Psychology - Research methods

AQA A-level Research Methods

experiments

can identify an IV (change) and DV (measure)

types of experiment

1. lab

2. field

3. quasi

4. natural

how to identify lab experiments?

• IV been manipulated by the researcher

• takes place in an artificial place/is an artificial task

advantages of lab experiments

1. high levels of control so more certain about cause + effect

2. high internal validity

3. easy to replicate as standardised procedure

disadvantages of lab experiments

1. lacks ecological validity as cannot be applied to a range of real world situations

2. lacks mundane realism

3. chance of demand characteristics

how to identify field experiments?

• IV been manipulated by research

• isn’t an artificial task or in an artificial place

advantages of field experiments

1. higher mundane realism

2. reduced demand characteristics

3. high external validity

disadvantages of field experiments

1. lack of control over extraneous variables - causality is difficult and less precise

2. ethical issues - no consent so might be invasion of privacy

how to identify quasi experiments?

• IV not manipulated by researcher because it CAN’T be manipulated

advantages and disadvantages of quasi experiments

depends on setting taking place in e.g. if was artificial or natural setting

how to identify natural experiments?

• IV not manipulated by researcher but it COULD be

advantages of natural experiments

1. high external validity

2. provide research opportunities that may not otherwise be possible for practical or ethical reasons

disadvantages of natural experiments

1. rare

2. hard to replicate

3. lack of control

4. participants cannot be randomly allocated to experimental conditions

aim of research

a general statement of the purpose of the research

hypothesis

a precise and testable statement about the assumed relationship between variables

• MUST operationalise DV and give both conditions for the IV

directional hypothesis

“the (DV) in the (IV condition 1) is better/faster/bigger/less/fewer etc. than the (DV) in the (IV condition 2)”

undirectional hypothesis

“There is a difference between (IV condition 1) and (IV condition 2)”

hypothesis for correlation

“there is a positive/negative relationship/a relationship between (both co variables)”

null hypothesis

no relationship between 2 variables

random sampling

every member of the target population has an equal chance of being selected e.g. obtain all names of population, write on separate pieces of paper and draw the number required for the sample

adv of random sampling

least biased so representative of the target population

disadv of random sampling

still a chance it could produce a biased sample and is time consuming

systematic sampling

taking every nth person from a list of the population e,g, select every 10th person

adv of systematic sampling

unbiased method and it is likely to be representative (results = generalizable)

disadv of systematic sampling

by chance could generate a biased sample

stratified sampling

subgroups (or strata) identified within the population and then random samples are taken from each strata

• must know the proportions of the strata and then calculate how many people necessary and then pull necessary number out of a hat

adv of stratified sampling

more representative as there is a proportional representation of subgroups

disadv of stratified sampling

time consuming

opportunity sampling

makes use of people readily available and willing to take part

adv of opportunity sampling

quick and easy as no actual selection process and is sometimes the only available possible method

disadv of opportunity sampling

inevitably biased as the sample is drawn from a small part of the population

volunteer sampling

people volunteer or put themselves forward for research in response to a newspaper or on an advertisement

adv of volunteer sampling

fairly quick and easy and can target specific participants who are required

disadv of volunteer sampling

not likely to be representative of a target population as volunteers tend to be a certain type of person e.g. confident

pilot study

small-scale trial run of a research design in order to find out if any aspects of the design do not work - can be adjusted preventing large amounts of time and money being wasted

pilot studies in observations

make sure the behavioral categories are suitable, observers are consistent in what they see and interpret behaviors in the same way and cannot be seen

pilot studies in questionnaires/interviews

questions can check to be clear, unambiguous, not misleading and not offensive

pilot studies in experiments

check experimental design is suitable, instructions are clear and check if demand characteristics could become a problem

independent group design

each group of participants completes one condition

adv of independent group design

less chance of demand characteristics and order effects as only complete one condition

disadv of independent group design

potential problem of participant variables as different people and twice as many participants needed (increases time and money spent)

how to deal with participant variables

random allocation

1. names out of a hat

2. evenly distributes participant characteristics across the conditions of the experiment used random techniques

3. put names into hat, 1st into one, 2nd into the other until all assigned to a group

repeated measures design

one group of participants completes both (all) conditions of the experiment

adv of repeated measures design

no problem of participant variables as same people and less participants needed

disadv of repeated measures design

more chance of demand characteristics and order effects

how to deal with order effects?

