P2 - Research methods

Experimental methods

• laboratory

• field

• quasi

• natural

Laboratory experiments - strengths

• can establish cause-effect relationship

• IV=cause, DV=effect

• replicability - repeat and achieve same finding

• more objective than other methods

• highly controlled

Laboratory experiments - limitations

• lacks mundane realism

• lacks ecological validity

• lacks experimental realism

• know they’re being observed = demand characteristics (e.g. ‘screw you effect’)

• evaluation apprehension (nervous of judgement

• limited sample size - population validity

laboratory experiments - important

random allocation to conditions so there are no major differences that affect results

laboratory experiments - ethical considerations

need informed consent, consider long term effects (physical/psychological harm), need right to withdraw

Field experiments

• carried out in natural setting, e.g. school, work, etc.

• IV deliberately manipulated

• participants unaware

Field experiments - research

Shotland and Straw 1976

• Male and female confederates staged an argument

• 1 condition F: ‘I don’t know you!’, 2 condition: ‘I don’t know why I ever married you!’

• found less likely to help when ‘married’

• IV: shouted phrase, DV: number of people who attempted to help

• ethical: possible psychological harm

Field experiments - strengths

• no demand characteristics

• mundane realism = higher ecological validity

• experimental realism

• no evaluation apprehension

• scenario can be replicated

• can establish cause and effect

Field experiments - limitations

• many extraneous validity (lacks internal validity)

• lack of informed consent

• C-E relationship less clear

• random allocation is difficult

• ethics: almost impossible to offer right to withdraw or give debriefing

Quasi experiment

• when its not possible/unethical to randomly allocate participants/manipulate the IV

• resembles true experiments but weak on some characteristics, key differences in point 1

• use pre-existing group e.g. effects of divorce on young children or relation between heart disease and personality

Quasi experiment - strengths

• investigate effects of IV that would be unethical to manipulate

• participants behave naturally

Quasi experiment - limitations

• less control as IV not manipulated

• no random allocation

• difficult to establish C-E

• requires ethical sensitivity

Natural experiments

• type of quasi experiment

• use of naturally occurring event for research purposes, e.g. social/geographic

• experimenter has no control over changes in IV

• e.g. affects of stress after natural disaster/bereavement

• natural disasters, elections, wars, riots, terrorism, pandemic

Natural experiments - research

Kario et al. 2003 studied effects of Kobe earthquake, 6400 people died, measured stress of those closest to epicentre, increased rate of heart attacks and sudden death 24 hours after

Natural experiments - strengths

• participants often not aware they’re taking part in an experiment

• allows us to study effects on behaviour of IV that would be unethical (mostly impossible) to manipulate

Natural experiments - limitations

• participants have not been assigned at random

• IV not controlled

• cannot make causal inferences

• participants unaware of participation

• sensitivity - experimenter attitude

Observational techniques

involves observing behaviour covertly (natural) or openly (overt/controlled) or as a participant in the activity

• natural observation

• controlled observation

• participant observation

Natural observation

unobtrusive observational study conducted in a natural setting

natural observation - strengths

• participants unbiased

• mundane realism = higher generalisability

• flexible

• external validity

• don’t have to obtain consent

• works well with children/non-humans

Natural observation - limitations

• too many uncontrolled and unknown factors

  • extraneous variables, hard to establish C-E

• observer has to be natural (or response changes)

• ethical - participant doesn’t realise they’re participating

• training observer is time consuming and expensive

• impossible to replicate

controlled observation

observations whereby the researcher exercises control over environment in which the observation is conducted

controlled observation - strengths

• easily replicated

• good control of variables, establish cause and effect

• less risk of extraneous variables

• comparison of extraneous - rich in detail and more complete

controlled observations

• lacks mundane realism - hard to generalise

• investigator effects - experimenter expectations

• social desirability bias

• demand characteristics

  • awareness = change in behaviour

participant observation

• observers in natural setting where observer interacts directly with participants

covert observation

undercover

overt observation

obvious

Self report techniques

• questionnaires

  • open questions

  • closed questions

• interviews

  • structured

  • unstructured

Questionnaire

• survey that requires written answers

• given set of Qs with instructions about how to record their answers

• used to explain an endless range of issues

  • personality, attitudes, beliefs

closed questions

• closed/fixed choice e.g. Y/N or ratings of agreement

• easier to store/quantify but restricts participants answers e.g. less depth → quantitative

open questions

• more realistic as in everyday life we have more scope to answer questions in our own way

