of and interaction with the participants during the research process
ways to minimise extraneous / confounding variables
randomisation
standardisation
variables that need to be controlled
extraneous variables
confounding variables
demand characteristics
investigator effects
randomisation
the use of ‘chance’ in order to control for the effects of bias i.e. in a memory experiment that may involve participants recalling words from a list. the order of the list should be randomly generated so that the position of each word is not decided by the experimenter
Reliability and validity
they can affect the credibility of research findings. psychologists must consider these when designing and conducting research and they are used to assess how good a piece of research is
validity (2 defintions)
how accurate and representative the results are
the degree to which something measures what it claims to
what are the two types of validity
internal validity
external validity
internal validity
concerns whether the results are due to the manipulation of the IV and not affected by confounding variables
external validity
refers to the extent to which the results can be generalised to other settings
what are the two types of external validity
temporal validity
ecological validity
temporal validity
refers to how well we can generalise the results across different periods of time
ecological validity
refers to whether the experimental results can be generalised to other settings, particularly from artificial/controlled settings to real life environment
reliability
how consistent the results are
if the experiment is repeated, will the same or highly similar results occur again?
what are the three different types of reliability
internal reliability
external reliability
inter-observer reliability
internal reliability
refers to the extent to which a test is consistent within itself. for example, if someone was completing a questionnaire measuring high levels of obedience they should have the same score on the questionnaire for it to have internal reliability
external reliability
refers to the extent to which a test is consistent over time. for example if someone achieved 120 on an IQ test if they were tested again in say 8 months time, we would expect them to achieve the same results. this would show the test to have external reliability
inter-observer reliability
refers to the extent to which two or more observers are in agreement on the behaviours they observe. we check for inter-rater reliability by correlating the two or more sets of observations to see if they correlate positively. over or at 0.8 would be high inter-observer reliability and they have consistently observed the same behaviours
experimental design
how the participants in an experiment will be used
three ways a researcher can arrange his/her participants
independent groups
repeated measures
matched pairs
independent groups
when two separate groups of participants experience two different conditions of the experiment
strengths of independent groups
order effects are avoided (when participants become aware of or bored with an experimental procedure)
there are less likely to be demand characteristics because participants only take part in one condition of the experiment and are therefore less likely to pick up on cues
limitations of independent groups
individual differences between groups, otherwise called ‘‘participant variables’’, may affect the results ( what if one group has people who have a naturally higher IQ than people in the other group?) - to deal with this random allocation is used
a larger amount of participants are needed in this experimental design
repeated groups
where all participants take part in both the conditions
strengths of the repeated group
participants variable problems are avoided because all participants take part in both conditions. therefore, it doesn’t matter if they have different IQs or memory abilities because they are kept constant through both conditions
this experimental design requires fewer participants because the same group is re-used
limitations of repeated groups
order effects are very likely to occur; participants may become bored, aware of the aims or tired because they carry out a task twice. they would need to control for this by using counterbalancing
demand characteristics are more likely to occur because participants have been exposed to both conditions of the experiment and therefore may pick on cues or figure out the aim of the experiment
the researcher will need to ensure they have different test materials for condition 1 and 2. for example, they would not be able to use the same list of words in a memory test for both conditions. to control for this they have to use a different set of words but make sure they are of similar difficulty
counterbalancing
am attempt to control order effects in which half the participants take part in condition A then B and the other half take part in condition B then A
For example
Participant 1 = A - B
Participant 2 = B - A
Participant 3 = A - B and so on …
counterbalancing does not remove or prevent …
order effects but attempts to balance out the effects of order between the two conditions
matched pairs
where pairs or participants are first matched on a key variable (i.e. IQ). then one member is assigned to condition A and the other assigned to condition B
strengths of matched pairs
the issue of participant variables is greatly reduced
order effects are totally avoided
demand characteristics are less likely
limitations of matched pairs
