RESEARCH METHODS - psychology unit 4

what does a researcher pay attention to in pilot studies?

- Do participants understand the instructions? - Is the task/material used appropriate, e.g. can participants read the words on a word list - Are the timings ok? Do you need to give the participants more or less time to complete the task - When constructing a questionnaire or interview are questions clear and unambiguous? - Are participants getting bored? Is the task too long or are there too many things to do?

improving validity

- Internal validity can be improved by reducing investigator effects, minimising demand characteristics (e.g. by hiding the true aim of the investigation), using standardised instructions and a random sample. - External validity can be improved by setting experiments in more naturalistic settings and ensuring tasks have mundane realism (reflect tasks which occur in real life).

paradigms (kuhn)

- a shared set of assumptions about psychology and the methods appropriate to its study. A paradigm dictates what is studied and researched and the types of questions that are asked. - scientists collect data that fits the current paradigm, creating bias where scientific journals publish confirmatory examples of research rather than opposing ones. - occasionally a paradigm is replaced with a new paradigm which emerges from a minority position.

what should a consent form include?

- enough detail for the person to give informed consent - they should know exactly what will be expected of them, and any potential for psychological or physiological harm. - they should be informed of their right to withdraw at any point during or after the procedure - they should be reassured that their results will remain anonymous and confidential - they should be given the opportunity to ask any questions they may have

deception can be excused when...

- it does not lead to harmful consequences for participants. - the study is potentially very useful to society or our understanding - not deceiving participants may make the research invalid - in some studies, if participants are told exactly what the study is about, then they may behave differently (this is called demand characteristics).

features of science: theory construction and hypothesis testing (Popper, 1935)

- tentative theories are the first stage of the scientific process and are used to generate predictions. - predictions are put into the form of testable hypotheses and are experimentally tested. - data produced is statistically analysed to see if the theory is supported or falsified. - adjustments in the theory can then be made.

what should a debrief include?

- they should be thanked for their time - the full aim of the study should be revealed, and the results expected explained - they should be reminded of their right to withdraw their results and that their results are confidential - they should be given the opportunity to ask questions

protection from harm can be ignored as an ethic if...

-t he harm is unavoidable, and the study could not be carried out in any other way; - the harm is short-term and relatively minor; - the participants do not leave the study with any residual harm or distress.

strengths of correlations

Correlations don't require manipulation of variables so they can be used in situations where it would be unethical to carry out an experiment, for example looking at the relationship between number of hours revision and score on a psychology exam - it would be unethical to ask students to do no or little revision so you would need to ask people how much revision they had done after the exam.

skewed distribution

In some populations scores are not distributed equally around the mean - this is a skewed distribution. A positively skewed distribution is where there are a few high extreme scores. These outliers affect the mean, which is always higher than the median and mode in a positive skew. A negatively skewed distribution is where there are a few low extreme scores. These outliers affect the mean, which is always lower than the median and mode in a negative skew

overt observation

The participants know their behaviour is being watched and recorded and for what purpose.

participant observation

The researchers become part of the group or situation being observed.

naturalistic observation

These take place in a real life, natural environment where no manipulation is made and everything has been left as it is normally.

controlled observation

These take place in an artificial laboratory setting. The researcher manipulates some aspect of the environment.

thematic analysis

This type of content analysis also analyses qualitative data but is not worried about quantifying it, focusing instead on identifying themes or ideas in the data. the data is searched for themes and similar themes are sorted together.

type 2 error

Type 2 error - this is when the null hypothesis is accepted, but actually the null hypothesis was incorrect. This tends to occur when the probability value is too strict.

strengths of questionnaire

a cost-effective and time-efficient way of gathering data is they can be quickly administered to large numbers of participants at the same time which means that a more representative sample can be gathered. when completed privately and anonymously they can provide honest data, improving the validity of the research. the reduced involvement of the researcher reduces the impact of investigator influence on the behaviour of participants.

strengths of case studies

a large amount of rich, in depth data is gathered, giving a very detailed insight into the person or group under investigation. Information that may be overlooked using other methods is likely to be revealed. vital for investigating aspects of human behaviour and experience that are rare or cannot be replicated experimentally due to ethical constraints, e.g. cases of people with brain damage, or how people respond to terrorist attacks or natural disasters.

