PAPER 2- TOPIC 3 RESEARCH METHODS ✅ Flashcards

Question

the three factors affecting which inferential test to use

Answer 1

- data? (level of measurement) - difference? (testing for a difference or a correlation) - design (independent groups or matched pairs/ repeated measures ---> unrelated or related)

Answer 2

a more robust test, that may be able to identify significance that other tests can't MUST BE... - interval data - P's must be drawn from a normally distributed population - the variance between P's in each group must be similar

Answer 3

is the value that is produced by the statistical test

Answer 4

value that is gathered from the calculations table for the specific test - the cut off point between accepting and rejecting the null hypothesis

Answer 5

"gReater rule" if test has an R in it, the observed/ calculated value should be GREATER than or equal to the critical value e. g. unRelated t - Related t - chi- squaRe - peaRsons R - spearman's Rho all should have an observed value ≥ critical value to be significant (sign test, wilcoxon and mann whitney u must have observed value ≤ critical value, to be significant)

Answer 6

- presented in form of categories | - is discrete and non-continuous

Answer 7

- presented in orders or ranked - no equal intervals between units of data - lack precision as subjective for to what someone sees a "4" as - data is converted into ranks (1st, 2nd, 3rd) for statistical tests as raw scores are not accurate enough

Answer 8

- continuous data - units of equal, precisely defined sizes (often public measurement scales used - e.g. time, temperature) - most sophisticated precise data - hence their use in parametric tests

Answer 9

matched pairs | repeated measures

Answer 10

independent groups

Answer 11

type 1- false positive (said there was a significance when their wasn't 𝗼𝗻𝗲) type 2- false negative (𝘁𝗼𝗼 strict)

Answer 12

* find difference between two scores (+ - 0 ) * select lowest number of + or - as 's' observed value (same for Wilcoxon test) * calculate N (no. of participants - 0's) * use hypothesis, probability and N value to find critical value * s must be ≤ critical value, to be significant

Answer 13

* observed value 's' of 1.4 was ≤ critical value of 1.6 for N value of 10 at a probability of 5% for a one tailed test * therefore, we can accept the alternative hypothesis showing that 'the happiness score of toffees increases when Rafa is out, rather then when he is in'