it is pretty much impossible to match people exactly on every characteristics unless maybe they are identical twins - and even then, it is usually just matching physical characteristics
it is very time consuming to find lots of people that match each other so closely
example of using matched pairs design in psychological research
bandura et al
bobo doll
in order to control for naturally occurring aggression levels in the children (so it would not confound the DV) he got the children’s parents and teachers to rate their aggression on a 1-5 scale
he then matched the children on their aggression levels so each condition had the same number of highly aggressive children (5), medium aggression (4-2) and non aggressive children (1)
two types of self-report techniques
interviews
questionnaires
structured
structured interviews are made up of pre-determined questions that are asked in a fixed order
this is like a questionnaire but conducted face-to-face (or over the phone) in real time
semi-structured
many interviews are likely to fall somewhere between structured and unstructured
there is a list of questions prepared in advance, but interviewers can follow up answers (like a job interview)
unstructured
an unstructured interview is a lot like a conversation. there is no set questions, but there is an aim that a certain topic will be discussed
the interview will be free-flowing. the interviewee is encouraged to expand on their answers
strength of structured interviews
easy to replicate due to standardised format (increases reliability)
limitation of structured interviews
a problem is that it’s difficult to deviate from the topic or for interviewees to expand on their answers (lacks depth and therefore validity)
strength of unstructured interviews
much more flexible, an interviewer can follow up on points if and when they arise gaining more insight and understanding (increasing validity)
limitations of unstructured interviews
trying to analyse the data can be challenging often because open ended question are used (qualitative data)
there is always the risk of interviewees being untruthful for reasons of social desirability
three different types of interviews
structured
semi-structured
unstructured
questionaires
involve a pre-set list of questions (or items) to which the participants responds through written answers. these are used to assess a person’s thoughts and/or experiences
a questionnaire may be used as part of an experiment to measure the DV. there are different styles of questions that can be designed : open and closed questions
what are the two different styles of questions
open and closed questions
strengths of questionnaires
can be given to a large sample of people and so large amounts of data can be gathered relatively easily. they can also be done without the researcher being present i.e. postal questionnaires
limitations of questionaires
respondents wanting to show themselves in a positive light (social desirability) rather than beinf truthful. or respondents may show ‘response bias’ where they respond in a particular way i.e. always ticking ‘yes’ or answering ‘3’ on a scale of 5
open questions + example
open questions do not have a fixed range of answers and respondents are free to answer in any way that they wish.
open questions tend to produce qualitative data (rich in depth, but difficult to analyse)
for example “Why do you enjoy psychology A level course?”
closed questions + example
closed questions offer a fixed number of responses and produce numerical data by limiting answers respondents can give. they produce qualitative data (easy to analyse, but lack depth associated with open questions)
For example “Do you watch more than 10 hours per week of TV?’… ‘yes’ or ‘no’
or respondents may be asked to rate how often they watch soap opera on TV on a scale of 1 - 5
1 2 3 4 5
1 = never
3 = sometimes
5 = everyday
evaluation of open questions
repondents can expand on their answers, which increases the amount of detailed information collected
open questions can reveal unexpected answers; therefore researchers can gain new insight into people’s feelings and attitudes
they also provide qualitative (non-numerical data) which although may be rich in information, it can be more difficult to summarise and/or detect patterns to draw conclusions
evaluation of closed questions
they have a limited range of answers and produce quantitative data (numerical data). this means the answers are easier to analyse using descriptive statistics (mean, mode, graphical representation)
however, respondents may be forced to select answers that don’t represent their true thoughts or behaviour, therefore data collected may lack validity
what to avoid when designing questionnaires and interviews
overuse of jargon
emotive language and leading questions
double barrelled questions and double negatives
overuse of jargon
technical terms that only those familiar with the field will understand
emotive language and leading questions
guiding the respondent to a particular response
double barrelled questions and double negatives
two questions in one; respondent may agree with one half of the question but not the other and therefore would not know how to respond
the question could be hard to decipher and could be written in a much clearer way
interviews are also a …
self-report method
interviews are more likely to collect …
qualitative data than questionnaires, but certain types of interview will lead to quantitative data being gathered
a good interview will involve …
an interview schedule
recording
effective interviewer
no ethical issues
an interview schedule
a list of questions the interviewer intends to cover. this should be standardised for each interviewee to reduce investigator bias