assessing reliability: inter-rater reliability

a means of assessing whether different observers are viewing and rating behaviour in the same way. A correlational analysis of all the observers' scores is conducted. A strong positive correlation indicates that they are observing and categorising behaviour consistently.

self-report techniques

a method of gathering data where participants report on themselves by answering questions or responding to statements. These are used to establish the attitudes of people and find out what they think or feel about something. Two methods of doing this are questionnaires and interviews

hypothesis

a precise, testable prediction of what is expected to happen. for an experiment, it predicts the effect of the IV on the DV, and for a correlational study it predicts what type of relationship will occur. hypotheses should be operationalised.

systematic sample

a predetermined system is used to select participants, e.g. every 10th person from a phonebook.

questionnaires

a set of pre-determined questions for participants to respond to. They could be completed in the presence of the researcher or could be sent through the post or emailed for participants to complete on their own and send back

problems with peer review: detecting fraudulent research

a small number of cases peer review has failed to detect fraudulent research, including fabrication (where results are made up), falsification (where results are 'tweaked' to give a more significant result) and plagiarism (where someone else's work is copied).

pilot study

a small trial run of a study to test any aspects of the design. This allows the researchers to make any improvements or adjustments to design or procedures before the main investigation.

the double-blind procedure

a technique to reduce investigator effects. this involves neither participants nor the researcher knowing which condition participants are in. this prevents the researchers from unconsciously giving participants clues as to which condition they are in and therefore reduces demand characteristics.

what is peer review?

all published research has to go through a peer review. This is where the editor of the journal sends the research sent to them to other psychologists who are experts in that topic. these experts read the research carefully and return it to the editor with comments and a recommendation about its suitability for publication, whether it needs to be revised, or whether it should be rejected.

unstructured interviews

an informal discussion on a particular topic. Interviewers can explore interesting answers by asking follow-up questions.

strengths of interviews

any misunderstood questions can be explained and individual questions can be adapted so they are understood by all participants. structured interviews are very easily replicated to check the reliability of findings as the questions are standardised. (however unstructured interviews are more difficult to replicate.)

extraneous variable

any variable other than the IV which could affect the DV. these need to be controlled so they stay constant across all of the experimental conditions and between participants.

weaknesses of case studies

as case studies are one-off cases they cannot be replicated, so case studies lack reliability. the findings can't usually be generalised. Subjects of case study are usually unusual or special in some way so are not representative of the rest of the population. findings are likely to be subjective. Due to the longitudinal nature of many case studies the researcher will often form an attachment to the people they are studying which could bias their interpretations.

strengths of content/thematic analysis

as the method doesn't require direct contact with people, investigators cannot influence behaviour content analysis can be used to complement other methods and is especially useful as a longitudinal tool in looking at trends and changes over time.

sample

as the researcher can't study all of the people in the target population they need to select a smaller group

bar charts

bar charts allow data in the form of categories to be compared. Categories are placed on the x (horizontal) axis. The y (vertical) axis can show totals, means or percentages. The columns/bars should all be the same width and separated by spaces because the variable on the x-axis is not continuous. Bar charts can display two values together, e.g. male and female consumption of chocolate as shown by gender and age.

strengths of naturalistic observation

behaviour observed is likely to be natural so findings can be generalised to everyday life (ecological validity is high).

range (measure of dispersion)

calculated by subtracting the lowest value from the highest value in a data set. It is easy to calculate but is affected by extreme values and doesn't show the distribution of the data (e.g. are most values clustered around the mean with one extreme value).

spearman's rho (rs) test

can be applied to correlational data. to find out whether the relationship found is strong enough to be significant. If rs is equal to or greater than the critical value the relationship is significant.

weaknesses of random sample

can be time consuming to get a list of all members of a population and allocate them all a number. unbiased selection doesn't guarantee an unbiased sample; all females could be randomly selected by chance. Also not everyone selected will be able to be contacted or will be able to take part.

