Research Methods: Part 1 Understanding RM Flashcards
1
Q
What are the 4 types of research methods?
A
- Experiments
- Observations
- Self Reports
- Correlations
2
Q
Name the different types of experiments.
A
- laboratory experiments
- field experiments
- quasi / natural experiments
3
Q
Name the different types of observations.
A
- naturalistic / controlled
- participant / non-participant
- structured / unstructured
- covert / overt
4
Q
Name the different types of self report techniques.
A
- questionnaires
- interviews
5
Q
Name the different types of correlational analysis’.
A
- positive correlations
- negative correlations
6
Q
Describe laboratory experiments
A
- run in carefully controlled conditions
- use a standardized procedure and allows experimenter full control of variables
- minimize extraneous variables
- cause and effect can be established (IV and DV)
7
Q
Describe field experiments
A
- IV under direct control of experimenter
- DV measured
- setting more like PP’s natural and usual environment
- take place outside lab
8
Q
Describe quasi / natural experiments
A
- researcher makes use of naturally occurring variables
- not a true experiment because scientists cannot manipulate IV
- often used when other lab methods are impractical or unethical
9
Q
What are used to allocate participants to the different levels of the IV in an experiment, and name all the different types of them.
A
Experimental designs:
- independent measures design
- repeated measures design
- matched participants design
10
Q
Describe independent measures design.
A
- two independent groups of participants, so data of each is independent of each other
- eliminates demand characteristics✅
- may be individual differences among participants so could be a confounding variable e.g. age, IQ, gender❌
- researchers try to minimize this by having a large sample and randomly allocating PP’s to conditions✅
11
Q
Describe repeated measures design.
A
- each PP takes part in both conditions of experiment
- minimizes individual differences as uses same people✅ and also needs less people✅
- means demand characteristics, so people may perform better or worse because they have practiced test before, and may start to guess aim of experiment so changes behavior❌
- solution is to counter balance order in which PP’s do the tasks
12
Q
Describe matched pairs design.
A
- similar to independent group, but with participants carefully matched on a number of variables e.g. Age / IQ/ gender (identical twins are ideal)
- can control individual differences and order effects✅
- matching is extremely difficult, so may still be individual differences among PP’s, even twins!❌
13
Q
What is operationalizing variables?
A
- important that variables are clearly defined so that cause and effect can be established and studies can be replicated
- so IV must be stated precisely and accurately
- DV needs to be clearly measured and quantifiable.
14
Q
What are demand characteristics?
A
- any features of the experimental setting that indicate to participants the aim of the study so may change or influence their behavior
- so can seriously affect findings❌
15
Q
What is a single blind procedure?
A
- experimental procedures in which only experimenter knows the aim of the experiment
- so ensures pp’s are unaware of level of IV in which they are performing
- reduces demand characteristics✅
16
Q
What is a double blind procedure?
A
- experimental procedures that ensures neither the researcher or pp’s are aware of condition of IV they are in or the aim of the experiment
- reduces demand characteristics✅
- reduces investigator / researcher effects✅
17
Q
What are investigator / researcher effects?
A
- the researcher may influence the participant and therefore the findings, by for e.g. their presence, beliefs, gender or behavior etc
18
Q
What are extraneous / confounding variables?
A
- can be anything other than the IV that may also influence the DV e.g. Temp when running
- if DV IS influenced by an extraneous variable in anyway it becomes a confounding variable
- compromise a study’s validity and reliability❌
19
Q
Assess reliability in all experiments.
A
- lab experiments generally reliable due to high level of control, standardized procedures and treating all pp’s in same way
- field experiments + quasi / natural experiments are less reliable as variables of pp’s normal environment are harder to control
20
Q
Assess validity in laboratory experiments.
A
- extraneous variables tightly controlled so researchers can be confident changes to IV cause changes to DV so = high validity
- sometimes way DV is measured is quite artificial so has a lack of mundane realism so = lack of validity
- demand characteristics may mean behavior changes, also artificial nature of lab experiments = lower ecological validity.
21
Q
Assess validity in field experiments.
A
- higher ecological validity than lab experiments as less artificial and pp’s more like to display more natural behavior
- demand characteristics also less likely in natural settings so increases validity
- extraneous variable are harder to control and DV more difficult to measure precisely so internal validity can be compromised though.