recording
the interviewer may take notes throughout the interview (although this may interfere with listening skills). Alternatively, the interview may be audio recorded or videoed
effect of interviewer
one of the strengths of interviews over questionnaires is that the presence of the interviewer who is interested in the interviewee may increase the amount of information provided. this is because it allows a rapport to be built with interviewee
the interviewer needs to be careful with their non-verbal communication - not sitting with arms folded for example. Behaviour needs to be welcoming and encouraging i.e. head nodding & leaning forward. A further consideration for the interviewer is listening skills - an experienced interviewer will know when and how to speak i.e. not interrupting or using negative language
interview ethical issues
respondents should be reminded that their answers will be treated confidentially. this is especially important if the interview includes topics that may be personal or sensitive
questionnaires are a …
self-report method
they are usually used to produce quantitative data for statistical analysis, but can also be used to collect qualitative data
features of a good questionnaire
clarity - clear questions that are easy to understand for respondent (reader)
bias - questions do not lead respondents to give a particular answer
assumptions - avoid making assumptions about respondents e.g. about sexuality
non-intrusive - avoids questions that are too personal
checked - questionnaire is piloted to make sure questions are understood and interpreted correctly
non-experimental methods: observational methods
naturalistic and controlled observations
overt and covert observations
participant and non-participant observations
naturalistic observation
takes place in the participant’s natural environment. for example it would not make sense to study how employees and managers from Primark behave by dragging the workforce into an artificial lab setting. it would be better to study their ‘interaction’ in their normal working environment. this means the researcher does not interfere in any way with what’s happening
don’t confuse a naturalistic observation with a natural experiment
they are different
in a natural experiment there is an IV whereas in an observation there isn’t
strengths of naturalistic observation
naturalistic observations provide a realistic picture of behaviour therefore have a high external validity (findings can be generalised to everyday life). Although this may be less so if participants are aware of being observed.
weakness of naturalistic observations
one of the issues is due to the lack of control there may be uncontrolled extraneous variables that may actually influence the behaviour observed
also, naturalistic observations tend to be one off situations and makes replication of investigation challenging
controlled observation
takes place in a controlled environment provided by the researcher e.g. the strange situation
in this set up the researcher can at least control for some variables but it does reduce the ‘naturalness’ of the environment and behaviour being studied
strengths of controlled observation
means the researcher can focus on particular aspects of behaviour and also being controlled means extraneous variables are less of a problem and replication becomes easier
weaknesses of controlled observation
making an environment more controlled can sometimes impact on how the participants behave. this may be less natural because of the environment
overt observation
in both naturalistic and observations participants may be aware they are being observed, this is called an overt observation
since this is likely to have an effect on the ‘naturalness’ of the participants behaviour, observers try to be as obtrusive as possible. the participants would have given their informed consent beforehand
strength of overt observations
overt observations have an ethical advantage to covert observations because participants are aware of what is going on and have given consent
weakness of overt observations
the slight disadvantage is that having this awareness could mean participants behave differently to normal and behaviour is not as natural (weakens internal validity)
covert observations
the participants are totally unaware they are the focus of a study and their behaviour is observed in secret, say from across the room or from a balcony
participants are made aware after the study of what took place
observations can take place through a 2 way mirror (participant’s cannot see the observer)
strengths of covert observations
good internal validity because the participants are unaware of the observations, the behaviour will be natural (less likely to suffer from demand characteristics)
weakness of covert observations
ethics of these studies may be questioned, as people may not wish their behaviour to be studied without their initial consent
participant observation
sometimes it may be necessary for the observer to become part of the group they’re studying, this is participant observation
for example, a researcher may join the workforce at Primark (as mentioned earlier) to get first-hand account of the relations between staff and managers
strengths of participant observations
can provide real insight into participants being studied and this richness may not be gained in any other way (increases internal validity)
weakness of participant observation
there is a danger the observer may identify too strongly with those they’re studying and as a result lose their objectivity