Answer 14

``` abstract introduction method results discussion referencing ```

Answer 15

- short summary of the study - includes all major elements: aims, hypothesis, method, results, discussion - written last, at start of report

Answer 16

• large section of writing - outlines relevant theories, concepts and other research- & how they relate to this study - state aims an hypotheses

Answer 17

✰ section explaining how experimental is carried out, split into: * design - experimental design (e.g. IG, MP, RM) ; experimental method (overt, naturalistic); IV & DV; and validity and reliability issues * participants - sampling technique, who is studied (biological and demographic), how many P's, target population * apparatus/ materials needed * procedure - step by step instructions of how it was carried out, include briefing and debrief to P's * ethics - DRIPC, how this was addressed

Answer 18

✰ summary of key findings, split into : • descriptive statistics - uses tables, graphs and measures of central tendency & dispersion • inferential statistics - test chosen, calculated and critical values, significance level, if it was significant, which hypotheses accepted ••••• if gathered qualitative data, likely to be in the form of categories or themes

Answer 19

✰ large piece of writing where researcher summarises and interprets the findings verbally and the implication of them includes: - relationship to previous research in introduction - limitations of research- consider methodology and suggestions for improvement - wider implications of research- real world applications and the contribution of research to current theories - suggestions for future research

Answer 20

the full details of any source material mentioned in the report, are cited

Answer 21

surname, first initial (year published), title of book (italics), place of publication. publisher e.g. Copland, S (1994), 𝘛𝘩𝘦 𝘤𝘩𝘳𝘰𝘯𝘪𝘤𝘭𝘦𝘴 𝘰𝘧 𝘣𝘦𝘪𝘯𝘨 𝘴𝘶𝘴, California, Puffin books

Answer 22

author, date, article title, journal name (italics), volume (issue), page numbers e.g. Copland, S (1994) Effects of being sus on your ball knowledge, 𝘛𝘩𝘦 𝘜𝘭𝘵𝘪𝘮𝘢𝘵𝘦 𝘉𝘢𝘭𝘭 𝘒𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦 𝘵𝘦𝘴𝘵 , 11 (12), 231-237

Answer 23

- contains any raw data, questionnaires, debriefs, consent forms, calculations - evidence that don't fit in the main body of report

Answer 24

* aim * what they will do, and for how long * right to withdraw and confidentiality * ask for questions * place to sign & add date

Answer 25

* aims * discuss what went on in all conditions and any deception * findings * right to withdraw * remind confidentiality * where they can find more info * any questions?

Answer 26

• step by step of everything P has to do

Answer 27

whether the observed effect of a study is genuinely due to the manipulation of IV (measuring what they set out to) and accurately generalised beyond the research setting (e.g. across historical contexts, compared to well-recognised studies, measuring what set out to measure)

Answer 28

- internal - external - ecological - concurrent - face - temporal

Answer 29

extent to which findings have a correlation with the results from well-recognised studies with established validity

Answer 30

extent to which findings can be generalised to other historical contexts/eras

Answer 31

extent to which findings can be generalised to real life, outside of the research setting

Answer 32

extent to whether on the surface, a study looks like it will measure what it's set out t.

Answer 33

extent to which a study measures what it set out to measure | is observed effect on the DV, due to maniupulation of IV

Answer 34

the extent to which the findings reflects the real world (in terms of the population (population), the environment (ecological), the time era (over time)

Answer 35

• questionnaires - incorporate redundant questions to create a 'lie scale' (account for social desirability bias) - anonymity - remove ambiguous questions • interviews and case studies - structured interview reduces investigator effects, but reduce rapport and so answers less accurate - triangulate data - gain respondent validity by checking you understood the p correctly and use quotes in findings (increase interpretive validity) • experiments - control group - pilot study to expose extraneous variables - change experimental design to reduce order effects or effect of participant variables - standardise procedure - counterbalancing, double blind, randomisation • observations - familiarise with BC so don't miss anything - operationalise BC, so it is clear what you're looking for - use covert or non participant

Answer 36

a small scale trail run of the actual study completed before the real full scale research is completed

Answer 37

- can identify extraneous variables, that can be controlled for the real study - can help improve reliability (test-retest) - modify any flaws with procedure or design (reduce cost from messing up large scale) - can allow training of observers - can adapt or remove ambiguous or confusing questions in questionnaire or interview - can identify areas where further randomisation, counterbalancing, standardisation etc... can be used, to limit any observed order effects, bias, investigator effect or demand characteristics

Answer 38

assessment of research, done by other psychologists in a similar field, who provide an unbiased opinion of a study to ensure it is high enough quality for publication

Answer 39

- allocate research funding as people (and funding organisations) may award funding for a research idea they support - ensure only high quality, ... < useful >..... studies are published - suggest amendments, improvements or withdrawal before publishment

Answer 40