hypothesis: difference level of measurement: nominal research design: independent

chi-squared

primary data

comes directly from first-hand experience. it is new data collected by the researcher in a study. the main strength of this is the control the researcher has. they can design the data collection so it is valid and fits the aims and hypothesis of the study. however designing a study, recruiting participants, conducting a study and analysing data is time consuming and needs to be funded.

assessing validity: concurrent validity

correlating scores on a test with another test which is known to be valid. A strong positive correlation would suggest good concurrent validity.

weaknesses of correlations

correlations are not able to show cause and effect - we can't say for sure that one variable causes another. For example you could find a positive relationship between amount of revision and score on an exam, but we can't say for sure that revising CAUSES you to do better. It could be that those who are good at psychology want to revise more while those who dislike psychology or think they're bad at it might not bother

how to deal with weaknesses of repeated measures

counterbalancing - half the group does condition 1 first, the other half does condition 2 first.

qunatitative data

data in the form of numbers or quantities.

weaknesses of non- participant observation

data lacks the richness provided by participant observation, such as feelings and motivations of participants.

a single-blind procedure

demand characteristics can be reduced with this procedure. this is where the participants have no idea what condition of a study they are in. for example in a drug trial they wouldn't know whether they have been given the real drug or a placebo.

weaknesses of stratified sample

detailed knowledge of population characteristics is required which may not be available - there may be hidden subgroups the researcher is not aware of, or who are not accessible. the dividing of a population into strata and then randomly selecting from each can be time consuming.

assessing validity: temporal validity

do research findings remain true over time?

assessing validity: face validity

does the test simply look like it tests what it's meant to test?

strengths of opportunity sample

easiest method and takes less time to locate and recruit the sample than other methods.

strengths of field experiments

ecological validity is high as participants are in their natural environment and the situation/tasks are likely to reflect those in real life. therefore behaviour is likely to be natural. participants may not know they are participating in a study, therefore they are less likely to try to work out the aim of the research and display demand characteristics.

strengths of covert observation

ecological validity is high because completely natural behaviour will be observed.

strengths of overt observation

ethically sound as participants know they are being observed and will have given informed consent. They are able to withdraw from the study at any point.

strengths of lab experiments

experimenters can control all variables and the IV and DV are precisely operationalised. this allows replication - other researchers can repeat the experiment to check the reliability of the results. As all other variables are controlled, any change in the DV must be caused by the manipulation of the IV, so a cause and effect relationship can be determined.

features of science: measurable variable

for a study to be scientific there should be a measurable variable. this is difficult in psychology as many of the factors studied are unobservable, e.g. memory. therefore a valid measurement is needed, e.g. with a memory test.

right to withdraw

giving people the opportunity to leave the study at any time if they no longer want to take part. This means that they can withdraw from the study and that the researcher will not use any of their data. participants should be told about their right to withdraw and if they are being paid for participating they should be informed that they will still be paid if they drop out.

closed questions

have a predetermined range of answers, e.g. multiple choice and likert scales. this produces quantitative data, making it easy to analyse using statistics and graphs. however the data collected may lack validity as participants can be forced into giving an answer which doesn't reflect their true feelings. There can also be response bias, where participants tend to select 'don't know'.

strengths of controlled observations

high level of control. The behaviour being observed can be isolated and the environment can be manipulated to make measurement of this behaviour easier and more objective.

structured interviews

identical pre-determined questions are read to participants, with the interviewer writing down the answers.

when should we use p≤0.01?

if it could be dangerous if the null hypothesis was rejected incorrectly, e.g. for drug trials.

for chi-squared and spearman's...

if the calculated value is equal to or greater than the critical value then the results are significant and the alternative hypothesis is accepted and the null hypothesis is rejected.

for mann-Whitney and wilcoxon...

if the calculated value is equal to or less than the critical value then the results are significant and the alternative hypothesis is accepted and the null hypothesis is rejected.

time sampling

if the order of events is important in the observation then the researchers can occur when each behaviour occurred. For example, observations can be taken at predetermined intervals of 30 seconds which are then recorded chronologically.

open questions

invite participants to provide their own answers, producing qualitative data. These sometimes provide unexpected answers, allowing researchers to gain new insights. as participants can express their attitudes or feelings in their own words open questions collect valid data. however participants who are less literate may be put off and most participants avoid giving lengthy answers. data collected via open questions can be difficult to analyse objectively.

