Paper 2 but all - RM Flashcards
Describe tests for correlational designs and hypotheses
Spearmans when the data is nominal/ordinal
Pearsons when the data is interval
Describe tests for independent designs looking for a difference in hypothesis
Man whitney U - nominal/ ordinal
Unrelated T - interval
Describe tests for dependent designs looking for a difference in the hypothesis
Wilcoxon T - nominal/ordinal
related T - interval
Give the circumstances needed to conduct a chi test or a sign test
Chi - observational
Sign - between observations, finding a difference
Define:
- Aim
- Hypothesis
Aim - a statement referring to what a researcher intends to discover
Hypothesis - prediction of the studies outcome
- directional: stating the direction of the results - exactly how the IV effects the DV(higher, lower)
- non directional: when there is no previous research in the area - unclear how the IV will effect the DV
- null: stating that the IV will have no effect on the DV
Describe different types of variables
IV - what the researcher manipulates
DV- what the researcher measures
Confounding - varies systematically with the IV and may cause meaningless results
Extraneous - does not vary with the IV but affects the DV
Control variable - held constant to prevent any invalid results
What is operationalisation?
converting the DV into something that can be measured scientifically
e.g. intelligence through score on IQ
Give types of extraneous variables and how to prevent them
Demand characteristics: ppts change their behaviour as they know they are being observed (research will lack validity)
- prevent by using a single blind trial - do not tell the pp an aim or use a natural experiment
Investigator effects: unconscious or conscious cues by the experimenter that the ppt may pick up on and change their behaviour
- prevent by using double blind - neither the interviewer or the ppt is aware of the experiment
Define validity
when the research measures what they intend to measure
- only the IV effects the DV
Give different types of validity
Internal - concerns what is happening within the experiment (control variables)
External - extent to which the results can be applied to different people places and time
e.g:
Ecological - application to a real life setting
Population - application to different groups (age)
Historical/ temporal: application to different time periods
What is informed consent?
Ethical issue - revealing the true aims of the study before the ppt agrees so their agreement is valid
problems - the ppt may show demand characteristics reducing the validity of the experiment
benefits: basic human right to give consent
Dealings;
- presumptive consent: asking similar samples if they would agree ad then generalising (individual difference??)
- retrospective consent: asking after experiment
- general consent: ask if they would agree to similar experiments
What is deception?
Ethical issue - giving false information to the participants regarding the aim
Problems - deception is essential in order to produce natural results
benefits - no deception allows ppt to make informed consent and won’t be distressed or believe that psychologists are untrustworthy
Dealings;
- debriefing: revealing the truth after the experiment, clearly explains the aim and procedure, providing contact info etc
- however this will not change the ppts emotions towards experiment
What is protection from harm?
Ethical issues - ppts cannot experience any harm that they wouldn’t experience in their day to day lives
problems - a researcher is not interesting in ordinary behaviour and needs the ppt to be distressed in order to analyse abnormal behaviour
benefits - remains unharmed
Dealings;
- debriefing
- cost-benefit analysis: listing pros and cons and judge in terms of society and implications to people - judged by a committee to see whether it is worth it
What is the right to withdraw?
Ethical issues: ppts should be reminded that they can leave throughout the experiment
problems- biased results because the sample remaining is more obedient
benefits - they should be able to leave whenever and still keep the award (especially when deceived)
Dealings;
- informed that they can withdraw at any point
What is confidentiality?
Ethical issue - information cannot be traced back to the ppt
problems: n/a
benefits: ppts are free to act naturally and won’t be effected by social desirability bias
Dealings;
only using initial
What is privacy?