- research is sent to an anonymous peer to objectively review all aspects of written investigstion - they look for: • clear and professional methods & design • validity • originality (not copied) and significance of the research in that field of psychology • results - the statistics chosen and the conclusions

Answer 41

•bury ground breaking research - may slow down rate of change - also if research contradicts paradigm or mainstream research, itsmay be buried or resisted •publication bias - editor preferences may give false view of current psychology - some publishers only want to publish positive news or headline grabbing research to boost the popularity of their journal (may ignore valuable research) •anonymity - the peers reviewing stay unidentified - researchers competing for funding may be over critical (some publishers now reveal after who is reviewing to combat this) - reviewers may also resist findings that challenge their previous research

Answer 42

how consistent a study's data is, and the extent to which it would produce similar results if the study was repeated

Answer 43

• inter-observer reliability (inter-rater for forms like content analysis) • test-retest reliability

Answer 44

the extent to which there is an agreement between two or more observers observing the same behaviour, using the same behavioural categories

Answer 45

the extent to which there is a correlation between the scores of the same P’s on a test or questionnaire, completed on different occasions

Answer 46

- complete observation again with two or more observers watching the same observation and using same behavioural categories - compare the results of the different observations using a spearman's rho test (0.8+ for a strong correlation)

Answer 47

- administer same test or questionnaire to the same P on different occasions - not too soon - prevent recall of answers - not too long - prevent the views or ability being tested changing - use a correlation to compare the results, (0.8+ for a strong correlation)

Answer 48

• questionnaires - closed questions - clear, unambiguous questions • interviews and case studies - same researcher - limits leading or ambiguous questions - structured interviews • experiments - standardised procedure and instructions • observations - operationalise and familiarise/train observers with behavioural categories - two or more observers and compare results -

Answer 49

• 'eyeball' test to measure face validity of the test or measure OR - pass to expert to measure face validity AND • compare with well-recognised test with established validity to create a correlation co-efficient, measuring concurrent validity (close agreement in +0.8)

Answer 50

when we assume a technique that works in one cultural context will work in another

Answer 51

combination of research and findings from several studies on the same topic - findings are weighted for its sample size

Answer 52

aka types of experiment lab field natural quasi

Answer 53

IV is a naturally occurring 𝗯𝗶𝗼𝗹𝗼𝗴𝗶𝗰𝗮𝗹 difference (existing between people) - --> not manipulated or changed - measuring affect of naturally occurring IV on DV - can be field or lab Bahrick- studied duration of LTM of those within 15 years of graduation, and those within 50 years of graduation

Answer 54

IV is a naturally occurring event, not manipulated - someone or something caused the IV to vary (not the researcher) - measuring affect of naturally occurring IV on DV - can be field or lab Rutter's Romanian Orphans study

Answer 55

carried out in a natural, everyday setting (the usual environment the P's are in) - IV manipulation is under less control Piliavin- studied how people react differently when they see someone collapse on the train, due to drinking or due to injury

Answer 56

- carried out under controlled conditions - researcher manipulates IV to see effect on the DV Milgram

Answer 57

- control of variables - --> easier to replicate - --> an establish cause and effect of IV and DV as variables exactly manipulated (high internal validity) - --> more accurate results

Answer 58

- tasks may be artificial or trivial - low mundane realism - P's aware they are being studied- DC (low external validity) - artificial environment- may behave differently - hard to generalise to real world (low external validity)

Answer 59

- higher mundane realism in natural environment - P's unaware being studied (no DC) - higher external validity

Answer 60

- ethical issues- P's can't consent, and their privacy is invaded - less control over variables - less valid, harder to replicate and harder to establish cause and effect between IV and DV

Answer 61

- opportunity for research that would not be possible, for practical and ethical reasons - high external validity, as study real world scenarios and problems

Answer 62

- IV isn't deliberately changed so can't say that IV caused the observed change in DV (cant claim cause and effect) - events rarely occur, so hard to generalise findings - P's can't be randomly allocated to conditions - ---> may experience confounding variables

Answer 63

completed in a field or lab setting - if in field: P's unaware being studied (no DC), mundane realism - if in lab: control over variables, replication

Answer 64

- P's can't be randomly allocated to conditions - ---> may experience confounding variables (e. g. all those who been in car crash may have higher trauma level than control group) - IV isn't deliberately changed so can't say that IV caused the observed change in DV (cant claim cause and effect)

Answer 65

results are representative of everyday life

Answer 66

different ways P's can be arranged into experimental conditions - RM - IG - MP

Answer 67

one group of P's experience both conditions of the experiment - mean of condition A compared to mean of condition B

Answer 68

- two separate groups experience two different conditions of the experiment - performance of each group is then compared

Answer 69