non-participant observation
in most cases, the observer is merely watching (or listening) to the behaviour of others and remains separate from the participants in the study. This is non-participant observation
strengths of non-participant observations
observers are more likely to remain objective because they aren’t part of the group being studied
weakness of non-participant observation
they may lose valuable insight into the participants because they are too removed from the people and behaviour (decreased validity)
unstructured observations
the researcher records all relevant behaviour, but has no system. they may simply write down everything they see
…clearly there may be too much to record as well as recording behaviour that may not be that important
structured observations
it is preferable to use these observations; they aim to be objective and rigorous
the researcher uses a pre-determined list of behavioural categories and sampling methods
developing behavioural categories
for structured observations
researchers need to be very clear on exactly what behaviour they’re looking for. it is operationalising - breaking up the behaviour in a set of components so it can be measured
what should the behavioural categories be and explain them
objective - the researcher should not have to make guesses about behaviour. The categories must be observable
no waste basket - in other words all possible behaviours are covered and avoiding a ‘waste basket’ category, in which loads of different behaviour is thrown in because it’s unclear where the behaviour should be categorised
independent of each other - categories should not overlap, meaning that the researcher has to mark two categories at one time
two sampling methods of structured observations
event sampling - this involves the times a particular behaviour (event) occurs in an individual or target group
time sampling - this method records behaviour within a particular time frame. for example noting what an individual is doing every 30 seconds, or some other time frame
evaluation of observational methods
structured v unstructured - structured are designed to use behavioural categories that make the recording of behaviour easier, data is likely to produce quantitative data which means analysing and comparing the behaviour observed is straightforward
unstructured tend to produce qualitative data, which may be harder to analyse, also higher risk of ‘observer bias’ as behavioural categories aren’t used. researchers may record behaviour that simply ‘catches their eye’ and could also miss important behaviours
behavioural categories - having categories can make data collection easier, adds structure and they’re objective but they need to be very clear avoiding ‘waste basket’ category
sampling - event sampling is useful when the target behaviour or even happens infrequently and could be missed if time sampling was used however if the event is too complex the observer may overlook important details if using event sampling
time sampling, effective in reducing the number of observations that have to be made however those instances when behaviour is sampled might be unrepresentative of the observation as a whole
case studies
in-depth information, can lead to greater internal validity due to richness of information gathered. people being studies are normally pretty unique and are studied with the aim of answering difficult or important questions that cannot be investigated experimentally
might include medical records, attitude tests, school records and reports, interviews with family members or colleagues, tests of clinical symptoms
advantages of case studies
rich in detail - provide great depth and understanding about individuals
the only possible method to use - case studies allow psychologists to study unique behaviours or experiences that could not have been studies any other way. the method also allows ‘sensitive’ areas to be explored, where other methods would be unethical, like the effects of sexual abuse
useful for theory contradictions - just one case can contradict a theory
disadvantages of case studies
not representative - as no two case studies are alike, results cannot be generalised to others
researcher bias - researchers conducting case studies may be biased in their interpretations or method of reporting, making findings suspect
reliance on memory - case studies often depend on participants having full and accurate memories
correlation
a way psychologists can measure the strength between two or more co-variables (things that are measured)
types of correlation
scattergram
positive correlation - where one co-variable increases and so does another
negative correlation - where one co-variable increases and the other decreases
no correlation - the variables have no relationship
correlational hypothesis
predict a relationship between two variables not a difference (like in experiments), and therefore are worded differently to experimental hypotheses
directional correctional hypothesis
states whether the relationship will be a positive or a negative correlation
example - there will be a significant positive correlation between temperature and ice-cream sales
non-directional correlational hypothesis
simply states that there will be a correlation
example - there will be a significant correlation between average time spent reading per week and scored on an I.Q test
correlation co-efficients
the strength of the correlation
the sign tells you the direction of the correlation, positive or negative
the number tells you the strength : closer to 1 means stronger, closer to 0 means weaker