22
Q
Assess validity in quasi / natural experiments.
A
- have the least control and so less certainty that it’s changes to IV causing effects to DV, so lowering validity
- however behavior is more natural, so demand characteristics are less likely, increasing validity
- ecological validity = very high as behavior is occurring in its natural environment and observed differences are the representation of real life.
23
Q
Evaluate lab experiments.
A
✅- can confidently say IV causes change in DV with careful control, so has internal validity
✅- experiment is replicable and therefore reliable
✅- variables are easier to control under lab conditions
❌- more controlled, the less natural so lacked ecological validity
❌- sometimes impossible to use lab experiment
❌- demand characteristics may reduce confidence in results
24
Q
Evaluate field experiments.
A
✅- less artificial so mundane realism usually higher so higher internal validity
✅- avoids participant effects because they aren’t aware, so increases internal validity
❌- extraneous variables less easy to control as experiment is taking part in real world, so reduces internal validity
❌- may still be demand characteristics
25
Evaluate quasi / natural experiments.
✅- allows research where IV can't be manipulated for ethical or practical reasons
✅- enables psychologist to study 'real' problems so increases mundane realism and validity
❌- cannot demonstrate casual relationships as IV is not directly manipulated
❌- many extraneous variables which is a threat to validity
❌- participants may be aware of beings studied and cause participant effects and demand characteristics
26
Evaluate independent measures design.
✅- diff participants used in each level of IV so no possible order/practice effects
✅- demand characteristics reduced as participants see task only once
❌- individual diffs between pp's in diff levels of IV e.g. Age, gender, IQ etc can distort results making them less accurate and not as consistent, so diff to compare and conclude from
❌- more pp's needed, so may be less ethical or harder to find
27
Evaluated repeated measures design.
✅- individual diffs unlikely to distort effect of IV as pp's do both levels
✅- uses fewer pp's so is good when pp's are hard to find or recruit, e.g when they know true nature of experiment less may volunteer
❌- order effects e.g. Practice fatigue and extraneous variables can affect results
❌- pp's see experimental task more than once, increasing demand characteristics I.e. May figure out they are being tested on something the didn't know about
28
Evaluate matched participants design.
✅- pp's see task only once, reducing demand characteristics e.g. Second guessing true nature of experiment
✅- controls for individual diffs, e.g. Identical twins are ideal as have identical genes so likely to answer similarly if put in same condition, so should provide reliable results when split in diff conditions as almost like one person
✅- results cannot be distorted as there are no order effects e.g. Practice fatigue
❌- similarity between pairs is limited by matching process
❌- matching pp's is time consuming in order to be accurate, and difficult as necessary info required
29
Describe naturalistic or controlled observations.
Naturalistic: people are observed in their everyday (natural) setting / environment.
Controlled: some variables are controlled / manipulated by researcher, may be conducted in pp's normal environment or an artificial situations e.g. Lab
30
Describe structured/systematic or unstructured observations.
Structured/systematic: has an aim to record and categorize behavior as accurately as possible, using predefined and decided categories, often using a coding system
Unstructured: researcher records a wide range of all relevant behavior but has no particular system in which to do this.
31
Describe Participant or Non-participant observations.
Participant: researcher is engaged with pp's as part of the social setting
Non- participant: researcher is not engaging with participants and collects data by simply observing but not interacting too.
32
Describe covert or overt observations.
Covert (non disclosed): pp's are totally unaware they are being watched.
Overt (disclosed): research in which the role of the observer is know to pp's.
33
What methods can be used to collect data in observations?
Event sampling
| Time sampling
34
Describe event sampling.
Uses a checklist of possible activities which are tallied as they occur.
35
Describe time sampling.
3 types:
- instantaneous scan sampling: only behavior during time interval is recorded
- predominant activity sampling: recording behaviour performed most during time interval
- one-zero sampling: recording whether behaviour occurred or not during interval.
36
Assess reliability in observations.
- Clearly operationalized coding systems increase reliability
- check inter rater reliability to increase reliability
- coding systems improve reliability
- using pilot studies to check behaviour categories increases reliability
37
Assess validity in observations.