• qualitative but subjective

questionnaire issues

• complexity: same questions may be too difficult to understand

• ambiguity: items can be interpreted in more than one way

• double-barrelled items: contains 2 questions and asks the participants for a Y/N response, ppt may want to give a yes response to one part but no to other

• leading questions: contains implications that a certain response is expected

what should be considered when designing a questionnaire

• aim: easier to write questions to address this

• length: short and to the point to decrease drop out rate

• use successful past questionnaires as basis

• question formation should be concise and unambiguous

• pilot study - relevance

• measurement scales - Likert type sd-d-n-a-sa

If the questionnaire is good

• standardisation: given to large representative sample so individual scores can be compared

• reliability - extent to which findings are consistent

• test-retest - individuals given same questionnaire on two different occasions and scores are correlated (high=0.8+=reliability)

• split half technique - questionnaire split in 2 and scores from one half are compared with other (correlation)

• validity - extent questionnaire measures what is claims to measure

Interview

• verbal research method that in which the participant answers a series of questionnaires

• one on one but can still be virtual

• collect thoughts, feelings, attitudes, and opinions of the participant

structured interview

• uses set of preprepared closed/open questions, or combo

• responses written/recorded

• does not veer from script

• qualitative data → through follow up questions e.g. can you explain why?

• quantitative data → e.g. number of ‘yes’ responses

structured interview - strengths

• use of standardised questions - replicable

  • minimise researcher effects

• more quantitative than unstructured

  • can be statistically analysed and increases reliability

structured interviews - limitations

• pre-determined questions may be restrictive

• participants may say something which should be explored further

• limits usefulness

unstructured interview

• no preprepared questions, keep open mind

• writes/records interview

• treated as a conversation, freedom with responses

• generally starts with open question or posing an idea

  • ‘lots of people think there should be harsher punishment… what do you think?’

• qualitative data only

unstructured interview - strengths

• high ecological validity - complete freedom

  • tailored to individual, open expression without manipulation

• flexibility to explore interesting areas that emerge

  • topics discussed from many POVs

  • original can be abandoned, may bring new insight

unstructured interview - limitations

• easy to derail - ppts may want to go into detail about irrelevant topics, easy to mix details & lose narrative

  • limits reliability

• researcher may lose objectivity, especially if its more than one session

  • may feel too close/identify with participant or wish to present participant in best light (social desirability bias) = compromised validity

Designing an interview

1. interview schedule determines: nature/number of questions asked, type of interview and interviewer

• use reflexivity - identify role in research and any existing ideas you have that could influence

2. find environment where ppt feels safe to disclose what may be sensitive information

• e.g. neutral room, quiet, comfy seats

3. establish rapport with ppt before interview - ppt more relaxed, could be informal chatting, given consent form and told they have the RtW, confidentiality and anonymity

4. ask and answer questions - ensure reliable way of recording

• ensure clear, coherent and on topic

• do not pass judgement or make them uncomfortable/compromised

Correlational analysis

correlation illustrates the strength and direction of an association between two or more co-variables, instead of IV and DV]

CORRELATION DOES NOT INFER CAUSATION

positive correlation

an increase in one variable leads to an increase in another variable

negative correlation

as one variable increases the other decreases

zero correlation

there is no correlation

How can correlation be measured

• scatter graphs

• correlation co-efficient

as these are statistical methods using quantitative data, need to operationalise variables

scatter graphs

one variable on the x-axis and one on the y

correlation coefficient

numerical representation of strength and direction of the relationship between two variables

• anywhere between -1.0 to +1.0

• ±1.0 = perfect (strong) correlation

• 0 = no correlation

correlational analysis - strengths

• allows researchers to analyse situations that could not be manipulated experimentally

• can produce reasonably definitive information about causal relationships if there’s no correlation

• can collect great amounts of data quickly

• easy and quick to analyse

• allows us to see relationship between two variables

correlational analysis - limitations

• cannot establish cause and effect

• researchers cannot manipulate variables

• confounding variables other than the ones you are measuring could have an effect

• ethical issues - often study controversial/sensitive issues hence need to be aware of social sensitivity

content analysis

• analysis of behaviours, written or spoken word into categories (top-down or bottom-up), known as coding units

• once in categories, the data can be counted (qualitative to quantitative)

why use content analysis

• carried out in order to understand the change in the trend of the content overtime