correlation co efficient numbers
-1.0 perfect negative correlation
between -1.0 and -0.8 is moderate negative
between -0.8 and -0.5 moderate negative
between -0.5 and 0 weak negative
0 = no correlation
between 0 and 0.5 is weak positive
between 0.5 and 0.8 is moderate positive
between 0.8 and 1.0 is strong positive
and at 1.0 perfect positive correlation
correlations are designed to investigate …
the strength and direction of a relationship between two variables
the closer the correlation coefficient to 0 …
the weaker the correlation
the closer the correlation coefficient is to 1 (or -1) …
the stronger the correlation
strengths of correlation
correlations are useful as a tool of research as they provide a strength and direction of a relationship between variables and can be used as a starting point to assess the relationship between variables before committing to an experimental study
they allow researchers to look at the relationship between variables that you would not be able to experimental investigate e.g. obesity and junk food consumption
they are also quite quick and economical to carry out because there is no needs for a controlled environment and no manipulation of variables is required, so they are less time consuming than the planning and execution of setting up an experiment
they can also use secondary data (data collected by others) which means they are less time consuming
limitations of correlation
correlation doesn’t provide a cause and effect relationship: we cannot conclude that one variable is causing the other to change. this can sometimes lead to correlations being misinterpreted or misused
although correlations can tell us the strength and direction of variables, they cannot tell us why the variables are related
it may also be the case that another untested variable is causing the relationship between the two co-variables, this is known as the third variable problem
content analysis
a type of observational research technique in which people are studied indirectly via the communications they have produced
the forms of communication that may be subject to content analysis are wide -ranging and may include spoken interaction such as conversations or presentations, written communications texts or emails, broader examples from the media books magazine tv programmes or films
coding and quantitative data
categorising of information into meaningful units, needs to be done because often the data set to be analysed can be extremely large
may involve counting the number of times a particular word or phrase appears in the text to produce a form of quantitative data
may also involve gathering qualitative data
thematic analysis and qualitative data
process of coding and identification of themes are closely linked as themes may only emerge once data has been coded
a theme in content analysis refers to any idea, explicit or implicit, that is recurrent (reoccurs often), these are likely to be more descriptive than coding units
once the researcher is satisfied that the themes they have developed cover most aspects of the data they are analysing, they may collect a new set of data to test the validity of the themes and categories
strengths of content analysis
high in external validity because the data is obtained from real life experiences e.g. journal entry
content analysis is flexible in the sense that it may produce quan. and qual. data depending on the aims of the study
limitations of content analysis
communication that people produce is normally analysed outside of the context within which it occurred. this is a limitation as the researcher may attribute opinions and motivations to the speaker or writer that were not originally intended
can be difficult to decide on appropriate categories/codes/themes and time-consuming to carry out
it may suffer from a lack of objectivity, especially when more descriptive forms of thematic analysis are employed
population
large group of individuals that a particular researcher may be interested in studying
target population
target of people but also subset of the general population
too large to study therefore the researcher selects a sample of this target population
sample
a group of people who take part in research and is taken from the target population
researchers aim to obtain a representative sample so that …
the findings can be generalised
bias (in the context of sampling)
bias can occur if certain groups may be over or under represented within the sample selected, this limits the extent to which generalisation can be made to the target population
generalisation (in the context of sampling)
the extent to which the findings and conclusions from a study can be applied to the population, this is made possible if the sample of participants is representative of the population
opportunity sampling
where a researcher decides to select anyone who is available and willing to participate in their study. students are often used in psychological research for this reason
strengths of opportunity sampling
convenient as it saves time, effort and is less costly
limitations of opportunity sampling
the sample is likely to be unrepresentative of the target population as it’s drawn from a specific area such as one street in one town
the researcher has complete control over the selection of participants, they may simply avoid people they don’t like the look of (researcher bias)
random sample
a form of sampling in which all members of the target population have an equal chance of being selected
to select a random sample, firstly a complete list of all members of the target population is obtained