counterbalancing

1. participants split in half

2. one half do condition 1 then 2

3. other half do condition 2 then 1

matched pairs design

two groups of participants are matched to each other on relevant characteristics then one from each pair completes one condition

adv of matched pairs design

less problem of participant variables as similar people and less chance of demand characteristics and order effects as one condition only

disadv of matched pairs design

time consuming to match participants and may not be entirely successful (unless identical twins)

behavioural categories

used as a ‘checklist’ as the behaviour is observed in an observation - impossible to analyse a whole stream of behaviour and needs to operationalise behaviour (tally chart)

time sampling (observation)

observations made at regular time intervals (e.g. every 10 mins for an hour) and record any behaviour occurring (good if one participant and want a comprehensive idea of their behaviour)

event sampling (observation)

observation lasts for a certain length of time and includes the whole time so don’t miss any behaviours - use of categories and tally behaviour

closed questions

choice of pre-determined answers e.g. yes, no, strongly agree etc (always include these options)

adv of closed questions

easier to analyse and represent in a graph

disadv of closed questions

not always representative as may not be one of the options etc.

open questions

allows the respondent to answer in their own words

adv of open questions

more descriptive and involve the collection of qualitative data and often quantified using content analysis

disadv of open questions

difficult to analyse due to range of results so time-consuming

design of interviews

1. how will it be recorded?

2. interviewer effects may impact outcome so must consider characteristics

operationalisation

making variables measurable

extraneous variables

unwanted variables that could affect DV so must be controlled (may confound/confuse the results)

examples of extraneous variables

1. participant variables (IQ, age etc.)

2. environmental variables (distraction, noise etc.)

3. experimenter variables

confounding variables

uncontrolled extraneous variables that confuse the results by affecting the DV

standardisation

controlling the extraneous variables - same room, same experimenter, same instructions

randomisation

present all conditions muddled up so the order occurs by chance

participant effects

if know they are being studied they may behave differently due to social desirability bias

demand characteristics

participants try to figure out the aim of the study and act accordingly so may be too cooperative etc.

single blind procedure

participants are not fully informed of the true nature of the research

investigator effects

influencers may influence the participants e.g. leading questions

double blind procedure

participants and investigator unaware of the research aim

informed consent (ethical issues)

participant must be told sufficient details so make their own choice (may cause participant variables = cause deception where incorrect details given or information withheld)

how to deal with informed consent

1. presumptive consent - gained from similar group of people

2. prior general consent - agree to be deceived without knowing how

3. retrospective consent - asking participants after they have participated (ONLY if confident yes)

4. debriefing - occurs after the study and involves giving all relevant details including aim of study

protection from harm (ethical issues)

protected from physical and psychological harm e.g. anxiety, lowered self-esteem - leave in same state they arrived in

how to deal with protection from harm

debriefing - occurs after the study and may include support/counselling as required if participant are harmed

right to withdraw (ethical issues)

should be able to leave the study at any time even if agreed to continue

how to deal with the right to withdraw

state this option at the start and remind them during the study - be paid for their contribution even if they drop out

confidentiality (ethical issues)

all data should be protected and kept confidential and must agree if information published

how to deal with confidentiality

numbers or pretend names used

privacy (ethical issues)

should not be observed if unaware

how to deal with privacy

only occur in public places where a participant could expect to be observed anyway

quantitative data

numerical, lacks detail, from experiments, correlations, observations which use categories and closed questions

qualitative data

non-numerical, rich in detail and occurs through descriptions of behaviours or attitudes e.g. open questions on interview or case studies

mean

adding all the numbers in a set of data and dividing by the size of the sample

median

worked out by first putting all scores in order of size, starting with smallest

mode

score that occurs most often

adv of mean

1. every bit of data is used in its calculation (sensitive measure) and most accurate

2. measure of central tendency that should be used with interval data

disadv of mean

1. distorted by extreme scores (outliers)

adv of medians

1. not distorted by extreme scores (outliers)

2. measure of central tendency with ordinal data

disadv of medians

1. not a sensitive measure as does not include all scores

adv of mode

1. data is collected in categories (nominal data) so a mean would not make sense

disadv of mode

1. can be two modes - bimodal or no modal valuerang

measures of dispersion

measure of the (variability) spread of scores

range

subtracting the lowest value from the highest value (smaller = more reliable)

standard deviation

measure of the spread of a set of scores from the mean (larger = larger spread of scores)

adv of range

simple to calculate

disadv of range

distorted by extreme values

adv of standard deviation

sensitive measure as every score is used

disadv of standard deviation

more complicated to calculate as requires a formula

correlational data

ranges from -1.0 to +1.0 (correlation coefficient) - the closer to 0 = weaker the correlation (0.1-0.3 = weak, 0.4-0.6 = moderate, 0.7-0.9 = strong with 1 and -1 being perfect)

scattergrams

display correlationdal data as there is no IV or DV (just co-variables) - cannot establish cause and effect - positive correlation = as one goes up, the other goes up with negative = as one goes up, the other goes down