Research Methods Flashcards

Question

Evaluating Experiments - Generalise

Answer 1

To apply from one situation or group of people to another - we want to be able to do this so we have a point to the experiment

Answer 2

The extent to which the findings from the study are accurate/true Internal validity - the extent to which the findings from the study actually measure what they claim to measure - i.e is the experiment ‘true’ to what it is studying - more control = more internal validity External validity - they extent to which the findings of the study can be generalised (and are valid/accurate/true) outside of the original context in which the study was conducted - do the results only show us how those specific participants behaved in that setting at that time Ecological validity - are they true in the real world

Answer 3

Consistency Internal reliability - was the research consistent within itself - e.g. did all the participants have the same experience - more control the better External reliability - if the research is repeated, are the results consistent over time?

Answer 4

Does x cause y? More control = more sure you can be

Answer 5

- high control over EVs as the researcher can control the environment and research. It is therefore possible to establish a cause and effect relationship between the IV and the DV - high reliability as they can be easily repeated due to the controlled conditions and so it is possible to check for consistent results

Answer 6

- high chance of demand characteristics as participants are aware they are being studied. Therefore they are likely to pick up on clues as to the nature of the research and may change their behaviour to help/hinder, meaning their behaviour is not natural - low ecological validity as the environment is highly controlled and artificial. It is therefore difficult to generalise the findings to the behaviour in real life situations

Answer 7

- high ecological validity as the experiment is carried out in a real life environment/setting and therefore it is possible to generalise the findings to real life behaviour - less chance of demand characteristics ad participants may be unaware they are being studied. They are therefore unlikely to pick up on clued and their behaviour is more likely to be natural

Answer 8

- difficult to establish cause and effect between the IV and the DV as the research is carried out in a real life/natural environment. There is therefore low control over EVs meaning these could be responsible for the effect on the DV - lacks reliability as there is low control over the research meaning it is difficult to get the same circumstances to repeat the research and check for consistent results

Answer 9

- high ecological validity as there is no artificial manipulation of the IV. It is therefore possible to generalise the findings to behaviour in real life - allows researchers to study areas which would be unethical or impractical to manipulate e.g. comparing sighted to non-sighted children. Therefore valuable for studying certain behaviours

Answer 10

- difficult to establish a cause and effect between the IV and the DV as the researcher deoesn’t have control over or directly manipulate the IV. Therefore there could be EVs responsible for the effect on the DV - lacks reliability as there is low control over the research meaning it is difficultly to get the same circumstances to repeat the research and check for consistent results

Answer 11

- Independent Group Design - Repeated Measured Design - Matched Pairs Design

Answer 12

The different ways in which participants are allocated to the different conditions (IV levels) in an experiment - i.e. how we allocate participants to the experiment and control conditions in our experiment Notes: 1. You only have an experimental design when you have an experiment (not in experimental research) 2. Don’t mix up ‘experimental design’ with ‘type of experiment’

Answer 13

- participants only take part in one condition - they are randomly allocated to either the control condition OR the experimental condition - also called independent measures design

Answer 14

- participant take part in both conditions - they are in the control condition AND the experimental condition

Answer 15

- participants only take part in one condition - however, before being allocated to EITHER the control condition OR the experimental, they are matched with another participant on key variables relevant to the experiment - e.g. people with similar IQs may be placed in opposite group

Answer 16

- there will be no order effects - ppts only complete one condition so they will not become bored or get better (practise effect) - increases the internal validity - less chance of demand characteristics - ppts take part in only one condition so they are less likely to work out then purpose of the study and change their behaviour to help or hinder - this increases the internal validity of the study

Answer 17

- indiviudal differences/participant variables i.e. the different characteristics of the participants such as age, gender may affect the results - lowers the internal validity - requires more participants than as the researcher will need to get two separate groups of ppts to end up with the same amount of data

Answer 18

- requires fewer participants as ppts take part in all conditions - meaning a potentially larger sample can be used - increasing external (Population) validity - participant variables are removed e.g. IQ/Age as ppts take part in all conditions - increasing the internal validity

Answer 19

- results may be affected by order effects - ppts may become bored by the second condition and so do less well or do better (practise effect) - lowers the internal validity - increased chance of demand characteristics as when ppts do the second condition they may work put the purpose of the study and change their behaviour to help or hinder the experiment - lowers the internal validity

Answer 20

- controls for individual differences/participant variables - ppts are matched on characteristics such as age, IQ an d gender that may affect the results - increasing the internal validity - there will be no order effects - ppts are only in one condition so they will not be more bored or get better (the practiser effect) - increases the internal validity

Answer 21

- the most time consuming experimental design as the researcher needs to identify all important participant variables and spend time matching participants - impossible to control all participant variables because only key ones can be matched for - there could be others which could effect the research - this lowers the internal validity - requires more participants as the researcher will need to get two separate groups of ppts to end up with the same amount of data

Answer 22

Some of the weaknesses of the experimental designs we can’t do anything about, such as how time consuming the nature of matched pairs But there are ways we can deal with some of them, through techniques like random allocation and counterbalancing

Answer 23

Used to reduce participant variables in independent measures designs It works my reducing the impact of participant variables through random allocation - the laws of probability suggests groups SHOULD end up reasonably comparable - not necessarily going to work out like that - can have an ‘accidental bias’ and groups end up very different, however that is statistically unlikely

Answer 24

Reduce order effects Also known as the ABBA technique - A stands for condition a (control) - B stands for condition b (experimental) Half of the group will do condition a followed by condition b The other half will do condition b followed by condition b This cancels out the order effects and balances them out

Answer 25

When carrying out research, most psychologists have a target population - a group of people that share a set of characteristics about which the researcher wishes to draw conclusions In an ideal world, you would carry out research on everyone in your target population, but this isn’t actually possible To solve this, you select a representative sample - a group typical of the target population You would then generalise your findings from the sample to the rest of the population You can only generalise your findings if the sample is representative

Answer 26

Every person in the target population has an equal chance of being selected For this, we need a sampling frame - a list of every one in the target population The probability sampling methods are: Random sampling Systematic sampling Stratified sampling

Answer 27

Each person in a given population stands an equal chance of being selected This means it is necessary to have the names of every person in the target population The sample could then be selected by drawing names out of a hair or by entering all the names into a computer random generator and clicking ‘select’ the number of times for participants needed

Answer 28

The researcher first randomly picks the first participant from the population They will then select each n’th participant from the list (also a randomly generated number) - e.g. person 3 is randomly selected as the first participant and the n’th is 10 - that means person 13, 23, 33, 43, 53 and so on until enough participants have been selected

Answer 29

A random and systematic sample is potentially unbiased - as everyone in the target population has an equal chance of being selected, the laws of probability would suggest that the sample should be overall representative of the target population There is also a reduced researcher bias - as is is the selection technique itself that selects the participants rather than the researcher themselves, this means that the researcher is not selecting only participants only - they bare objective sampling techniques