secondly all the names are assigned a number
thirdly the sample is generated through the use of some lottery method (computer-based randomiser or picking numbers from a hat/container)
strengths of random sampling
it is free from researcher bias. the researcher has no influence on who is selected therefore selecting people who people who they think may support their hypothesis
limitations of random sampling
very difficult and time-consuming to conduct, a complete list of the target population may be extremely difficult to obtain
participants selected may refuse to take part (resulting in further bias)
systematic sampling
a form of sampling with every nth member of the target population is selected
a sample frame is produced which is a list of people in the target population organised into, for instance, alphabetical order. the researcher then works through selecting every 5th, 3rd, 9th person etc
strengths of systematic sampling
avoids researcher bias, once the system for selection has been established the researcher has no influence over who is chosen
it is usually fairly representative
limitations of systematic sampling
process of selection can interact with hidden ‘traits’ within the population. if the sampling technique coincides with the frequency of the trait, the sampling technique is neither random, no representative. for example if every fourth property in a street is a flat occupied by a young person, then selecting every fourth property will not provide a representative sample
stratified sampling
a sophisticated form of sampling. from the wider population, a sub-group is created (strata) based on age, social class etc. , then the population is randomly sampled within each strata
to carry out a stratified sample the researcher first identifies the different strata that make up the population
the proportions needed for the sample to be representative are worked out
finally the participants that make up each strata are selected randomly the way its done for random sampling
strengths of stratified sampling
technique avoid researcher bias, once the target population has been sub-divided into strata, the participants that make up the numbers are randomly selected
this method produces a representative sample because it’s designed to accurately reflect the population, which means generalisation of the findings becomes possible
limitations of stratified sampling
requires a detailed knowledge of the population characteristics, which may not be available
can be very time consuming dividing a sample into a strata and then randomly selecting from each
volunteer sampling
also known as self-selected sample
where participants select themselves to be part of the study
a researcher may place an advert online or newspaper or noticeboard for example, and people respond wanting to take part in the study
strengths of volunteer sampling
requires little effort from researchers (other than producing an advert) as the participants volunteer themselves
limitations of volunteer sampling
sample will be bias and unrepresentative as volunteers tend to be a certain ‘type’ of person. this makes results difficult to generalise to a target population
volunteers are eager to please, which increases the chance of demand characteristics, for example participants giving the answer they think is required
pilot studies
small-scale trial run of the actual investigation, takes place before the real investigation is conducted
its aim is to check that the procedures and materials work, and the instructions to participants are clear
allows the researcher to make any changes if necessary before the investigation is carried out
pilot studies what
not restricted to experiments, can be used for self-reports like questionnaires or interviews: in this case it may be useful to try out question in advance and remove and replace words or questions that may be confusing
also with observational studies, a pilot study would be a good way to check the behavioural categories are effective before the real observation takes place
ethical issues
conflicts arise between the rights of participant in research studies and the goals of researchers to produce valid data
the BPS code of ethics is a legal document instructing psychologists in the UK about what behaviour is and is not acceptable when dealing with participants
what are the major ethical issues
deception
informed consent
protection from harm
privacy (confidentiality)
right to withdraw
deception
deception means deliberately misleading or withholding information from participants. despite this, there are occasions when deception can be justified if it doesn’t cause undue distress
for example, Milgram deceived his participants by telling them it was a study on ‘learning’ not obedience
informed consent
participants should be aware of what they’re doing. informed consent is making participants aware of the aim of the research, the procedures and their rights. they can make an informed decision on whether they want to take part
For example, Loftus didn’t tell participants the true aim of the study (to investigate the effect of anxiety on EWT)
protection from harm
participants should not be placed in any physical or psychological risk (i.e. feeling embarrassed, inadequate or placed under undue stress)
For example, participants in Zimbardo’s study were subject to psychological harm. The study was stopped after 8 days because of this
Privacy (confidentiality)
participant’s data should not be disclosed to anyone unless agreed in advance. Numbers should be used instead of names. Participants shouldn’t be able to identify themselves either.