meta- analysis

involves combining data from lots of smaller studies in a particular research area into one large data set. T this allows the identification of trends and relationships that wouldn't be possible from individual smaller studies. this is particularly useful when a number of smaller studies have found contradictory results in order to get a clearer view of the overall picture.

nominal data

involves counting frequency data, for example how many days of the week were rainy and sunny, how many people are smokers or non-smokers. Tally charts are used to record this type of data.

the mean

is calculated by adding up all the bits of data and dividing by the number there are. It is the most sensitive measure of central tendency because it takes account of all the values in the data set. However the mean is easily distorted by extreme values, so may not represent the data as a whole.

the median

is the middle value when the data set is put in numerical order. If there is an even number of bits of data there will be two central values so need to be added together and divided by 2. the median is not affected by extreme values so can be useful when the mean is inappropriate. However it isn't as sensitive as the mean.

the dependent variable

is the variable that the experimenter measures to see whether the IV had had any effect on it.

weaknesses of matched pairs

it is time consuming and it is difficult to have groups matched on every characteristic that could affect the results.

weaknesses of lab experiments

laboratory experiments lack ecological validity. the laboratory setting is artificial and unlike real life. Therefore participants don't behave naturally and it can be difficult to generalise the results to other situations. participants are aware they're being experimented on so may alter their behaviour and display demand characteristics.

strengths of non-participant observation

lack of direct involvement ensures greater objectivity when interpreting behaviour.

weaknesses of covert observation

lack of informed consent means this type of observation raises ethical issues. However it is ethically acceptable to observe people without their knowledge as long as they are in a public place.

assessing validity: predictive validity

look at how well a test predicts future behaviour, e.g. do average GCSE results accurately predict A level results?

hypothesis: difference level of measurement: ordinal research design: independent

mann-whitney

assessing reliability: test-retest method

measures external reliability by giving the same test to the same participants on two occasions. If the same result is obtained then reliability is established.

assessing reliability: split-half method

measures internal reliability by splitting a test into two and having the same participants do both halves. If the two halves of the test give similar results this indicates that the test has internal reliability.

weaknesses of natural experiments

natural and quasi experiments can't demonstrate causal relationships because the IV is not directly manipulated. random allocation of participants into conditions is not possible, therefore there may be uncontrolled confounding variables, reducing the internal validity of the study. natural experiments are very difficult, if not impossible, to replicate as exactly the same conditions are unlikely to occur again.

strengths of natural experiments

natural experiments allows research where the IV can't be manipulated for ethical or practical purposes. allow researchers to study real life issues, so ecological validity is very high.

weaknesses of participant observation

objectivity of observation is affected by becoming part of the group being observed. The observer is likely to form their own opinions about the participants which could affect their interpretation of their behaviour.

observational techniques

observations involve watching and recording behaviour. They are often used as part of an experiment as a means of measuring the dependent variable so are often a technique which is used alongside other methods.

consent forms

often used to gain informed consent. it should explain the objectives of the study and what it will involve. anything that might influence the participants' willingness to take part should be disclosed. the participants should also be given the opportunity to question the researcher about the study. the information sheet also provides an opportunity to tell the participants that they can withdraw from the study at any time.

event sampling

one way of recording behaviours is to record on the checklist of behavioural categories the number of times a particular behaviour occurs. The chronological order of events is not recorded, there is just a frequency count for each behaviour

standardisation

one way to control extraneous variables and prevent confounding variables from affecting the results of the study. all participants in the study should have exactly the same experience. So procedures should be standardised to make sure of this, for example the instructions given to participants should be exactly the same, any stimulus should be shown in the same way for exactly the same length of time.

counterbalancing

one way to control extraneous variables and prevent confounding variables from affecting the results of the study. if participants are going to do all conditions then half should do condition 1 first and half should do condition 2 first. This avoids order effects, for example where participants improve in the second condition due to practice.

randomisation

one way to control extraneous variables and prevent confounding variables from affecting the results of the study. if there are two different conditions participants should be randomly allocated to the conditions to ensure participant variables such as intelligence don't become confounding variables. E.g. if you have 20 participants and you need ten in condition 1 and ten in condition 2 you can use a computer programme to randomly choose who is in each condition. Participant variables should then be evenly spread across the two conditions.