- covert observations increase validity as pp's don't change their behaviour
- double blind observations reduce observer bias so increase validity
- clearly operationalized coding systems increase validity
38
Describe inter-rater reliability.
High if researchers observe and code behaviour in same way as each other.
Will be low of they're not observing the same event or coding behaviour is vague and lacks validity.
39
Describe designing observations using behavioural categories.
- involves identifying clearly defined behaviour which can be observed and recorded
- may be place on a checklist and tallied
40
Describe designing observations using coding frames.
- coding frames allow for more specific behaviors to be observed with in a behaviour category
- codes and abbreviations used to record the severity of behaviors or a diff sub type within a category
41
Evaluate time sampling.
✅- can give an indication of order in which events happen
✅- instantaneous scan sampling can be highly reliable
✅- one zero sampling can record occurrence of infrequent behaviors
❌- even with computers, diff to record as many different behaviors as can be recorded in event sampling
❌- predominant activity sampling only provides a relative estimate of time spent on each behavior, not an actual measure
42
Evaluate event sampling.
✅- can record every occurrence of each behaviour to give a complete record
✅- records are easy to obtain and easy to analyze as they are just totals
❌- gives no indication of relative time spent on each behavioural category
❌- gives no indication of order in which events from each behavioural category occur
43
Evaluate natural observation.
✅- participants not aware they are being observed so more likely to react in realistic and genuine ways
❌- extraneous variables threaten validity as can rarely be controlled
44
Evaluate controlled observation.
✅- extraneous variables can be controlled so validity can be high compared to natural observations.
❌- participants are in an unfamiliar environment so may react differently to normal behaviour
45
Evaluate structured / systematic observation.
✅- operational definitions can be developed in a pilot study before real data collection so improves validity
❌- predetermined behavioural categories may be limiting if new behaviors become apparent during the study
46
Evaluate unstructured observation.
✅- without limitation of specific behavioural categories, any relevant behaviors can be recorded so data are richer and more complete
❌- by attempting to record everything, observers may ignore or miss important aspects
47
Evaluate participant observation.
✅- being involved in social group can give observer insight into participants real emotions and motives
❌- can make observer subjective
48
Evaluate non participant observation.
✅- equipment can be used for data recording to make detailed and accurate records if non participant observers are covert e.g. behind a screen
❌- if pp's are unaware they are being observed it raises ethical issues and questions about informed consent and right to withdraw
49
Evaluate overt observation.
✅- as participants are aware of observers presence, the technique is more ethical than covert
❌- pp's are aware they are being observed so responses may not reflect what they would normally do
50
Evaluate covert observation.
✅- pp's less likely to be aware they are being observed so more likely to react in a natural way
❌- if covert observers are also participant observers, may be difficult for them to record data accurately and unobtrusively without pp's finding out
51
What are the two types of self report techniques?
- Interviews: 3 types of these
| - Questionnaires
52
Describe interviews and the different types of them.
- method in which participants reply verbally to questions asked by an interviewer
- like a questionnaire but responses are oral
- usually face-to-face
- can be structured, semi structured or unstructured
Structured: involves mainly closed q's in a fixed order, q's are normally scripted so are standardized.
Semi structured: uses a fixed list of open and closed q's however interview can introduce additional questions and topics if required.
Unstructured: generally begin with a standard q for all pp's but from there on q's depend on individuals responses, may have a list of topics to cover though.
53
Describe questionnaires.
- pp's record their answers to a list of pre-written questions
- can include open q's (where free responses are allowed), or closed (where a fixed no of responses are offered to choose from)
- rating scales where respondents are asked to give their views on a number scale, or maybe likert scales
54
Describe quantitative research.
- to do with quantity
- focuses on collecting numbers
- V controlled
- analysis is objective
55
Describe qualitative research.
- to do with quality
- focuses on collecting opinions
- is less controlled
- analysis can be subjective
56
Describe rating scales.
- a numerical scale on which a participant selects a number to indicate their choice
- provides quantitative data
- can give a numerical answer to a question, or just indicate the extent to which pp agrees with statement.
57
Describe likert scales.
- type of q that measures attitudes using a statement
- pp's respond to statement by choosing an option (typically from choices of 'strongly agree', 'agree', 'don't know', 'disagree', 'strongly disagree', to indicate how far they agree with the statement.