For example, important case studies in memory are often referred to by initials HM, KF
right to withdraw
participant should be aware they can leave a study at any time, and even withdraw their data after the study is finished
for example, Milgram’s study participants were made aware that they could leave the study at any time but due to the nature of the research participants felt they did not have this right
BPS
british psychological society
BPS code of conduct
The BPS has its own ethical guidelines. psychologists have a professional duty to observe these guidelines. the guidelines are closely matched to the ethical issues above and attempt to ensure all participants are treated with respect and consideration during a piece of research
dealing with informed consent
participants should be issued with a consent letter/form detailing the relevant information that may affect their decision to take part
assuming the participant agrees, then the consent form is signed. for investigations involving children under 16, a signature of parental consent is required
alternative ways of getting consent if you can’t get informed consent
presumptive consent - rather than getting consent from the participants themselves, a similar group of people are asked if the study is acceptable. if the group agree, then consent of the original participants is ‘presumed’
prior general consent - participants give their permission to take part in a number of different studies - including one that will involve deception. By consenting, participants are effectively consenting to be deceived
retrospective consent - this involves asking participants for consent after they have participated in the study (debriefing). they may not have been aware of their participation. However they may not consent and have already taken part
ways of dealing with deception
at the end of the study, participants should be given a full debrief. this means they should be told the true aims of the research, the various conditions of the research, and what their data will be used for. they should be told they can withhold their data if they wish
ways of dealing with privacy
if personal details are held these must be protected. however, it’s more usual for researchers to use numbers rather than names
peer review
before a piece of research can become part of a journal it must be rigorously checked
the research is scrutinised by a small group of usually two or three experts (peers) in the particular field. these experts should be objective and unknown to the author. this helps any research intended for publication is of high quality
main aims of peer review
allocation for research funding - research is paid by various charitable bodies for research that is seen as worthwhile, overall budget for science set at £5.8 billion in 2015 - 2016
assess the quality and relevance of research - all elements of the research is assessed for quality and accuracy: if the hypotheses, research method, statistics and conclusions are appropriate and relevant
suggesting improvements - peer reviews may suggest minor changes to the work to therefore improve the report that’s been submitted. in extreme circumstances they may conclude the work is inappropriate for publication and should be withdrawn
assessing the research rating of university departments - the funding universities get depends upon the good rating they receive from the peer review process
evaluation of peer review
finding an expert - it isn’t always possible to find an appropriate expert to review a research proposal (research to be done) or report (research already done)
anonymity - the process can be done so that the ‘peer’ remains anonymous, so that an honest and objective appraisal can be achieved. However it’s not unheard of where a minority or reviewers may use their anonymity as a way of criticising rival researchers who may have crossed them in the past. nowadays, peer reviews may be ‘open’ which is where both the author and reviewer know each other’s identity
publication bias - the editors of journals want to publish significant ‘headline grabbling’ findings to increase the circulations of their publication. this means they may prefer to publish research with significant (positive) results. this could mean research that doesn’t reach these criteria could be ignored - This creates a false impression of the current state of Psychology if editors are being selective/bias in what they publish
burying ground breaking research - the peer review process may supress ground breaking research that may contradict the views of the reviewer. established scientists are the ones likely to be chosen as reviewers, but this may mean results of research that coincide with current opinion are more likely to be passed than new, fresh and innovative research that poses a challenge to the established order
qualitative data
is expressed in words, rather than numbers or statistics. it may take the form of a written description of the thoughts, feelings and opinions of participants. For example, a transcript from an interview, an extract from a diary or notes
quantitative data
data expressed numerically, this form of data usually gathers numerical data such as individual scores from participants such as the numbers of words a person was able to recall in a memory experiment. data is open to being analysed statistically and and can be expressed using graphs, charts etc