- there are a separate group of P's for each condition of IV - but each P is matched to one other, based on certain shared characteristics relevant to the study e. g. complete IQ tests, prior to actual study and match the no.1 score with no.2 score - each P does one condition and their scores are compared directly against their partner

Answer 70

+++++++ • no effect of individual variables • more economical - only need one group of P's (less time/£ on recruitment) ---------- • order effects (boredom, fatigue, practice), also 1st condition may affect the 2nd condition (e.g. effects of coffee in C1 may continue into C2 where they have water) • may realise what study is aiming to find, therefore DC may be present

Answer 71

+++++++ • reduced order effects • less chance of realising aim of study - no DC ---------- • individual variables • less economical- twice as many P's needed

Answer 72

+++++++ • accounts for individual variables • reduced order effects • unlikely to realise aim of study ---------- • more expensive and time consuming to find suitable matching P's, also may need a pre-test • although matched, individual variables will still be present

Answer 73

- keeping procedures in a research study the same - all participants treated the same - (so they have the same experience) • removes experimenter bias also makes the study replicable and easy to complete again accurately

Answer 74

where half of P's do the first condition first followed by the second, and the other half do the second condition first and the first condition second - control for order effects

Answer 75

each participant has an equal chance of being in each group/condition - control for participant variables

Answer 76

individual characteristics that may influence how a participant behaves

Answer 77

- use of chance wherever possible to reduce bias or investigator effects (conscious or unconscious)

Answer 78

double blind: demand characteristics and experimenter bias single blind: demand characteristics

Answer 79

- all members of target pop have an equal chance of being selected METHOD - randomly selects the list of P's in the sample - assign all names in sampling frame a number, input all names into computer, and randomly generate n numbers, the P's that these numbers correlate to, are in the sample

Answer 80

+++++++ • no bias, as everyone has equal chance (confounding and extraneous variables are equal divided across groups) ---------- • may still be unrepresentative as its random • hard to obtain complete list of target pop

Answer 81

every nth member of population is selected METHOD - organise sampling frame e.g alphabetically - begin from randomly generated number and select every nth member after until sample complete

Answer 82

+++++++ • unbiased and objective, once n is selected investigator has no effect ---------- • may still be unrepresentative • time consuming and unrealistic to obtain full sampling frame

Answer 83

composition of sample is weighted to reflect the proportion of people in certain subgroups in the target population (e.g. race, religion, what football team support) METHOD - identify strata in population, work out the % of population the strata contains, use same % representations in sample

Answer 84

+++++++ • highly representative of target population - can generalise findings • no bias ---------- • very time consuming and costly • hard to determine what variables to split the population up, based on

Answer 85

selecting the first people that are willing and able to take part in the study - whoever, is available and around at the time of them completing their study

Answer 86

+++++++ • convenient - time and low cost (don't need full sampling frame) • more likely to be people you know, so easier to conduct study ---------- • researcher may only go up to/ ask a certain type of person (only 'approachable' people) (researcher bias) • unrepresentative, as only from one specific area

Answer 87

researcher advertises the study (or even ask people to raise their hand) and P's select themselves to take part

Answer 88

+++++++ • minimum input from researcher, P's come to you • all of the P's who put themselves forward will be willing to participate (all engaged too) ---------- • attract a certain personality of those who want to help, or please the investigator • cost of advertising • risk not enough people willing to take part

Answer 89

- selected P's may refuse to take part, ends up more like opportunity sampling

Answer 90

- a group of people that the researcher is interested in studying - a small subset selected from the target population to take part in the study, through a particular sampling method (assumes to be representative of target population)

Answer 91

- if sample is representative, the findings from the study can be applied to the wider target population

Answer 92

when certain subgroups of a target population are under or over represented within the sample

Answer 93

- naturalistic - controlled - covert - overt - participant - non participant - structured - unstructured

Answer 94

• naturalistic - completed in field setting, where target behaviours would normally occur - investigator doesn't interfere with setting or variables • controlled - completed in lab setting under controlled conditions - investigator controls E and C variables, and manipulate variables to observe effect

Answer 95

covert (covert operation- under cover) - participants are NOT aware they are being observed - behaviour must be in public and happening anyway (to be ethical) overt - participants ARE aware they are being observed

Answer 96

participant - observer involves themself in the group of people they are observing - sometimes only possible if involve (e.g. to see how factory workers are treated, join them) non participant - observer remains separate from group they are observing, and don't interfere

Answer 97