Answer 30

You may still end up with a biased band unrepresentative sample - there is always the possibility of accidental bias - while random and systematic sampling is more likely to produce a representative sample, it is not guaranteed These sampling techniques can be difficult and time-consuming - for example, a complete list of the target population scan be difficult to obtain - selected participants may refuse to take part

Answer 31

It involves dividing the target population into important sub-categories (called strata) and randomly selecting participants within each sub-category in proportion that they occur in the population - e.g. if 40% of the target population is male and 60% is female, then the sample needs to reflect these proportions

Answer 32

A stratified sample is going to provide us with the most representative sample - this is because there is a proportional representation of subgroups There is also a reduced researcher bias - as it is the selection technique itself that selects the participants rather than the researcher themselves, this means that the researcher is not selecting certain participants only - this is another objective sampling technique

Answer 33

Stratified time sampling is very time consuming - this is because you have to identify important subcategories, work out the proportions and then you have to select the participants in those proportions - you also need a sampling frame that can be very hard and time consuming to obtain It will still not provide your with a fully representative sample - even though it will provide you with the most representative sample compared to all of the sampling techniques, it is still likely that there will be some bias because you can’t know what each important subgroup and some will inevitably be left out

Answer 34

It involves the researcher selecting anyone who is available at this time This means it doesn’t require a sampling frame For example, the researcher uses anyone who is around at the time of their study e.g. in the street

Answer 35

It is appropriate to use when the sampling frame is unknown - for example, if you want to do a field experiment in a high street, you wouldn’t have a sample frame An opportunity sample is a confident method of gaining participants because: - it is less costly than random, systematic and stratified sampling because you do not need to source a complete list of your target population - it is less time consuming than stratified sampling because you do not need to identify subgroups and work out proportions

Answer 36

Opportunity sampling is likely to result in a biased sample (an unrepresentative sample) because: 1 - not everyone in your target population will be present when you do your study - e.g. if you got to a high street at 10am on a Monday, large groups will be issuing, such as people at work or school 2 - you approach p-eople who you want to take part which will inevitably lead to researcher bias - they are more likely to select people that they believe will fit in with their hypothesis - this is usually done unconsciously

Answer 37

Participants select themselves This means you do not need a sampling frame For example, in response to an advertisement in a newspaper or a poster

Answer 38

A volunteer sample is a convenient method of gaining participants - this is because you just put the advert up and let people come to you It is a useful way to locate willing participants - this means that participants are less likely to drop out of the research It can be seen as a more ethical sampling technique - you are surer that you have consent from participants as they behave volunteered - this is compared to something like random sampling where participants may6 feel obliged to take part because they were selected

Answer 39

Volunteer sampling is likely to result in a biased sample - it is prone to volunteer bias as it may be that certain types of people are likely to volunteer - e.g. those with the time/specific interest to the area of the study Volunteer sampling may be more likely to result in demand characteristics - this is because the participants are so interested in the study, they are thinking really hard about the aim might be

Answer 40

“Concerns with what is deemed acceptable human behaviour, with what is good or bad, right or wrong with human conduct in pursuit of goals or aims” - penguin dictionary of psychology 1985

Answer 41

Ethical issues occur when their is conflict between what the researcher need to do to conduct a useful and meaningful research and the rights of the participants For example, a researcher might not tell the participants the aim of the research to avoid demand characteristic, but is it acceptable to leave the participants in the dark?

Answer 42

Ensuring that the prospective participants in the study know what they are getting themselves into before they get into it. It involves making participants aware of the aims of the research, the procedures, their rights (including the right to withdraw partway through the study should’ve they wish to) and what their data will be used for From the researcher’s point of view, asking for this may make the study meaningless as participant’s behaviour will not be ‘natrual’ as they know the aims of the study From a participant’s point of view, they should be able to make an informed judgement about whether or not to take part without being coerced of feeling obliged

Answer 43

This involves deliberately misleading or withholding information from participants at any stage in the investigation From a researcher’s point of view, there will be times where deliberately misleading or withholding information about the study is necessary From a participant’s point of view, lying is wrong. Not only does it prevent them from giving full valid consent, it can also lead to them to see psychologists as untrustworthy, which might mean they are less likely to take part in psychological research in the future

Answer 44

Participants should be protected psychologically (e.g. stress, humiliation or anxiety) and physically (pain). They must not be placed at more risk than they would be in every day life From the researcher’s point of view, studying some of the more important questions in psychology may involve a degree of risk or harm (psychological or physical) to participants. It can be hard to predict what this harm might include, so it is difficult to guarantee protection from any risk of harm From a participant’s point of view, nothing should happen to them during a study that causes harm and they should leave the study in the same state as they ere before hand

Answer 45

This refers to the right participants have to control information about themselves. This extends to the area that the study took place. From a researcher’s point of view, it may be difficult to avoid invading this when studying participants without their awareness, like asking about personal thoughts and opinions in a field experiment From a participant’s point of view, they do not expect to be observed by others in certain situations, like in their home, nor do they expect to be forced into revealing information about themselves that they do not wish to share

Answer 46

This refers to our right, enshrined in law under the General Data Protection Regulation, to have any personal data protected and kept anonymous From the researcher’s point of view, it may be difficult to protect this because they may wish to publish the findings. Even if the researcher can guarantee anonymity, by withholding the names of participants ect, it may still be obvious who has been involved in the study, like if has been filmed or if it is a very specific group of participants From a participant’s point of view, this is a legal right and it is only acceptable for personal data to be recorded if it is not available in any format that identifies the participants

Answer 47

Participants should be issued with a consent form that details all relevant information that might affect their decision to participate If they agree to this, they need to sign a form If children under the age of 16 are required for the investigation, they need parental consent too However, using a consent form for informed consent might cause issues because not many people actually read terms and conditions

Answer 48

- thank the participants - the aim of the research (if you are withholding the aim you may not include this - in this case you will be gaining consent but it won’t be valid) - an outline of the procedure and how much of their time it will take - explain the ethical issues which have been accounted for, especially the right to withdraw (who they should speak to if the feel uncomfortable or want to withdraw) and confidentiality Remind them they can ask questions Include space where they can sign and date to show they consent (this is essential, it is your record that valid consent was sought)

Answer 49

There are alternative ways of getting consent to over come getting consent to overcome the problem that getting people to sign a consent form might make the study meaningless if they know the aim

Answer 50

Rather than getting consent from the participants themselves, a similar group of people are asked if the study is acceptable If this group agrees, then consent of the original group is presumed This might be a problem though because everyone has different ethics and morals The group asked might not have gone through a certain traumatic event that someone in the other group has - everyone has different triggers

Answer 51

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 However, they may have consented to an experiment that had minor levels of deceit, whereas the actual study might have very high levels of deception Consent should be given every time Participants may be expecting deception, which could lead to demand characteristics

Answer 52

Participants are asked for their consent after the study, during the debrief They may not have been aware of their participation or they may have been subject to deception However, they might not have been happy with being lied to They also might not have taken part if they knew the real aim