Ethical issue - if the study is natural, it should only be conducted in a public place
problems: unable to study independent behaviour
benefits: protects privacy of ppts, not embarrassed
Dealings;
- do not study anywhere in public - but there is no universal agreement
Describe and evaluate the use of a lab experiment
- highly controlled environment where the ppt are fully aware they are participating
Pros:
- control of all variables so high internal validity (establish cause and effect)
- standardised - can be repeated to find inter-rater reliability
- informed consent - meaning it is ethical
Cons:
- lacks ecological validity
- more chance of demand characteristics
- lacking mundane realism (artificial tasks)
Describe and evaluate the use of field experiments
- conducted in a real life setting where the ppts are not aware
Pros:
- high in ecological validity
- reflects mundane realism
- no demand characteristics or investigator effects
- mundane realism
Cons:
- lacks internal validity due to poor control of extraneous variables
- no informed consent; deemed unethical
Describe and evaluate the use of natural experiments
- conducted using a naturally occurring IV in a natural environment; cannot be created (still measures DV) e.g. abused victims vs unabused
Participants are aware
Pros:
- high in ecological validity
- highly ethical as there is no direct manipulation of the IV
Cons:
- cannot randomly allocate ppts (confounding variables)
- less control of the variables
- cannot demonstrate causal relationships because the IV cannot be manipulated
Describe and evaluate the use of quasi experiments
- IV is based on a pre-existing difference (no manipulation) e.g. gender or characteristics
Pros:
- standardised and controlled; high internal and repeatability
- less demand characteristics
Cons:
- lacking random allocation
- lacks ecological validity
Describe and evaluate naturalistic observations
- environment where target behaviour usually occurs
- experimenter does not interfere or control variables and leaves normal
Pros:
- high in ecological validity
Cons:
- difficult to replicate
- uncontrollable variables
- observation bias: difficult to be objective
Describe and evaluate controlled observations
- a specially designed environment which allows experimenter to control variables
- ppts are fully aware
Pros:
- controlled and can establish cause and effect
- standardised
Cons:
- possible demand characteristics (tends to be in lab settings)
- behaviour may not be entirely natural
- observational bias: difficult to be objective
Describe and evaluate overt observations
- ppts are aware they are in an observation
Pros:
- informed consent; ethical and have right to withdraw
Cons:
- observer effects; demand characteristics and less natural so lacking validity
- lacks validity
- may behaviour different
Describe and evaluate covert observations
- ppts are unaware that they are being watched - researcher is undercover
Pros:
- natural behaviour; they won’t pretend to be seen in a better light
- more valid : ppts unlikely to react
Cons:
- lacks ethics, no consent
- less control
Describe and evaluate participant observations
- observer takes part in the experiment as part of the group
Pros:
- good insight into the experiment
- revealing spontaneous info
Cons:
- bias and subjective - could have developed relationships
Describe and evaluate non-participant observations
- observer doesn’t take part in the experiment
pros:
- objective
- valid
cons:
- lacking insight and extra view
Describe sampling methods in observations
Time sampling:
- observing ppts in given time periods at different intervals
( saves time but missing behaviour when absent)
Event sampling:
- observing the whole time but only recording presentation of particular behaviours
(narrows down but missing alternative behaviour)
What is an independent groups design?
using separate groups of ppts for each condition
- no order effects
- less likely to have demand characteristics
- ppt variables: individual difference (confounding)
- more ppts are required
SOLUTION: random allocation of ppts to distribute variables equally
What is a matched pairs designs
- eliminates the ppt variables in independent groups do increase validity
- the separate groups are matched on specific characteristics that may act as confounding variables
What is a repeated measures design?
- comparing the DV between the same ppts who will experience both conditions
- results are not effected by individual differences
- problem of order effects which may produce demand characteristics or practise effects (extraneous)
SOLUTION: counterbalancing
- ABBA - splitting ppts and allow them to experience both conditions in opposite order
- gives them an equal amount of bias/ order effects
What is a sample?
a group of selected ppts which will be representative of the entire target population (the people you want to generalise the findings to and where the sample is being taken from)
Describe and evaluate different sampling methods
Volunteer: self selections; ppts choose to take part from a poster
P: informed consent, ethical
C: less valid; sample is biased to those who have spare time/motivation
Random: selecting a sample where each individual had an equal chance of being selected e.g. names in a hat or number generator
P: objective
C: time consuming; collecting list of members
Stratified: creating subgroups within the population and selecting sample in proportion (strata/total x sample size)
P: representative of everyone and avoid bias
C: time consuming
Systematic: pre-determined sequence; numerical intervals
P: objective system and avoiding bias
C: may be unrepresentative and only valid if you random sample first
Opportunity: selecting a sample based on availability and convenience for the researcher
P: saves time
C: biased and unrepresentative
Describe and evaluate correlations
- systematic association between two continuous variables to discover a link
Positive, negative and no correlation - lacks validity: unaccountable variables which the investigator has not manipulated
- could be third unidentified variables; lacks cause and effect
- ethical: researcher does not manipulate IV which may be difficult to achieve
- secondary data which is quick and economical to gain
Describe and evaluate the use of questionnaires
- pre-set questions which are used to asses the DV
- could be open Qs; allow detailed responses and qualitative data
- could be closed:
Fixed choice; list of responses with descriptions and selecting the one that applies
Likert scale; respondents are given a number which corresponds with a response
Rating scale; similar to likert but rating is only with extreme values
P: studying large samples in time and cost effective way; bigger sample, more representative and can be generalised
- experimenter isn’t present so no bias
C:
- social desirability: avoid by keeping anon
- demand characteristic
- response bias (hurry)
- could contain socially sensitive Qs which may effect individual
Describe and evaluate the use of interviews
- face to face interaction between the ppt and the research
- assessing thoughts and feelings
Structured: set of predetermined Qs
p - easy replication and standard, detailed qualitative
C - social desirability and possible interviewer bias who may interpret results to fit hypothesis (we should use oblivious one)
Semi- structured: predetermined Qs with room for scope and elaborations
Unstructured: conservation
- next Q is determined by the answer of the previous
p - more detailed info
c- more subjective
How can you improve self reports?
- avoid bias: use interviewer who doesn’t know aim
- good qs: avoiding double barrelled or double negative questions; clarity, good analysis
- avoid leading questions
- filter qs (distract from main purpose)
- sequences (easy - hard)
What is primary and secondary data?