debriefing

one way to make sure that participants leave the experiment without suffering on-going distress or harm is to debrief them, i.e. tell them exactly what the experiment was all about and reassure them that their behaviour in the experiment was 'normal'.

strengths of independent measures

order effects are not a problem as participants only do one condition of the study so will not be affected by practice or fatigue.

weaknesses of repeated meaures

order effects could influence the results as the same group of participants is used for each condition. So performance could get better in the second condition due to practice, or decline due to fatigue.

strengths of repeated measures

participant variables are not an issue as the same group of participants is used for each condition.

weaknesses of independent measures

participant variables could affect the results as there are different groups doing each condition, e.g. one group may have naturally better memories than the other.

demand characteristics

participants can also affect results. they might read things into the situation and start changing their behaviour responding to the perceived demands of the study.

weaknesses of interviews

participants may be affected by interviewer effects, for example women may be less willing to talk about sex with male interviewers. Interviews are also likely to be affected by demand characteristics and social desirability bias. unstructured interviews can be difficult to conduct in a way that collects meaningful data. Interviewers may need to be trained, especially when concerning sensitive topics.

weaknesses of overt observation

participants may not behave naturally if they are aware of being observed. They may show demand characteristics or show what they think is socially desirable behaviour.

demand characteristics: hawthorne effect/ screw-you effect

participants may try to guess what the investigation is about then behave in the way they think the investigator wants them to. they may deliberately try to behave in an unexpected way.

demand characteristics: wanting to appear normal

participants may worry about being in a psychological study and want to appear 'normal', this may change their behaviour and they may behave in ways they wouldn't do normally.

protection from harm

participants should not be harmed, either physically or psychologically when participating in research. In practice, however, participants often suffer distress or pain during the course of an experiment and this is considered acceptable as long as

why is peer review important?

peer review acts as a quality control, ensuring that only the best quality scientific research is published, increasing Psychology's credibility.

problems with peer review: bias

peer review is subject to bias. Work that is consistent with an existing theory is more likely to be accepted for publication. Peer reviewers may be strongly opposed to the views expressed in the report which could influence their views on the quality of the research. There is also institution bias where reviewers may be influenced by the university the research has come from, e.g. assuming that work from a prestigious institution will be high quality.

problems with peer review: file drawer phenomenon

peer review tends to favour positive results (where the results support the hypothesis). Many negative findings (where the null hypothesis is accepted) are not published. This may give a false impression of the strength of evidence for a particular theory.

opportunity sample

people who are most convenient or most available are recruited, e.g. people walking by in the street, students at a school

confidentiality

personal information given by participants during the course of a study, or the results of tests taken by participants, is kept confidential. results can be published, but they must not identify the individual, either by name, or by any other way. Initials, numbers or pseudonyms may be used. participants should always be told that personal data will be kept confidential.

researcher characteristics affecting the study

physical characteristics of investigators may influence results, such as age or ethnicity. for example male participants may not want to admit to sexist attitudes to a female researcher. less obvious personal characteristics of investigators like accent or tone of voice. for example participants may react differently to someone with a stern voice and demeanour. investigators may be unconsciously biased in their interpretation of data and find what they expect to find.

aim

precise statement of why a study is taking place.

directional/one tailed hypothesis

predicts the expected direction of the results (so which group will do better in an experiment, or whether a relationship will be positive or negative. they are used when previous research suggests that results will go in one direction, or when replicating a previous study

non-directional/ two tailed

predicts there will be an effect, but doesn't predict the direction of the results. Again, these are worded differently for an experiment and for a correlation. they are used when past research is contradictory, or when it is a new area of study with no previous research.

null hypothesis

predicts there will be no effect and that any apparent effect will be due to chance.

strengths of participant observation

provides greater insights into behaviour that may not be gained by an 'outsider'.

good implications that psychological research has on the economy

psychological research often has practical applications which can be used in everyday life for the betterment of society. many of these contribute to the economy. For example the creation of effective therapies for mental disorders means that people can return to work and contribute more fully to the economy. Fewer people will need to spend time in a mental health institution, relieving pressure on the NHS.