• can be used to explain why there is special attention or focus on certain topics of content

content analysis - stages

1. researcher put forward one or more general hypothesis

2. researcher generally identifies categories of theoretical relevance into which gathered data will be placed (e.g. <30 or >30)

3. researchers need to decide which sources of information to use as their sample

4. it is good to have two or more judges or coders to assign the information into categories to ensure it is reliable and consistent

• ideally coders shouldn’t know hypothesis

5. result of the content analysis need to be related to the hypothesis that motivated the study

Cumberbatch 1990

• studied advertisements on British TV to look for sexist bias

• found only 25% of women in these adverts were over 30, compared with 75% of men

• 89% of voice overs were male, especially when communicating expert/official information

Top-down content analysis

researcher starts with pre-set categories(usually base on prior research/theory)

Bottom-up content analysis

research allows categories to emerge from the data(may lead to development of new theories

content analysis - strengths

• gives opportunity to understand people as rounded individuals in a social context

• can reduce complex behaviours/information into manageable categories(no reasoning though)

• Ease of application - content analysis easy and less expensive

• reliability as easy to replicate

• suggests interesting hypothesis that could be tested in subsequent research

• complements other methods e.g. verifying results, useful longitudinal tool(shows trends over time)

content analysis - limitations

• data may come from unrepresentative sample(one article), making it hard to generalise

• flawed results - limited availability of material hence observe trends that may not reflect reality

• social desirability bias

• if researcher has huge amounts of material, easily show bias by emphasising information that favours hypothesis

• lack of causality - not under controlled conditions

Thematic analysis

• qualitative analytic method, analysis and reporting themes(patterns) within data

• patterns identified through data coding(similar to content analysis)

• organises, describes, and interprets data

• identified themes become categories for analysis

• goes beyond just counting up words or phrases, and involves identifying ideas within data

• can involve comparison of themes, identification of co-occurrences of themes and using graphs to display differences

Thematic analysis - stage 1

familiarisation with the data

• involves intensely reading the data, to become immersed in it’s content

Thematic analysis - stage 2

Generating initial codes

• or labels, identify features of the data important to answering the research question

Thematic analysis - stage 3

Searching for themes

• involves checking potential themes against the data to identify patterns of meaning

Thematic analysis - stage 4

reviewing themes

• involves checking the potential themes against the data, to see if they explain the data and answer the research question

• themes refined → splitting and combining or discarding one

Thematic analysis - stage 5

defining and naming themes

• involves detailed analysis of each theme and creating an informative name for each one

Thematic analysis - Stage 6

writing up

• involves combining together the information gained from the analysis

case studies

• detailed and in-depth investigations of a small group or individuals

• allow researchers of those who have undergone unique/rare experience that would unethical or impossible to construct

case studies - collecting data

• qualitative data through interviews, observations, questionnaires - good at reporting subjective, individual experiences

• quantitative data through memory tests, IQ tests, closed questions

• if more than one method is used, called triangulation

• most longitudinal

case studies - strengths

• provide rich, in depth data high in explanatory power

• holistic, ideographic approach where the whole individual is considered

• high ecological validity

• studying rare disorders or conditions allows researchers to form conclusions as to how the majority of the population functions e.g. long term memory of HM

case studies - limitations

• findings only represent the person or group who is the focus of the study, cannot be generalised

• may suffer from the relationship between the researcher and the participant

  • may feel too close, lose objectivity, use of bias in report

Aims

• a general statement covering the theory that will be investigated

• identifies the purpose of the study(straightforward)

• outlines what is being studies e.g. the effect of caffeine on memory

hypothesis

• testable statement written as a prediction of what the researcher expects to find

• precise and unambiguous

• two types: null hypothesis, alternative hypothesis

alternative hypothesis

• should include the independent variable and the dependent variable

• IV and DV should be operationalised - specify how each is to be manipulated(IV) and measured(DV)

• two types: directional and non directional

operationalising the IV and DV

IV e.g. two conditions: 200ml caffeine versus water

DV e.g. number of correctly recalled items

directional alternative hypothesis

• predicts that the alternative will be higher/lower than the null

• e.g. 200ml of caffeine will increase the number of items recalled

nondirectional alternative hypothesis

• predicts that there will be a difference

• e.g. there will be a difference in recall of the people who consumed 200ml caffeine versus 200ml of water

null hypothesis

• what all research starts with, idea that the IV will not affect the DV

• the null hypothesis assumes ‘no difference’