``` OVERT +++++++ • can obtain informed consent --------- • more likely DC • likely susceptible to hawthorne effect (act differently as knew you're being watched) ``` COVERT +++++++ • unlikely for DC - increasing internal validity • less likely susceptible to hawthorne effect (act differently as knew you're being watched) ---------- • can't obtain informed consent, so are observing people without them knowing

Answer 98

NATURALISTIC +++++++ • higher mundane realism, external validity • P's more likely to act naturally ---------- • struggle to judge patterns of behaviour as cant control extraneous or confounding variables • harder to replicate ``` CONTROLLED +++++++ • easier to replicate • more control over variables, higher chance of cause and effect ---------- • DC • artificial environment- less mundane realism • hawthorne effect ```

Answer 99

PARTICIPANT +++++++ • may be necessary (e.g. to see how factory workers are treated, join them) • can understand reason behind behaviour as gain idea of emotions • can build rapport with P's and gain more insight and detail ---------- • observer may go native (too invested in experiment, they lose track of aims) • may miss important observations when involved • may lose objectivity if adopt the "local lifestyle" (identity too strongly with P's) NON-PARTICIPANT +++++++ • may be necessary when studying certain social groups (e.g. 50 year old man can't blend in and join a group of 15 year olds boys) • no risk of observers going native • objective ---------- • cant gain understanding of why people are behaving how they are, or may miss further insight)

Answer 100

pros - give actual insight into behaviour, people don't always act how they say they do cons - observer bias - depends on how observer interprets the situation and behaviour - cant demonstrate cause and effect as variables aren't as closely manipulated, may be confounding variables - cant understand why people are behaving how they are

Answer 101

structured - uses pre-determined behavioural categories to record the frequency that these target behaviours occur - ----> used in larger studies where there is too much going on unstructured - observe all relevant behaviour with no standardised checklist of behavioural categories - ---> used in small scale observation with few P's

Answer 102

STRUCTURED +++++++ • produce quantitative data- easier to analyse and draw conclusions • inter rater reliability increased - all observers looking for same behaviours --------- • not as detailed insight • may miss important behaviours UNSTRUCTURED ++++++ • rich in data • understand reason behind behaviour ---------- • hard to replicate, as interpretations of observer important • so much information - time consuming to analyse and make conclusions • observer bias - e.g. may only record behaviours that stand out

Answer 103

standardised checklist of operationalised target behaviours, that have been broken up to be more measurable and observable - only in structured observations

Answer 104

+++ • replicable • know what to look out for • allow for inter-observer reliability ----- • some categories may be wasted, or empty • must be clear, operationalised and not overlap otherwise up for interpretation • some behaviours may be missed, if no category for it

Answer 105

* continuous recording - making note of all target behaviours that occur (unstructured) * event sampling - tallying the number of times a particular behaviour occurs in * time sampling - recording all target behaviour that occurs at set intervals for x amount of time

Answer 106

* continuous recording - record all instances of target behaviour (unstructured) * event sampling - tallying the number of times particular target behaviours occur * time sampling - record behaviour that occurs within a pre-established time frame, at set intervals

Answer 107

+++ • easier to complete • focus on chosen behaviours, so don't miss anything ---- • if lots happen at same time, may be hard to keep up • may have some wasted categories

Answer 108

+ • time efficient - • may miss out in between periods

Answer 109

a live conversation where an interviewer asks the interviewee questions to assess their individual thoughts and experiences

Answer 110

• structured - standardised pre determine questions asked in a fixed order • semi-structured - list of questions to ask in advance, but can ask follow up questions • unstructured - no set questions in any fixed order - only a general aim to discuss a certain topic - interviewer encouraged to elaborate on answers, by interviewer - like a conversation

Answer 111

+ easier to analyse + replicable + interviewer requires less skill and training - answers may be restricted by questions - increase social desirability bias, as don't have too justify

Answer 112

+ more detailed insight, can understand reason behind + responses tend to be more honest, as have to be able to justify - hard to analyse and make general conclusions, as lots of different questions and answers - different questions may be interpreted differently by different P's - interviewer bias may affect what questions are asked

Answer 113

a set of written questions that a respondent answers to assess their individual thoughts, experiences and behaviour

Answer 114

* overuse of jargon - using specialist topic vocabulary and assuming the respondent knows more about the topic then they actually do * emotive language - author of Q portrays their emotion through the emotions of the word they use * leading questions - phrasing of the question indicates the respondent to answer in a particular way * double barrelled questions - contains two Qs in one * double negative - two forms of negative in same Q .... all decrease clarity, increase confusion and misinterpretation

Answer 115

open questions are Qs that allow the respondent to answer how they wish, with no sort answers to choose from closed questions are Qs with a fixed number of responses