Primary - information collected and observed directly by researcher
P: control of data, ethical, prevents bias and is designed specifically
C: time consuming, expensive, ethical
Secondary - information is collected by someone else but used by psychologist
P: quick and cheap, less equipment and no ethical issues
C: may not entirely fit data or could be biased, unaccounted for variables
What is qualitative and quantitative data?
Qualitative: cannot be numerical - worded and descriptive data
P: detailed and accessing thoughts and feelings, no restrictions
C: difficult to analyse and draw conclusions from
Quantitative: data is numerical
P: easy to analyse and conclude form
C: oversimplification, tick boxing which may not be representative - data could be meaningless
Describe methods of measuring central tendency
Mean: adding and dividing by total
P: most sensitive
C: effected by extreme values
Median: middle value in ordered data
P: not effected by outliers
C: insensitive, only one value
Mode: most common value
P: not distorted easily
C: insensitive
Describe methods of measuring dispersion
Range: arithmetic distance between the largest and smallest
C: doesn’t acknowledge distribution and effected by extreme values
Standard deviation: spread of data (large S.D is more spread)
P: precise
C: may hide certain characteristics
What is: - nominal -ordinal - interval data?
Nominal: categorical
Ordinal: rank/order
Interval: numerical; exact differences
Describe different graphs used to represent data
Bar graph - bars DO NOT touch
- y axis: frequency
- nominal data
Histograms - bars touch
- continuous and interval data
Line graph - continuous and interval
Scatter - measuring correlation
What is normal distribution?
- bell shaped curve - midpoint = mean
- 30% lie within 1 s.d. above and below
100% lies within 3 s.d.
Negative skews (right) Positive skew (left)
What are inferential statistics?
assumptions in stats
- using probabilities of less that 5% (0.05) - shows data is significant (only analyse data of those who would have less that 5% chance of it occurring naturally)
What are type 1 and type 2 errors?
Type 1: optimism; falsely accepting alternative and rejecting null
Type 2: pessimism; falsely accepting the null
What is the sign test?
Look at table: directional hypothesis is one tailed non-direction is two tailed N = the number of ppts S = lowest calculated value (lowest number of signs +/-) Critical value on table
S must be equal to or less than the critical value to show significance
Describe and evaluate peer review
- assessment of scientific work by experts in the same field
- ensures high quality research before publication
WHY:
- appropriate allocation of funding; identifying which require more or less to prevent waste of resources
- publication in academic journals should be validated and stops publishing weak finding
- university ranking: higher ranking for higher quality research
Cons:
- difficult to find expert in the same field with same knowledge
- anonymity - could be bias towards the person marking
- publication bias - only positive findings and therefore misinform the public
- publishing poor research is still common
- status quo; government uncomfortable with challenging results
Define reliability
- refers to consistency; you should get the same results in replicates (but this doesn’t mean it’s valid)
How may we asses reliability?
Test-retest:
- administrating a test (questionnaire and interview)
- giving the same test at a later date to see if there was correlation
Inter-rater reliability:
- getting two independent observers to look at the same behaviour to see if they agree with the findings
- everyone gets the same training and same behaviour categories
- record everything
- pilot study; making sure categories make sense
80% and similar is high inter-rater reliability
What is a pilot study?
small scale preliminary study to identify errors, formulate possible hypotheses, testing questions and believability of confederates
How are behavioural categories used in observations?
- taking target behaviours and operationalising them
- must be objective, non-overlapping ad standardised
e. g. aggression (looking for hitting, spitting, kicking)
these can be tallied in observations
How may we improve reliability?
lab experiments: control all extraneous
standardised: all ppts are given same instructions and timings
replication of the study
Observations:
- clear categories
- no ambiguity
Questionnaires:
- test-retest
- remove open Qs and ambiguous
Interviews:
- same interviewer
- same training
- structured
- no leading Qs
How do we asses and improve validity?
face validity: from face value, does it make cognitive sense? clear what is happening?
Concurrent validity: compare the results from study to well established study - if they are similar, data is valid
predictive validity: extent to which research can make valid predictions about future behaviour (Ainsworth)
What is content analysis? Evaluate
taking qualitative data and making it quantitative to be analysed
Indirect observations: analysing content of something (often media) and creating categories to measure representations
Steps of content analysis;
- sampling method: pick material and how frequently (TV)
- behaviour categories (content analysis of songs)
- coding: taking raw qualitative data and giving them codes and labels (e.g. rap=1) - these aid analysis
- pilot studies: making sure behaviour categories are clear and non-overlapping
P:
- high ecological - current, real life content and not artificial
- standardised categories
- reliable, can be replicated with same categories for the same effect
C:
- biased categories which may be supporting the psychologists hypothesis
- culture biased - different perspectives in different cultures
What is thematic analysis?
- technique used to analyse qualitative data; identifies themes and data is organised around them
- impose an order to represent perspectives
STEPS:
- reread and read transcript
- break them into meaningful units
- assign a label to the theme (heartbreak for love)
- count the number of times it occurs