  • e.g. in the number of items recalled

• if experiment shows difference, null hypothesis is rejected

correlation and hypothesis

• instead of difference, have to use relationship/correlation

• e.g. there will be a [positive/negative/no] relationship between the number of cups of caffeine drank and the number of hours slept per night/week

independent variables

• the only variable that should be changed

• required to observe the effect it has on the DV

• laboratory experiment must use an IV that has been implemented by the researcher

  • cannot be naturally occurring(e.g. gender) → experiments that do=quasi/natural experiments

dependent variables

• variables that is measured to determine the outcome/access effect of the IV

• must be quantitative, numerical data can be displayed in a graph and analysed statistically

extraneous variables

• any factor that adversely effects the DV

  • e.g. time of day → morning people more alert

  • e.g. temperature → affecting performance on task

  • e.g. mood → events that affect mood and hence performance

• usually controlled, same effect across all conditions

  • ensure neutral ground

• researcher responsibility to control as much as possible to ensure it is objective and unbiased

• if not controlled, can become confounding

types of extraneous variables

Situational

• unfamiliar environment

• sound, light, temperature

• time of day

Participant

• substances

• general mood

• learning condition

• internal processes(illness, period)

• amount of sleep

• age

• general preferences

confounding variables

• affect DV and negatively impact research findings

  • e.g. time of day → run each trial at 8am, many unable to concentrate

  • e.g. temperature → room too cold for all trials, ppts more focused on keeping warm

• may not be apparent until after research process completed hence researcher should acknowledge in discussion part of report

demand characteristics

• interference between the research process and the participant that can adversely affect the research findings

how may demand characteristics occur

• e.g. ppts pick up on cues indicating what is expected of them and assume aim

• lab setting may cause unnatural response

• any communication - implicit or explicit

• toward researcher - to please/annoy

controlling demand characteristics

single blind procedure

• where ppts do not know which condition they’ve been assigned to

Investigator effects

• when researcher’s presence/behaviour interferes with process and becomes source of bias

how may investigator effects occur

• characteristics: age, gender, ethnicity could influence how ppts interact

• accent, tone of voice, non-verbal communication and what they’re wearing can impact participant reaction

• e.g. accent=stereotype, excited tone=lack neutrality, vibrant/patterns/slogans too personal=distraction

• could be biased in the way they instruct ppts/lead a task

controlling investigator effects

double blind procedure

• participants and researcher do not know which condition each participant has been assigned, hence unable to exercise any forms of bias

randomisation

• deliberate avoidance of bias to keep research as objective as possible

• keeping ppt allocation to conditions random, e.g. draw out of a hat or computer generated

• procedure aspects e.g. random lists of words(avoids unconscious bias)

benefits of randomisation

• eliminates investigator effects

• minimises participant variables

standardisation

• identical procedure set up in an experiment (e.g. questions in self report) across all conditions/participants → no unfair advantage

  • allows research to be replicable = more reliable

  • instructions, briefing, debriefing, number of participants/conditions, timing with each condition, materials(unless that is the IV)

pilot studies

small-scale trials run to test some or all aspects of the proposed investigation e.g. Milgram’s obedience study

• conducted before the research to identify any issues which could arise e.g. flaws in design, ethical issues, feasibility issues, to test for reliability/validity

pilot studies - reasons

• if issues are found, opportunity to fix/find suitable alternatives

• financial reasons - evidence to obtain funding

• if alternatives made must run another

• identify if its worth the time, money, and effort

quantitative data

focus is on numerical data, such as closed question questionnaires and experiments

• objective

• normathetic - general laws, large samples

qualitative data

focus is on non-numerical data, such as verbal reports

e.g. interviews and focus groups

• subjective, ideographic, private and personal, small samples, case studies

quantitative data - strengths

• objective - no bias

• quicker to administer and gather data

• can create visual representation

• easier to compare and verify

• can generalise to wider populations(tend to be large samples)

quantitative data - limitations

• does not provide rich and detailed info

• complex behaviour is reduced into numbers, so we’re not getting the full explanation

• emotions and feeling ignored

qualitative data - strengths

• provides rich and detailed information

• prospect of understanding rounded individuals

• subjective

• data often suggests testable hypothesis

qualitative data - limitations

• less accessible

• harder to transcribe, analyse and compare

• time consuming and expensive

• subjective, bias can be introduced (social desirability)

• harder to generalise, smaller samples

experimental methods

• allow us to establish cause and effect relationships

• can be in the form of laboratory, field, and quasi/natural experiments