Answer 53

*where possible, deception should be avoided completely* At the end of the study, participants should be given a full debrief This should involve: Making the participants aware of the true aims of the study and any details they were not supplied with/deceived about In addition, the debrief should also tell participants what their data will be sued for and give them the right to withhold their data if they so wish

Answer 54

Research should be planned to ensure participants are not exposed to any more harm than they would be exposed to in everyday life If harm is inevitable participants should be made aware of this from the start of the research The participants should be reminded they have the right to withdraw from the investigation at any point In the debrief, they should be reminded that their behaviour is normal In extreme cases, counselling may be needed after the investigation, which the researcher should provide access to

Answer 55

Research should be planned to ensure participants are not being studied in locations in which they could reasonably expect privacy - e.g. their homes, changing rooms, shops, public toilets etc Participants should be reminded they have the right to withdraw from the investigation at any time

Answer 56

Researches should maintain anonymity in their research - this means recoding no personal details that could potentially indentify the participants - e.g. name, location - you should refer to participants using numbers or initials when writing up the investigation Participants should be reminded of the fact that their data will be preceded throughout the process and told that identifying information will not be shared to others

Answer 57

The BPS have published the ‘Code of Human Research Ethics’ These tell psychologists which behaviours are acceptable and give guidance on how to deal withy certain ethical issues that may arise in research Psychologists are to consider these ethical issues when planning and conducting research The BPS regularly updates its guidelines: 1st edition - 1978 3rd edition - 2017 If these guidelines are not followed they can be punished by the BPS can decide to ban the person from practising as a psychologist - expelled from the BPS - however it is not a legal issue The BPS are based on the 4 main principals as: - respect - competence - responsibility - integrity

Answer 58

Nowadays, before any type of research can take place, there must be an ethics document submitted to an ethics committee for approval What are they? - set up in research institutions to assess the cost-benefit analysis of research proposals Who is on them? - members of the committee include experts and lay people How do they manage risk of ethical issues? - request that the researcher changes the study’s design or procedure, or in some cases deny approval for the study What are the benefits of them? - a second check on ethical issues (and BPS guidelines), ensure rights are met and involve individuals who represent the participants (the lay people, aka non psychologists)

Answer 59

*non-experimental research* An experiment is not always the most suitable way of stuffing human behaviour There are a number of non-experimental research methods available to psychologists - i.e. research methods that do not have an IV One important non-experimental research methods are observations An observation involves watching a participant. And recording behaviour for later analysis

Answer 60

Regardless of the type of observation we might use, all observations have the following general strengths and weaknesses: S - they capture what’s people actually do - what people say vs do is actually very different W - Thor is the possibility of observer bias - our interpretation of what we can see can be influenced by our expectations W - you cannot establish cause and effect with observations

Answer 61

Naturalistic observations involve watching and recording of spontaneously occurring behaviour in the participant’s own natural environment - there is no interference by the observer - they avoid intrusion For example, Anderson, in the 70s, was interested how far young children would run from their parents in a park, before they returned to them (max 15ft)

Answer 62

Controlled observations involve the watching and recoding of behaviour within a structured environment in which the conditions are manipulated by the researcher For example, Ainsworth and the strange situation used to discover the quality of attachment between a child and their caregiver

Answer 63

Naturalistic observations have higher external validity because they record real behaviour in a natural environment

Answer 64

Controlled observations have lower external validity because they observe behaviour in a highly manipulated setting, and so there is no guarantee that they would behave like this in the real world

Answer 65

Controlled observations are very replicable as the situation is very controlled, so the experiment can be easy easily replicated to prove the reliability of the data It allows you to check for consistency and outliers

Answer 66

Naturalistic observations are not very replicable as they are naturally occurring events, so the environment is very hard to be repeated to prove the reliability and consistency of the data You can’t control the real world

Answer 67

Covert observations involve watching and recoding of behaviour without the knowledge or awareness of the participants and the observations is hidden from the participants For example, you may observe a therapy session or police investigation behind a one-way mirror, under cover

Answer 68

Overt observations involve watching and recording of behaviour with the knowledge or awareness of the participants and the observer is clearly visible to the participant For example, a classroom observation from a higher up member of staff from the school

Answer 69

Overt observations are more likely to produce demand characteristics as the person will want to behave desirably towards the observer, and therefore decrease internal validity as the behaviour isn’t true

Answer 70

Covert observations are less likely to produce demand characteristics as the person does not known they are being observed and can’t know to change their behaviour, increasing internal validity

Answer 71

Covert observations are more likely to produce ethical issues as you do not know you are being watched, and therefore you cannot give consent. This may cause privacy issues too, as you may be being observed in a place you would expect privacy, such as a therapy session, which should also be confidential

Answer 72

Overt observations are loess likely to cause ethical issues as you know you are being watched and therefore you can give consent

Answer 73

Participant observations involve the watching and recording of behaviour with the observer becoming a member of the group whose behaviour they are watching and recording For example, someone joined in with a group of football hooligans “On being sane in insane places’ - seeing how staff treat patients in psychiatric hospitals in the 70s

Answer 74

Non-participant observations involve the watching and recording of behaviour bewitch the observer remaining outside of the group whose behaviour they are watching and recording For example, watching children in a play ground

Answer 75

Participant observations are more likely to increase insight and understanding of the behaviour as you are immersing yourself and experiencing the behaviours first hand, increasing the internal validity

Answer 76

Non-participant observations are less likely to increase insight and understanding, as you are watching from outside and not fully experiencing the behaviour, decreasing internal validity

Answer 77

Participant observations are more likely to decrease observer objectivity as being a part of the group may cause you to build close relationships with who you are observing, potentially causing bias and decreasing internal validity It is harder to be neutral

Answer 78

Non-participant observations are more likely to increase observer objectivity as you are watching the group from a distance and not having any interactions with those who you are observing, decreasing the chance of bias and increasing the internal validity

Answer 79

When planning an observation, you have a number of design decisions to make Obviously the first steps are too decide which behaviour(s) you are going to observe and which type of observation it will be Once those decisions are made, you now have further design decisions to make

Answer 80

Will you record every behaviour of interest? This is known as an unstructured observation S - appropriate when observations are small in scale and involve few participants, e.g. observing interactions between a couple and a therapist in a marriage counselling session S - richness and depth of detail in data collected (qualitative data) W - there may be too much going on in an observation to record it all Or will you focus on recording specific behaviours only? This is known as structured observations S - recording data is easer and more systematic (quantitative data) W - lack of richness and depth of data

Answer 81

If you decide to do a structured observation, you then need to break the target behaviour up into a set of behavioural categories - way the target behaviour might be observed This is an operationalised checklist of all the ways in which the target behaviour may occur This means: - we clearly and unambiguously state what it is we are observing - each item on our checklist must be observable and exclusive (i.e. not overlap) - different observers would need to each be able to observe the same behaviour and ‘check it off’ the same way

Answer 82

Once we have produced our behavioural categories, we must then choose which way we will sample our observations *sampling in this way is not the same as ‘sampling techniques’* It’s refers to deciding when and how often to record the behaviour There are to ways to do this: - event sampling - time sampling