Answer 116

+ cost effective + wide geographical use + convenient (don't need researcher present) + can generate large volumes of data easily + easily analyse and compare data - may be subject to social desirability bias - response bias (e.g. acquiescence bias or responding without fully reading the Q

Answer 117

• the problems that arise from how of P’s in the study are treated these exist when there is conflict between the aims of the experiment to produce valid research data, and the rights or safety of P's code of issues created by British Psychological Society (BPS)

Answer 118

DRIPC Deception- deliberately withholding information and misleading P's at any time Right to withdraw- the rights of P's to be able to withdraw themselves and their data from the study at any time Informed consent- making sure P's know what the aims of the research, the procedure, their rights and show their data will be used Protection from harm- idea that P's emotional and physical health is top priority and should be protected Confidentiality- the right of P's to not share their data to the public, or doing so anonymously to protect their identity

Answer 119

• consent form - a form that a P signs to confirm they know the aims of there experiment and what they will have to do if impractical to get consent form: • retrospective consent - ask P's for consent after participation, during debrief • prior general - P's give their permission to take part in a number of studies, • presumptive consent - ask people of a similar group to P's if they would consent to the study

Answer 120

* debrief - after study, explain true aims of experiment and any information that was withheld from them * reminded they can withdraw themselves and their data at any time * provide support through counselling or therapy if P's need it * re-assure that their behaviour was normal/typical

Answer 121

• usually don't record any personal data to maintain anonymity ---> often use initials or numbers to describe P's • during briefing and debrief, remind P's their data is private and won't be shared

Answer 122

numerical data that involves statistics - easy to analyse - lack detail

Answer 123

non numerical data expressed in words - hard to analyse ---> hard identfying patterns - interpreted--> bias - rich in detail --> true feelings--> greater external validity

Answer 124

data gathered first hand by the researcher, particularly for the research project - gathered through observations, interviews, questionnaires • can be designed in a way to target collecting the data that they need for the investigation

Answer 125

data collected by someone else, other than the researcher, that already exists before the research project e.g books, journals, websites, government statistics • may be outdated

Answer 126

any element of an investigation, that can change or vary - the variable that is manipulated, so that the effect on the D can be measured - the variable that is measured by the researcher

Answer 127

- type of extraneous variable - P's think they may have guessed the aims of the research and therefore act in a different way - either help or hinder the experimenter to find what they want)

Answer 128

* the unwanted influence of the investigator's actions on the behaviour of P's * the differences in behaviour of P's as they try to adjust to the study environment

Answer 129

* demand characteristics - P's act different as they think they know the aims of study * hawthorne effect - P's act different as they know someone is watching * social desirability bias - P's act alter their behaviour to be seen as more socially acceptable

Answer 130

• experimenter bias - experimenter affects the results through interpretation, body language, facial expressions and gender bias (male researchers prefer female Ps) • interviewer bias - ways in which the interviewer influences the response of the P ---->nodding, leading questions, interpretations • greenspoon effect - interviewer affects the way the P responds in interview by making affirmative noises

Answer 131

aspects of the environment and situation a P is in, that might affect their behaviour e.g. light, temp, noise

Answer 132

* Bowlby's theory of monotropy, and further research into role of father * development of treatments for disorders

Answer 133

- Bowlby suggested only mothers can form the monotropic attachment bond with their baby - This co-erced mums to stay at home due to law of accumulated separation and law of continuity, and meant the father, who may be earning less, had to work - recent research into role of father (e.g. Frodi found fathers and mothers show same physiological responses to babies crying on clips) - --> has allowed more flexible working arrangements (fathers to be stay at home, and mothers work if they earn more) - ---> because it shows both parents are suitable to provide necessary emotional support - can maximise income and contribute positively to economy

Answer 134

- creates a more healthy workforce, contributing to labour - workers can manage their symptoms and return to work efficiently - also cost of mass producing drugs outweighs the £15 million cost of absence from work for the economy

Answer 135

- presumptive consent - ----> ask similar group of people how they would feel about taking part, if they agree, there assume real P's would to - prior general consent - ----> asking P's to give permission to take part in a number of different studies, including one that will involve deception

PAPER 2- TOPIC 3 RESEARCH METHODS ✅ Flashcards

(160 cards)