clarity of questions

questions need to be written so that the participant understands what is being asked. they should not be ambiguous and the use of jargon should be avoided. It is also important that questions are not leading as this could bias the participants' answers

strengths of volunteer sample

quick and easy for the researcher as all they need to do is create the advert and wait for participants to respond. as participants are eager to take part they are likely to take the research seriously, reducing the chance of the 'screw you' effect.

how to deal with weaknesses of independent measures

randomly allocating participants to conditions should mean characteristics are evenly spread between the conditions. Use a matched pairs design to make the groups as similar as possible to reduce participant variables

strengths of matched pairs

reduces influence of order effects AND participant variables as different groups do each condition but the two groups are similar

reliability

refers to the consistency of research. If a study is repeated using the same method, design and procedure and similar results are obtained then the results are said to be reliable.

validity

refers to the degree to which a study measures what it claims to and the extent to which findings can be generalised beyond research settings.

improving reliability

reliability can be improved by developing more consistent forms of measurement and by clearly operationalising variables. inter-rater reliability can be improved by training observers in the observation techniques being used and making sure everyone agrees with them. Behaviour categories should be objectively operationalised. this means all observers know exactly what behaviour they are looking for. For example, "aggressive behaviour" is subjective and not operationalised, but "pushing" is objective and operationalised.

weaknesses of field experiments

reliability is lower than in a laboratory experiment because the researcher can't control the environment completely. Therefore it is more difficult to replicate the study to check the consistency of the results. lack of control means that extraneous and confounding variables could affect the dependent variable.

informed consent

researchers should tell potential participants exactly what is going to happen to them in the experiment, so they can make an informed choice about whether they want to take part. children are often considered to be too young to give their own consent, and the consent of their parents or teachers should be gained.

statistical significance

researchers use statistical tests to work out how probable it is that something might occur. research studies should have an experimental hypothesis and a null hypothesis which are being tested. carrying out a statistical test allows us to either accept or reject these hypotheses (depending on the result). a significant result is one where there is a low probability that chance factors were responsible for any observed difference or correlation in the variables tested. psychologists have concluded that for most purposes, the 5% level of significance will be used (p≤0.05). If a result is significant at this level it means there is less than 5% probability that the results were due to chance alone.

weaknesses of questionnaires

response rates can be poor when administered in the absence of the researcher. it can then be difficult to generalise findings as people who bother to return a questionnaire may be psychologically different (e.g. more organised or motivated) to those who don't. the results only represent one type of person. it can be difficult to phrase questions in ways that are not leading or open to interpretation by the participants. Non-anonymous questionnaires may result in participants lying in order to give socially desirable answers. This is particularly a problem when researching a sensitive topic such as sexual behaviour or drug taking activity.

weaknesses of opportunity sample

sample is biased and unrepresentative of the target population because some members are excluded, e.g. asking people who walk past you in town on a week day excludes those at work or college. also biased by the researcher as we tend to approach "people like us".

standard deviation

shows the average distance between each piece of data from the mean. The bigger the number, the greater the variability in the data. You won't be asked to calculate standard deviation in the exam, but may be asked what it shows

frequency polygons (line graphs)

similar to histograms in that the data on the x-axis is continuous. Rather than drawing bars, a line is drawn from where the top of the bar would be in a histogram. A frequency polygon is used when you want to compare two or more frequency distributions on the same graph.

hypothesis: correlation level of measurement: ordinal research design: repeated measures

spearman's rho

observational research ethics

studies that observe people in their natural environment must respect the privacy and psychological well-being of the individuals studied. Participants should either: (i) give their consent to being observed or (ii) be observed in a situation where they would normally expect to be observed by strangers. E.g. in a public park or on a bus.

stratified sample

subgroups ('strata') within a population are identified, e.g. males/females, age groups, ethnicities. participants are randomly selected from each of the strata in proportion to how common they are in the population.

field experiment

take place in the participant's' natural environment. the IV is manipulated. The effect on the DV is measured.

lab experiment

takes place under controlled conditions in an artificial environment the IV is manipulated (changed).The effect on the DV is measured.