Answer 83

Counting the number of times a particular behaviour (the ‘event’) occurs across a whole event For example, event sampling affection during a date would mean counting every time the couple on the date showed the agreed on behavioural categories for affection, for the enTire dates

Answer 84

Counting the number of times the behaviour occurs within a pre-established time frame For example, time sampling affection during a date would mean recording what is happening every 10 minutes

Answer 85

A non-experimental method that measures the strength and direction of a relationship or link between two co-variables These can be positive or negative There is no IV and DV

Answer 86

Involves measuring two or more co-variables Doesn’t not involve and IV or DV Can be quickly used to analyse relationships in large amounts of data Identifies the direction of the relationship - positive, negative or zero Identifies the strength of the relationship - weak, moderate or strong

Answer 87

As one variable increases, so does the other one For example, as your narcissism score goes up, so too does the number of photos you post on social media

Answer 88

As one co-variable increases, the other decreases Form example, as your introversion score goes up, then number of social events you atten goes down

Answer 89

There is no connection between the two variables For example, there is no connection between your IQ and the amount of tea you drink

Answer 90

Sometimes relationships are more complex than positive or negative correlations For example, the Yerkes-Dodson law or arousal - where the performance is low during low arousal, highest for medium arousal and back lowest at highest arousal

Answer 91

Correlation is a non-experimental research method In an experiment: - the researcher manipulates the IV - they measure the DV This allows them to develop a cause and effect relationship In contrast, in a correlation there is no IV being manipulated and no DV being measured We instead have two co-variables that we are looking to see if there is a link or association between - we are not looking for a cause or effect relationship For example, if there was strong correlation found between aggression in parents and aggression in children, was can’t assume that the aggression in the parents is the cause of th aggression in the children as it could actually be the other way around - the aggression in the children may be the cause of the aggression in the parents - there may be other factors

Answer 92

Writing a directional alternative hypothesis to a correlation study: - identify the operationalised co-variables in the scenario - you will need to include these in any hypothesis - say that there will be a significant positive or negative correlation/relationship between the two named operationalised co-variables Writing a non-directional alternative hypothesis to a correlation study: - say that there will be a significant correlation/relationship between the two named operationalised co-variables - but do not state it will be negative or positive Tips on writing a null hypothesis to a correlation study: - say that any correlation/relationship between the two names operationalised co-variables will be due to chance

Answer 93

A useful preliminary tool for research - often used as a starting point to assess possible relationships or links between variables before researchers commit to an experimental study A useful research method to use when it is not possible or not ethnically to carry out an experimental study into the area

Answer 94

Correlations cannot establish cause and effect - for example, imagine we are studying the link between stress and illness and we have found a positive correlation between the two - but we don’t know ‘which comes first’ - is stress causing illness or is illness causes stress - there so may be intervening ‘hidden’ variables, e.g. poor diet causing illness and not the stress This means that correlations can be misinterpreted by the media and society when a link has been found between two variables - some may assume a conclusion can be made about the causes for the relationship (e.g. one co-variable caused a change in the other co-variable), which can then be misused by the public to support or contradict an argument - for example, the media may present correlation findings as casual facts

Answer 95

Self-report techniques describe methods of gathering data where participants provide information about themselves This is done by asking participants questions They are non-experimental methods They are often used to investigate: - experiences - beliefs - attitudes - thoughts or feelings of individuals

Answer 96

Researcher David Buss wanted to investigate if people ever fantasise about killing someone He simply asked 5,000 people about this He couldn’t have used another research method to study this aim because you wouldn’t have been able to find out peoples’ thoughts without asking them

Answer 97

You can asked closed questions These are questions for which the researcher has determined the range of possible answers There are different types of closed questions: - likert scale - rating scale - fixed choice option

Answer 98

You indicate your agreement/disagreement with a provided statement on a scale of ‘strongly agree’ to ‘strongly disagree’ - extent to which you agree/disagree

Answer 99

Gets respondents to identify a value that presents their strength of feeling about a particular topic - e.g. on a scale of 0 to 10, with 1 being ‘no pain’ and 10 being ‘worst pain possible’, rate your pain

Answer 100

This includes a list of possible options and respondents are required to indentify those that apply to them

Answer 101

Quantitive data - numerical E.g. Buss found that 91% of men and 84% of women had experienced ‘at least one vivid fantasy of killing someone’ These findings can be illustrated in graphs - in this case, these results could be illustrated in a bar chart

Answer 102

Quantitive data is quick to analyse, which means you can gather large amounts of data if you used closed questions in your self-report method This is an advantage because: - it is quick to analyse - you can gather lots of it quickly - you can afford to send close questions out to a lot of people, meaning you can generalise findings as the sample is bigger and therefore more representative

Answer 103

You are not going to get in depth insights into the question you are asking - you are missing out on a lot of the rich, informative qualitative information

Answer 104

Open questions are your asking questions within no prompts to answer the question - you have to go into greater depth and detail with your answers Open questions will provide you with qualitative data - this is non-numerical, word based data

Answer 105

You are getting more detailed, in depth answers and explanation for thoughts - you are getting that rich, informative data

Answer 106

It takes a long time to process the data/information so you can only use small groups to gather data - this means the data will be less representative and therefore any findings will be harder to generalise

Answer 107

Two examples of self-reports techniques are questionnaires and interviews - (you can also use the use of diaries) Questionnaires and interviews can be used as ‘stand-alone’ research methods, although they are often used as a way of gathering further information about participants’ responses in other types of studies

Answer 108

With any self-report method, there isn no guarantee that participants will answer truthfully - they may answer in a way that makes them ‘look good’ This can create a bias known as social desirability bias This is a problem because it’s means we are not gathering valid (accurate) data

Answer 109

Questionnaires are written methods of gaining data from participants The researcher does not need to be present when a participant completes a questionnaire as they can be carried out online or distributed via post or email

Answer 110

Because the researcher does not have to be present while the participants are completing them, and they an be sent out online, it means we can study people who are geographically distinct This is an advantage because it means we are going to get a comprehensive system - the more data we can get, and from then more places around the world you can gather this, the more we can generalise our findings

Answer 111

Questionnaires rely on people completing and returning them to the researcher - this means that you may not actually get a sample as large as you originally planned - this then in turn means you cannot generalise your findings as much as you had intended Questionnaires require a certain level of literacy - this means that you will excluded people who are not to that literacy level - therefore your sample won’t be as representative

Answer 112

1. Type of data you want - quantitive or qualitative 2. Type of questions required - open or closed 3a. Open questions start with ‘describe’ or ‘explain’ etc 3b. Closed questions starts with ‘what rating’ etc 4. Include distract or questions (so won’t figure out aim and cause demand characteristics) 5. Decide an order - e.g. easy questions to begin with 6. Carry out a pilot study

Answer 113

Interviews involve direct verbal questioning of participants by the researcher There are three broad types of interviews: 1 - Structured Interviews: - list of (usually closed) questions, and you do not deviate 2 - Unstructured Interview: - knowing you want to discuss a general topic, but not having planned questions 3 - Semi-Structured Interview: - Structure to the majority of the interview - you have planned out questions but deviate from them

Answer 114

Issues to consider: - what sort of interview is it - do you’ve need a standardised interview schedule? - how will you record the interview response - e.g. note taking, record the interview for later review etc? - one-to-one or group interview? - location of interview and layouts of the room? - starter questions? - you need to put them at ease - ethics?