limitations of qualitative data
can be difficult to analyse, it does not lend itself to being summarised statistically so that patterns and comparisons can be drawn
a consequence of this is that conclusions may rely on subjective interpretations of the researcher which may be subject to bias
strengths of qualitative data
it offers the researcher more richness and detail which can provide unexpected insights upon behaviour. this is due to the fact the participants have more license to develop their thoughts, feelings and opinions on a given subject. the data does have more internal validity
strengths of quantitative data
as the data is numerical, its objective and less subject to bias. it is also far easier to analyse and draw conclusions
limitations of quantitative data
however, it is much narrower in scope and meaning than quantitative data and therefore not fully representative of real-life
primary data
data that has been gained directly from the participants, it would be specifically related to the aims and/or hypothesis of the study. the data conducted from participants doing an experiment, questionnaire, interview or observation would be classed as primary
secondary data
data collected by someone other than the person conducting the study. this may be data that already exists before the psychologists begin their research
examples would be data in journal articles, books, websites, government statistics etc.
a piece or research that uses secondary data is a …
meta-analysis
meta-analysis
a type of research method that uses secondary data
the researcher uses the data from a large number of studies, which have involved the same research questions and methods. the results of all these studies are analysed to give an overview and conclusion
the researcher may simply discuss the findings/conclusions - which is a qualitative analysis. or they may perform a statistical analysis on the combined data. this may involve calculating the effect size
effect size
gives an overall statistical measure of the difference or relationship between variables across a number of studies
strengths of primary data
the researcher has control of the data in that it can be designed to fit the aims and hypothesis of the study
limitations of primary data
requires time, effort and can be expensive
conducting an experiment, for instance requires considerable planning, preparation and resources, considering secondary data which can be accessed within a matter of minutes
strengths of secondary data
inexpensive and easily accessed
the data has probably already been statistically tested and peer reviewed
limitations of secondary data
the content of the data may not exactly fir the needs of the study. it may be incomplete or out-dated and therefore not match the researcher’s objectives
there may be substantial variations in the quality and accuracy of secondary data
descriptive statistics
ways of summarising and analysing quantitative data in order to draw meaningful conclusions
measures of central tendency, measures of dispersion, graphs
measure of central tendency
‘averages’ which give us information about the most typical values in a set of data
three to consider: mean, median, mode
strengths of the mean
the most sensitive as it includes all the scores/values in the data set within the calculation
due to the above point, it’s more representative of set of scores
limitations of mean
easily distorted by extreme values
median strengths
not affected by extreme scores
once arranged in order the median is easy to calculate
median limitations
it is not as sensitive as the mean, as not all scores are included
mode strength
very easy to calculate
unaffected by extreme values
limitations of mode
not very useful if there are several modes
measures of dispersion
how ‘spread out’ the data is
how far scores vary and differ from one another
range and standard deviation
range
an incredibly easy measure of dispersion to calculate
most useful when assessing how representative the median is a typical score
this is because the median only takes into account the one score in the middle of the set
range median correlation
the higher the range, the less representative the median is because it would indicate that the scores are spread widely from that figure
standard deviation
sophisticated measure of dispersion. a single value tells us how far all scores deviate (move away from) the mean
high standard deviation
suggests a greater spread of scores around the mean. for example in an experiment looking at the amount of words recalled after an interference task, a large standard deviation would suggest not all participants were affected by the IV in the same way, because the data is widely spread
what does a large standard deviation suggests
not all participants were affected by the IV in the same way, because the data is widely spread
a low standard deviation
suggests the scores are clustered close to the mean. we could imply from this that participants responded in a similar way. this would indicate that mean is more representative as a typical score. this is because a low score indicates a low average distance between each score and the mean