independent groups

testing different groups of people for each condition of the experiment.

matched pairs design

testing separate but similar groups of people for each condition. Each member of one group is matched on age, gender, or socioeconomic background to a member of the other group.

repeated measures

testing the same group of people for each condition - the same people are used repeatedly.

natural experiment

the IV is not manipulated, it changes naturally (e.g. studying a culture's aggression before and after the introduction of TV). the effect on the DV is measured.

case study

the detailed study of one individual (e.g. Clive Wearing) or a small group who all have something in common

weaknesses of naturalistic observation

the environment can't be controlled so extraneous variables could be affecting the participants' behaviour. Difficult to replicate as exactly the same situation is unlikely to reoccur.

weaknesses of controlled observation

the environment is artificial so behaviour observed may be unnatural and not reflect how people behave in real life (ecological validity is low).

internal reliability

the extent to which a study or test is consistent within itself, e.g. if a word list is being used to test memory then all the words should be equally difficult to remember.

external reliability

the extent to which a test measures consistently over time, e.g. an IQ test should give the same results if the same person completes it on more than one occasion.

internal validity

the extent to which results are due to the manipulation of the IV and have not been affected by confounding variables such as investigator effects, demand characteristics and bias.

deception

the guidelines say that you should try to avoid intentionally deceiving (lying to) participants about the purpose and nature of the investigation. However in certain circumstances, deceiving your participants is acceptable.

the independent variable

the one that the experimenter manipulates/changes. The different variations of the IV are called the conditions of the experiment.

covert observation

the participants are not aware they are being observed. The observer may have a hidden viewpoint or be behind a two-way mirror. the observer may disguise themselves as a member of the group being observed and record via a secret camera.

volunteer sample

the researcher advertises for participants, e.g. in a newspaper, on a noticeboard or on the internet. The participants are self-selecting.

non-participant observation

the researchers do not become actively involved in the behaviour being studied and observe from a distance.

strengths of stratified sample

the sample should be able to be generalised because all the subgroups in the target population are proportionally represented. as random sampling is used to select participants from the sub-groups of the population it should be free from bias.

weaknesses of volunteer sample

the sample will be biased and unrepresentative as people who like to volunteer for things are likely to be psychologically different to non-volunteers (e.g. more confident or extroverted). as volunteers are eager to please the chance of demand characteristics is increased.

how to interpret correlations

the strength of a correlation can be displayed numerically or graphically. the correlation-coefficient is a number between -1 and +1 which shows the direction and strength of a relationship. the closer the number is to -1 or +1, the stronger the relationship is. 0 represents 'no correlation'.

the mode

the value that occurs most frequently. If two values are equally common the data is bimodal - both scores are the mode. This is the only measure of central tendency which can be used when the data is in categories. However it is not useful when there are several modes.

experimental design

the way in which the participants are 'organised' into groups

content analysis

the way of quantifying data which is non-numerical (e.g. words and pictures). This is commonly used to analyse the media.

strengths of random sample

there is no bias in the selection, increasing the chances of getting a representative sample which can be generalised.

weaknesses of content/thematic analysis

there may be a lack of objectivity because researchers have to make decisions about what behaviour or themes to look for and how to categorise it. it is a description of behaviour rather than an explanation of it ('what' but not 'why).

histograms

these look a bit like bar charts but are different because they are used for continuous data (e.g. scores on a test) rather than that in the form of categories. The scores are placed along the x-axis and the frequency is on the y-axis. There should be no spaces between the bars and their width should be consistent.

features of science: falsification

this is an important part of the scientific process. A theory or hypothesis needs to be able to be empirically tested in order to have credibility. A theory can be accepted as being validated if research evidence supports it but one finding of it not being true leads to its falsification.

probability

this is expressed as a number between 0 and 1, where 0 means an event will definitely NOT happen and 1 means an event definitely WILL happen. for example when you toss a coin the probability of getting heads is 0.5, there is an equal chance that it will or won't happen. Probability can also be expressed as a percentage. To convert a percentage expressed as a decimal to a percentage just multiply it by 100. For example 0.5 becomes 50%. So when you toss a coin there's a 50% probability that you will get heads.