Answer 115

Structured interviews, like questionnaires, are straight forward to replicate due to their standardised format The standardised format also reduces differences between interviews b There is much more flexibility in an unstructured interview - the interviewer can follow up points as they arise and is more likely to gain insight into the world view of the interviewee, including eliciting unexpected and questions A skilled and experienced interviewer should be able not create a rapport between the researcher and participant, allowing for the discussion of personal and sensitive topics

Answer 116

Given the nature of structured interviews, it is not possible for interviewers to deviate from the topics or explain their questions - this will limit the richness of the data collected as well as limit unexpected information Analysis of data from and unstructured interview is not straight forward as the researcher may have to sift through much irrelevant information, and drawing firm conclusions may be difficult There is a risk of bias, including both interviewer bias, where the interviewer might control the discussing and interpret responses in the way that fits with their expectations, and also social desirability bias when the interviewees lie to ‘look good’

Answer 117

Cyril Burt was an eminent British psychologist who in the 1950s published research that was used to show that intelligence is inherently He started withy 21 pairs of twins, who were raised apart - this was later doubles to 42 pairs He found a correlation coefficient of +0.771 in intelligence - this highly suggest intelligence is genetic Burt replicated the research in 1966, increasing the sample size to 534 sets of twins and he reported an identical correlation coefficient This consistency was suspicious and he was accused of inventing data This was confirmed when a reporter from The Sunday Times attempted to find 2 of Burt’s research assistants - they couldn’t be found because they didn’t exist

Answer 118

This discovery was worrying because the findings of the research had been used to shape social policy - such as 11+ exams to filter people to the correct educational facilities, such as grammar and public schools

Answer 119

In light of such prominent cases of prominent misconduct, John et al (2012) surveyed 2155 psychologists, asking them anonymously to report their involvement in questionable research practises - 70% said they cut corners is reporting data (1508.5 psychologists) - nearly 1% admitted to falsifying data ( nearly 21.55 psychologists) Conclusion of this research: - a surprisingly high percentage of psychologists admit to having engaged in questionable research practises

Answer 120

Peer review is the assessment of scientific work by others who are experienced in the same field to ensure that the papers that are published are valid and unbiased

Answer 121

Without peer review we do not know what is mere opinion and speculation or fraudulent from what is rigorously researched data This is why peer review is an essential part of the scientific process

Answer 122

*Peer Views Are Surely Of Benefit* 1 - is it acceptable to PUBLISH? 2 - is it a VALID piece of research? 3 - was the methodology APPROPRIATE? 4 - are the findings of SIGNIFICANT interest? 5 - does it show ORIGINALITY? 6 - does it BACK UP other findings?

Answer 123

As research is paid for by various government and charitable bodies, there is a duty is a duty to spend this money responsibly - there was a £5.6 billion budget for science researching in 2015-16

Answer 124

All universities science departments are expected to conduct research and this is assessed in terms of quality, using the Research Excellence Framework (REF) Future funding for the departments depends on receiving good ratings from REF peer reviews

Answer 125

A means of preventing incorrect or faulty data entering the public domain

Answer 126

While the benefit of peer review appears clear, certain features of the process can be criticised “Peer review is slow, expensive, profligate of academic time, highly subjective, prone to bias, easily abused, poor at detecting gross defects and almost useless at detecting fraud” - Richard Smith, editor of the British Medical Journal

Answer 127

When the researcher gathered data in numerical form which can be put into categories, or in rank order, or measured in units of measurement This type of data can be used to construct graphs and tables of raw data For example: - data in experiments - closed answer questionnaires - tallies oh behaviours in an observation - content analysis - number of occurrences of categories

Answer 128

Data which is descriptive rather than quantified or counted, therefore it is observed or reported rather than measured For example: - case studies - open answer questionnaires - description of behaviours seen in observations - content analysis

Answer 129

Can produce graphs from the data More likely to be objective Easy to analyse as averages and ranges can be produced Results from different studies can be compared Easier to draw conclusions from the data

Answer 130

Can over simplify complex behaviour Phenomena can be forced to fit a set of data

Answer 131

Represents the complexity of human behaviour Can gain access to thoughts and feelings Participants have freedom of expression Data is rich and detailed

Answer 132

Difficult to draw conclusions and detect patterns Can be affected by subjective analysis

Answer 133

Data collected or observed directly by the researcher from participants which is specifically for the of the research Examples include: - data collected in an experiment - interviewing - making your own questionnaire

Answer 134

Data collected by someone other than the researcher of the study and purpose was for something other than the aims of the study Examples include: - government statistics - journal articles - literature review - meta-analysis (this refers to a form of research method in which a number of studies are identified which have the same aims/hypothesis. The results of these studies can be pooled together and a joint conclusion produced)

Answer 135

It fits with the research Authentic data obtained from participants themselves for the purpose of a particular investigation Research methods (such as questionnaires and interviews) can be designed in such a way they target the information that the research requires

Answer 136

Requires time and effort on part of researcher - e.g. conducting an experiment requires considerable planning, preparation and resources You may not have the funding to carry out primary research

Answer 137

Inexpensive and easily accessed, requiring minimal effort - the researcher may find that the desired information already exists and so there is no need to conduct primary data collection

Answer 138

May be substantial variation in the quality and accuracy - information may appear promising and valuable at first, but upon further investigation, it may be incomplete or outdated They content of the data may not quite match the researcher’s needs or objectives - this may challenge the validity of any conclusions

Answer 139

These are measures of averages The 3M’s (mean, median and mode) Calculating the Mean: - add all up and divide by how many values there are Calculating the Median: - order the values and find the middle value Calculating the Mode: - most often/common

Answer 140

*least sensitive measure to most sensitive measure* 1 - mode 2 - median 3 - mode

Answer 141

*most common value in the data set* S - the mode is the simplest measure of central tendency to work out S - it is a good measure to use when the set of data contains exceptionally high or low scores as it is unaffected by extreme scores L - it is the least sensitive measure of central tendency L - it also tells us nothing about the other scores n L - a set of data might also have no mode or be bimodal

Answer 142

*put all the values in the data set in order and central value = media - if there are two values in the centre, add the numbers together and divide by 2* S - the median is easier than the mean to calculate S - the median is a good measure to use when the set of data contains exceptionally high or low scores as it is unaffected by extreme scores L - as it ignores most of the scores, the median is not always representative L - it does not work well with small sets of data