type 1 error

this is when the null hypothesis is rejected as a result of the statistical analysis but actually the null hypothesis was correct. When p≤0.05 there is a 5% of making this error. If result is significant at p≤0.01 there is only a 1% chance of making a type 1 error. We are likely to make a type 1 error when the p-value is too lenient.

random sample

this is where each member of a population has an equal chance of being selected. E.g. give all members of a population and number and use a computer to randomly select numbers to choose the sample.

operationalisation

this means making a variable clear and unambiguous. It is the process of devising a way of manipulating or measuring something so that another person knows what has been done. it allows the study to be REPLICATED.

how do researchers decide if they should break ethics?

through a cost/benefit analysis is carried out by the researcher involved and an ethics committee. - Benefits may include the advantages of the new knowledge and the advancement in understanding offered by it. - Costs might include harm or embarrassment to the participants or the social sensitivity of the findings.

features of science: replicability

to be scientific a study should be able to be replicated in exactly the same conditions. therefore research needs to be tightly controlled and free from extraneous variables.

features of science: objective

to be scientific research should be objective and free from bias. different people should be able to look at the same data independently and draw exactly the same conclusions. gathering quantitative data increases objectivity. variables should be clearly defined and operationalised so it is clear exactly how they were manipulated or measured.

strengths of systematic sample

unbiased because participants are selected using an objective system. Therefore the sample should be representative of the population and able to be generalised.

confounding variables

uncontrolled extraneous variables become confounding variables and mean that firm conclusions can't be drawn about cause and effect.

behavioural categories

used to reduce observer bias. People often interpret the same behaviour differently so it is important to be clear about what the target behaviour actually is. Using behavioural categories means the behaviour is operationalised which allows observations to be made efficiently and objectively. for example 'aggression' can be defined in a number of ways. Researchers who want to measure aggression in an observational study need to create a list of all the possible behaviours that could fit the definition of aggression (e.g. kick, push, punch etc). These can then be used to record the observations by tallying every time each behaviour occurs.

secondary data

was collected for another purpose, e.g. data the researcher collected for a previous study or data collected by another researcher. It also includes Government statistics or data held by an institution such as a hospital. the benefit of this is that it is simpler and cheaper to access pre-existing data. However the data may not exactly fit the needs of the study.

bad implications that psychological research has on the economy

when conducting research psychologists need to keep in mind that ethical considerations come before profit and that psychology should not be used to exploit people. For example drug trials have been successful in identifying drugs which could relieve symptoms of some disorders. However therapy is often a more effective long-term solution. However companies who make the drugs make huge profits from psychiatrists prescribing them, even if this might not be in the patient's best interests.

normal distribution

where most scores are on or around the mean with decreasing amounts away from the mean. There are defining features of a normal distribution: The mean, median and mode are all at the same midpoint. When plotted on a graph there is a symmetrical, bell shaped curve. The dispersion of scores either side of the midpoint is consistent.

negative correlations

where one co-variable decreases as another co-variable increases, for example heating bill decreases as temperature increases.

positive correlations

where one co-variable increases as another co-variable increases, for example number of ice creams sold increases as the temperature increases.

hypothesis: difference level of measurement: ordinal research design: repeated measures

wilcoxon test

external validity

the extent to which the results can be generalised to other settings (ecological validity), other people (population validity) and over time (temporal validity).

target population

the group the researcher is interested in

ordinal data

involves data which can be ranked in order but is not objective. This type of data is often collected via rating scales.

interval/ratio data

involves standardised measurement units like time, weight, temperature, distance.

demand characteristics: social desirability

Participants might just try to 'look good' and behave out of character or not tell the truth. This can be a problem for questionnaires on sensitive issues

qualitative data

non-numerical data, usually in the form of words but can also be pictures or images

semi-structured interviews

a combination of structured and unstructured techniques. There is a set of pre-determined questions that all participants are asked, the interviewer can then add in additional questions depending on their answers.

weaknesses of systematic sample

not truly unbiased/random (the first person on the list is unable to be selected) unless you select a random number and start with this person.

experimental method

involves the manipulation of an independent variable (IV) to see what effect it has on the dependent variable (DV). the aim is to try to determine a cause and effect relationship