Answer 143

*add up all the values an divide by how many there are* S - as the mean makes use of all of the data, it is a very powerful measure of central tendency S - it is a good measure to use when a set of data contains no exceptionally high or low values L - if there is an exceptionally high or low value, then the mean is not a good measure of central tendency to use as it can then be misleading

Answer 144

These are measures of how spread out your data is The two measures of dispersion are: - range - standard deviation These are useful to calculate in addition to a measure of central tendency - this is because it is useful to do some additional descriptive statistics as it would allows us to distinguish between sets of data when all measures of central tendency are all of the same

Answer 145

*the highest value subtract the smallest value* S - the range is a simple calculation compared to the standard deviation L - it only takes into account the two most extreme values and this may be unrepresentative of the data set as a whole L - the range does not indicate whether the values are closely grouped around the mean or spread out - we need the standard deviation to know this

Answer 146

The standard deviation tells us the average amount all scores deviate from the mean - i.e. how spread scores are from the mean value The larger the standard deviation, the greater the dispersion, or spread, of scores around the mean - this may suggest there are a few anomalous results (they are not consistent) A smaller standard deviation reflects the fact the data is tightly clustered around the mean, which might imply that all participants responded in a fairly similar way - this means the results are more consistent While this isn a little more time-consuming to calculate, this is the most sensitive measure of dispersion, using all the data available, and there are no important disadvantages

Answer 147

Graphical representation refers to how psychologists summarise their data pictorially Psychologists may choose to do this as it is easier to interpret than a list of raw data

Answer 148

Summary tables are really useful ways of providing an overview of data to the researcher Note that summary tables are exactly that - they summarise the descriptive statistics gathered in a piece of research - they are not tables of raw data/raw scores - i.e. they may include means and standard deviation etc

Answer 149

Line graphs show information that is connected in some way - e.g. change over time

Answer 150

Bar charts display data by using bars of different heights that illustrate the frequency of data occurrences in separate conditions (groups) It is the height of the bars that provides us with information Bar charts must: - have a title - have gaps between the bars - have all bars labelled - have all bars the same width - have a y-axis with regular intervals

Answer 151

Histograms are a special form of bar chart They differ from ‘standard’ bar charts in three different ways - there are no gaps between the bars because the data is continuous rather than discrete - it is the area of the bar rather than the height of the bar that gives us detail about the size of the category - in a histogram, both the x-axis and the y-axis have a scale (compared to a bar chart where only the y-axis has a scale)

Answer 152

Scattergrams are used to show associations between two sets of data - two co-variables

Answer 153

Frequency distribution that has a classic bell shaped curve The mean, median and mode are all in the exact midpoint Most scorers will be closely distributed near the midpoint The distribution of a frequency is symmetrical around the midpoint Y-axis=frequency X-axis=item of interest Justification of use: - for data that is evenly distributed

Answer 154

Majority (mode) of scores have low values However, there is a ‘tail end’ of high scores These extreme high scores then skew the mean to the right hand side of the mode - i.e. the mean is greater than the mode

Answer 155

Majority (mode) of scores have high values However, there is a ‘tail end’ of low scores These extreme low scores then skew the mean to the left hand side of the mode - i.e. in a negative skew the mean is smaller than the mode

Answer 156

When you have completed a piece of psychological research, you can do one of two things: 1 - accept the H1 and reject the H0 - your result was significant, i.e. was not due to chance 2 - accept the H0 and reject the H1 - your result was insignificant, i.e. was due to chance But at this point in time, with just your descriptive statistics, you don’t have enough information to know whether your results are significant or not This is why we need inferential statistics as these allow us to draw conclusions based on the probability that our results could have arisen by chance

Answer 157

We can never be certain 100% certain in a result - there is always a possibility a result is due to chance or luck or a fluke So in psychology, to decide if a result is significant, we want to be 95% in a result before before we can accept H1 and reject H0, and so we will allow a 5% possibility that the result is due to chance In psychological research, this 5% is expressed as a decimal - you will therefore hear psychologists talk about using a level of probability (or a level of significance) of 0.05

Answer 158

One of the most simple inferential statical tests is the sign test - this will allow us to draw conclusions about whether or a result is significant or not To use the sign test, we need to be looking for a difference (rather than a correlation) between two sets of data - i.e data from an experiment We need to have used a repeated measures design We need to have data that is organised into categories and every response can only fall into one of these categories - this is a type of data known as nominal data

Answer 159

In this example, we are seeing if the levels of anger from families of victims of killers change after a trial You would produce a table of the results, saying whether the anger levels of the families increased, decreased or stayed the same You then give these results a symbol, such as +,- or = Omit any that stayed the same Add up how many times each symbol appears Figure which symbol occurred the least number of times and how many times it occurred - this number gives us our observed value of S (or Sobs) We now compare Sobs to Scrit (the critical value we obtain from a sign test critical value table) The critical value table will always be obtained Find the n(number of data sets) - remember some might be omitted and remember your level of significance (assume 0.05 unless stated otherwise) Cross reference the the n and level of significance to find your critical value If the observers value is smaller than the critical value, then you have a significant result and vice versa

Answer 160

The strength and direction of correlation relationships can be analysed by using statistical tests, which calculate the correlation coefficient Thus is a numerical value between -1 and +1 and it’s this value that tells us about the strength and the direction of the relationship between co-variables What the correlation coefficients tell us: Perfect = +1, -1 Strong = +0.9,+0.8,+0.7,-0.9,-0.8,-0.7 Moderate = +0.6,+0.5,+0.4,-0.6,-0.5,-0.4 Weak = +0.3,+0.2,+0.1,-0.3,-0.2,-0.1 Zero = 0

Answer 161

1-2 marks: A case study involves the in-depth study cover time (longitudinal)of a ‘case’, which is usually a single individual or small group Higher mark questions: An in-depth study of an individual, group or event Often used ton study an exceptional or unique set of circumstances that could not be tested using experimental research methods Often over a long period of time - longitudinal Often uses more than methodology - e.g. interviews, observations, IQ tests Can collect qualitative (usually) and quantitative

Answer 162

When you have completed a piece of psychological research you can do one of 2 things: 1. Accept H1 and reject H0 - your result was significant (not due to chance) 2. Accept H0 and reject H1 - your result was not significant (was due to chance) But at this point in time, with just your descriptive statistics, you don’t have enough information to know whether your results are significant or not This is why we need inferential statistics as these allow us to draw conclusions based on the probability that our results could have arisen by chance

Answer 163

This is done by answering 3 questions about the research and consulting the choosing a statistical test table: 1. Is the research looking at a test of difference between conditions or a correlation between co-variables? (Is it an experiment or correlational study) 2. If it was a test of difference, which experimental design was used? 3.what level of measurement was used? (What type of data was collected)

Answer 164

Is the research looking at a difference between conditions or a correlation between 2 Co variables? This should be obvious from the wording of the aim/hypothesis provided in the scenario Words such as ‘difference’ with reference to an experiment or 2 conditions/groups = test of difference Word such as ‘correlation’ or ‘relationship’ or ‘link’ = test of correlation (in this context, a test of correlation can include investigations looking for an ’association’)

Answer 165

If it is a test of difference, which experimental design was used? A remainder - experimental design refers to how participants are allocated to the experiment and control conditions inn a test of difference There is no experimental design in a test of correlation

Answer 166

Hat level of measures was used? - quantitative data - type of numerical data collected Not all quantitative (numerical) data is the same There are different kinds of ‘numbers’ that researchers may gather Therefore the types of quantitative data very in how precise they are ‘Levels of measurement’ refers to the differences in precision There are 3 levels of measurement: - nominal - ordinal Interval/ratio (Least precise to most precise)

Answer 167

Most basic level of measurement Described as category data as it is basically a head count - it tells us how many people are in each group It is discrete in that one item can only appear in one category It5 gives us very little information For example, 100 smokers who owed to quit smoking were followed for 6 months and after the 6 months they were asked if they had continues to not smoke or if they had relapsed This is nominal data because these are discrete and separate categories - they can only pick one category

Answer 168

Data that is ordered in some way - scale data The intervals between each rank are not equal, e.g. giving a score of 4 out of 5 for how hard it was to quit smoking does not mean you found it twice as hard to quit smoking as someone who gave it a 2 - everyone’s interpretation will be unique - subjective scale of difficulty For example, on a scale of 1-5, with 1 being easy and 5 being hard, how hard have you found it to quit smoking ? This is ordinal data because it is based on subjective opinion rather than objectives measures

Answer 169

Data that is measured using fixed units of equal measurements This means that there is an equal gap (interval) between each unit on the scale being used This type of data is the most precise Interval data can go into minus numbers (e.g. temperature, money) Ratio data starts at zero For example, if you relapsed, how many days was it between quitting and smoking again? This is ratio data because you can’t be quit for a negative number of days

Answer 170

Test of difference Independent measures: - nominal data - chi squared - ordinal data - Mann-Whitney - interval or ratio data - unrelated t-test Repeated measures/matched pairs: - nominal data - sign test - ordinal data - Wilcoxon - interval or ratio data - related t-test Test of association or correlation: - nominal data - chi squared - ordinal data - spearman’s rho - interval or ratio data - Pearson’s r

Answer 171

> = grater than < = less than = = equal to <_ - less than or equal to >_ = less than or equal to

Answer 172

Once a statistical test has been calculates, you need to know how to interpret that results to allow you to know whether you can: 1. Accept H1 and reject H0 - your result was significant 2. Accept H0 and reject H1 - your results was not significant But before we can do that, we need to consider levels of significance and probability It is important to be aware that all any inferential test van tell us is how likely it is that’s pour hypothesis is ‘true’ or ‘correct’ We can never be 100% certain that our hypothesis is just how likely it is that our hypothesis is ‘true’ or ‘correct’, as there is always the possibility that true results might be due to chance So, before we interpret the results from the inferential test, we need to decide how certain do we want to be in the results before we accept the H1 and reject the H0 This is the equivalent to saying we will allow at most a 0.1% portability that our result is due to chance - this is our level of significance In psychology, the level of significance is expressed as a decimal - this would mean that in this case, our level of significance is 0.001 This is written as: p<_0.001 The level of significance that is usually always used in psychology is p<_0.05 - this means we are 95% certain in our results and there is a 5% chance that our results are due to chance

Answer 173

A level of significance of p<_=0.05 gives us a good balancer between Type I and Type II errors Type I Errors: These occur when our p is too lenient E.g. p<_=0.33 would mean there is a 33% chance of results being a fluke This gives us a false positive because we accept our hypothesis when in actual fact the null hypothesis was true So we claim something to be the case when it is not the case Type II Errors: These occur when our p is too strict E.g. p<_0.01 would mean there is a 1% chance of results being a fluke This gives was a false negative because we accept our null hypothesis when in actual fact the hypothesis was true So we claim something to not be thr case when it is in fact the case Type II errors are more common to Type I errors

Answer 174

Used when: - test of difference - independent measured - ordinal data To calculate critical value: - was the H1 directional or non-directional - what was the level of significance - N1 ( number of participants in first condition) and N2 (number of participants in second condition) If the observed value is smaller than the critical value than you have a significant result

Answer 175

Used when: - test of difference - repeated measures/matched pairs - ordinal data To calculate this, you will need: - was the H1 directional of non-directional - what level of significance to use - N(number of data sets) If the observed value is smaller than the critical value then you have a significant result

Answer 176

Used when: - test of difference - independent design - nominal data To find the critical value, you need to know: - what level of significance to use - df: degrees of freedom, calculated by multiplying (rows-1)x(columns-1) For this test, if the observed value is bigger than the critical value, you have a significant result

Answer 177

Used when: - test of difference - independent design - interval/ratio data To find critical value you need to know: - was H1 directional or non-directional - what level of significance to use - df: degrees or freedom, calculated by adding the number of people in group 1 to the number of people in group 2 and then subtracting 2 For this test, if the observed value is bigger than the critical value then your have a significant result

Answer 178

Used when: - test of difference - repeated measures/matched pairs - interval/ratio data To find the critical value you need to know: - was H1 directional or non-directional - what level of significance to use - df: degrees of freedom, calculated by grinding the number of participant there were and subtracting 1 For this test, if the observed value is bigger than the critical value then you have a significant result

Answer 179

Used when: - test of correlation - ordinal data To find the critical value you need to know: - was H1 directional of non-directional - what level of significance to use - n(number of data sets) For this test, if the observed value is bigger than the critical value then you have a significant result

Answer 180

Used when: - test of correlation - interval/ratio data To find the critical value, you need to know: - was H1 directional or non-directional - what level of significance to use - df: degrees of freedom, calculated by finding the number of data sets there were and subtracting 2 For this test, if the observed value is bigger than the critical value then you gave a significan result

Answer 181

The word reliability makes us think of dependent, trustworthy and consistent In general times, we can say that something is reliable is made more than once and produced the same result E.g. Clare has started a diet today at a starting weight of 13 stone. She weighed herself 4 times today and her scales have said every time that her weight is 13 stone. Her scales are reliable Mary has also started a diet . She is also 13 stone and she has also weighed herself 4 times today. Her scales read 13 stone, then 11 stone, then 16 stone and then 12 stone. Her scales are clearly not reliable While in psychology we tend to measure more abstract concepts such as happiness, IQ, aggression, memory, stress etc. rather than concrete things such as weight or length, we also want our scales to be reliable

Answer 182

Once you have planned your research and thought about how you could improve its reliability, you then need to assess your reliability, i.e. establish if it is reliable or not There are several methods of assessing reliability, such as: 1. Test-retest 2. Inter-observer reliability

Answer 183

Repeating the test again, on the same participants, but ensure sufficient time is left in between each test so order effects don’t occur If the test or questionnaire is reliable, then the results should be the same, or at least very similar, each time they are administered In the case of a questionnaire or psychological test, the 2 sets of scores would be correlated to make sure they are similar - the correlation coefficient should exceed +0.80 for reliability - this would demonstrate a strong positive correlation

Answer 184

One observer’s interpretation of events may differ widely from someone else’s - introducing subjectivity, bias and unreliability into the data collection process The recommendation would be that observations should be conducted in teams of at least 2, but inter-rather reliability must be established This involves a pilot study of the observation to check that observers are applying the behavioural categories in the same way Observers will watch the same event(s), record their data independently and then again correlate to make sure they are similar The correlation coefficient should exceed +0.80 for reliability - note that similar methods would apply tov other forms of observation such as content analysis and also interviews, although it would then be referred to as inter-rather reliability

Answer 185

Refers to the extent to which a piece of psychological research produces a results that is legitimate For example, Claire has started a diet at a weight of 13 stone. She weighed herself 4 times that day and they said that she was 13 stone each time. Her scales are reliable. She went to her GP who confirmed her weight is 13 stone. This shows her scales are reliable and valid Mary also started a diet. She is also starting at 13 stone named weighed herself 4 times that day and the scales read differently each time. Her scales are unreliable and invalid. Archie is also starting a diet. His scales are consistently saying that he is 14 stone, so they are reliable.. but then he goes to the doctor who says he is actually 15 stone. So his scales are reliable but invalid

Answer 186

There are 2 forms of validity: 1. Internal validity - does the research actually and accurately measure what it claims to measure? 2. External validity- can the research be legitimately generalised beyond the setting of the study?

Answer 187

Extraneous variables Confounding variables Demand characteristics Investigator effects Social desirability bias

Answer 188

(Happens in the planning stage) EVs and CVs - think about how EVs could be controlled for - trial run/pilot study of research to identify and EVs and CVs and control as appropriate e.g. experimental design, standardise procedure, environment etc Demand characteristics - single-blind study - this means ppts are not told the aim of a study before they take part/or they are lied to about the aim - reduced demand characteristics Investigator effects - use a double-blind study - neither the investigator or the ppts knows the aim of the study Social desirability (questionnaires, psychological tests) - confidentially - the ppts personal details are kept private and they are anonymous to the researcher - if ppts think their results are anonymous, they are more likely to be open and honest Potential internal validity issues with observations - ensure behavioural categories are clear and accurate - avoid behavioural categories that are to broad, overlapping or ambiguous

Answer 189

Unrepresentative samples - problems with population validity Problems with generalising the findings to other settings - problems with ecological validity Controlled/artificial setting/tasks - problems with ecological validity (mundane realism) Issues of whether findings or concepts hold true over time - problems with temporal validity

Answer 190

(Planning stage) Population validity - use a more representative sample - use stratified sampling Ecological validity - use a more natural setting - do research in the field e.g. in the real world Mundane realism - use a more natural task - testing memory by remembering shopping lists rather than nonsense syllables Temporal validity - test-retest over time and update/discard outdated theories - Freud’s work e.g. penis envy Interpretative validity - check the researcher’s interpretation in qualitative research moths is correct - triangulate by using a number of different sources e.g. with friends, personal diaries etc.

Answer 191

Ways for the researcher to test/see if a measure of behaviour e.g. a questionnaire, interview, or test measures what it intend to The first step for assessing your internal validity would be to look at the face validity of your measuring tool This involves a subjective assessment of whether or not a measurement tool appears to measures the behaviour it claims to You may therefore also want to consider the content validity of your measuring tool This is a more subjective assessment of your measurement tool to establish whether or not it relates to and measures the behaviour in question and involves you asking an independent expert on the topic of your research to evaluate your measurement tool The expert might suggest improvements or approve your method If you choose to develop your own measuring tool or test you should ensure that you validate your results with this new test by comparing your results to those of research using a well-established measuring tool or test This check for external validity is called concurrent validity

Answer 192

Studies have to be written up and published in peer reviewed academic journals In order to do this, you must use a set format

Answer 193

1. Abstract 2. Introduction 3. Method 4. Results 5. Discussion 6. Referencing

Answer 194

An ongoing debate in the field of psychology is whether psychology can be considered a science It is important to loom at all different aspects of what makes a science and how, it at all, psychology fulfils these expectations - Thomas Kuhn and paradigms - hypothesis testing - empirical method - falsifiability - objectivity - replicability - controlled - experimental method (The force)

Answer 195

The philosopher Thomas Kuhn (1962) argued that what separate a science from a non-scientific science is a universally accepted paradigm A paradigm is a set of shared assumptions and methods within a particular discipline For example, biology has evolution at its core Under this assumption, he suggested that psychology was best seen as a pre-science, separate from the likes of biology, chemists and physics (the natural sciences) This is, he argued, because unlike the natural science, psychology has too much disagreement at its core However, not everyone agrees with Kuhn here Most psychologists accept that the heart of psychology is the scientific study of mind and behaviour

Answer 196

Kuhn argued that the way in which a field of study moves forward from a pre-science is through model revolution This starts with a handful of scientists producing theories and evidence to challenges an existing, accepted paradigm Over time, more and more scientist start to challenge it, adding more research to contradict the previously accepted assumptions Eventually, a new paradigm is accepted - this is known as paradigm shift or paradigm change The model: Set out in a circular shape Normal science - accepted set of paradigms Model drift - start to question the accepted paradigms Model crisis - increasing change Model revolution Paradigm change - linked to model crisis leading to paradigm shift Pre-science - not a part of the circle, but coming off from normal science - no shared set of assumptions, lots of disagreement

Answer 197

Science uses hypotheses to test theories A theory is a set of general laws that have the ability to explain particular behaviours To construct a theory we need to collect evidence - science doesn’t allow for theories to based solely of ‘hunches’ Once the theory is developed, it will suggest a number of possible hypotheses (testable, predictable statements) These can be tested and findings will either strengthen or refute the theory

Answer 198

The most fundamental characteristic of a science is its reliance on empirical methods of observation and investigation This means that we have a body of rest each evidence from our theories Some examples of theories in psychology that have supporting research evidence is memory research

Answer 199

A key feature of the scientific method is to test hypotheses by falsifying them According to Popper (1959), if we can prove things to be false then we can rule them out as explanations and thus arrive at the truth by a process of elimination Any hypothesis that we cannot falsify is not scientific For example, the initial dopamine hypothesis of sz proposed that an individual with sz simply had too much dopamine However, this hypothesis was falsified by administering drugs that reduce the voles of dopamine and finding that these drugs had little to no effect on individuals who suffered mainly from the negative symptoms of sz This showed that the initial dopamine hypothesis was scientific

Answer 200

Scientific researchers must strive to maintain objectivity as part of their investigations All sources of personal bias must be minimised so as to not to distort or influence the research process Objective examples - bodo dolls, milgram obedience study, Loftus & Plasmer leading questions study Non-objective examples - case studies, interviews (especially unstructured), thematic content analysis

Answer 201

Scientific theories are replicable, they must be reliable

Answer 202

The most scientific research method is the controlled experiment (lab experiment) In a lab experiment, you manipulate IVs, measure DVs and control for EVs

Research Methods Flashcards

(245 cards)