58
Describe open questions.
- allows respondents to provide their ideas and views in detailed answers
- generate qualitative data
59
Describe closed questions.
- answers are listed and respondents choose their answer from these
- no opportunity to expand on answers
- generate quantitative data
60
Describe a pilot study.
- small scale trial run of a method to identify any problems with procedure before they can occur in large scale experiment.
61
Describe semantic differential scales.
- participants indicate their response views on a scale between two extremes (contrasting adjectives) e.g. Delicious, yummy, nice, okay, edible, disgusting, vomiting induced
(Delicious - vomiting induced).
62
Describe social desirability bias.
- respondents may not be honest when answering questions as afraid of being judged or not agreed with
- may want to please researcher with answers so give answers they think researcher wants to hear
- particularly true with sensitive topics e.g. Views on racism or disability or hygiene.
63
Describe internal reliability, external reliability and inter-rater reliability.
Internal: extent to which something is consistent within itself e.g. All questions on an IQ test actually measure IQ.
External: extent to which something is consistent overall on several occasions e.g. A questionnaire that produces same results over several occasions.
Inter-rater reliability: when researchers observe and code same behaviour in same way.
64
Assess reliability in self reports, and describe two tests that can be used to check it.
Test - retest: tests external reliability by testing same individual with same measures over a period of time, if results are same, reliability is high.
Split half: tests internal reliability by splitting results of a test in two and testing how consistent they are, if both have similar scores, reliability is high.
65
Assess INTERNAL validity in self reports, and describe the four types of it.
Internal validity: concerned with whether changes in DV are caused by only IV and not sources of error e.g. Extraneous variables:
- face validity: whether a measure appears, on face value, to measure what it intends to measure.
- criterion validity: extent to which test relates to or predicts some other variable
- concurrent validity: how well current questionnaire matches with a previously well established one on same topic
- construct validity: to what extent performance on psychological test measures the identified theory.
66
Assess EXTERNAL validity in self reports, and the sub type of it.
External validity: concerned with whether or not researcher can generalize to situations beyond current context:
- population validity: extent to which findings from one study can be generalized to whole population.
67
Evaluate interviews.
✅- structured interview data easy to analyze as is quantitative
✅- semi structured or unstructured enable interviewer to gain specific and detailed info
❌- structured are limited by fixed questions
❌- investigator bias is a problem, e.g. Social desirability in answering questions.
68
Evaluate questionnaires.
✅- data easy to analyze as usually quantitative
✅- relatively easy to administer as can be sent or emailed to participants, so time and cost efficient
❌- response bias e.g. Tending to always answer no or just ticking box on the left leads to invalid results
❌- limited and no flexibility like unstructured interviews
69
Evaluate rating scales.
✅- easy to respond to so large amounts of data can be collected quickly
✅- produce quantitative data so easy to analyze
❌- cannot be used to measure complex variables that require more than one simple numerical response
❌- risk of response bias e.g. Consistently giving middle scale answers
70
Evaluate likert scales.
✅- allow measurement of more complex attitudes than rating scales
❌- meaning of middle value is ambiguous, could indicate no opinion or undecided.
71
Evaluate open questions.
✅- produce qualitative and detailed data, expressing pp's views fully so raising validity
❌- findings are individual so may be less generalizeable
72
Evaluate closed questions.
✅- easy to respond to so collects more data quickly
✅- quantitative data produced so easy to analyze
❌- risk of response bias e.g. Saying yes consistently
❌- produce only quantitative data which lacks detail and meaning so lowers validity.
73
What are the types of correlations?
Strong positive: increase in one accompanies increase in other
Weak positive: "
Strong negative: increase in one accompanies decrease in other
Weak negative: ""
Perfect positive: "
Perfect negative: ""
No correlation
74
Describe scatter diagrams.
- graph used to display data from a correlational study
| - each point on graph represents participants score for each of two measured variables.
75
What is a casual relationship?
A relationship between two variables where a change in one is responsible for a change in the other.
76
What is a correlation coefficient?
- measure of the strength of a correlation
- usually expressed as a value between -1 and 1
- 0 - 1 means positive correlation
- -1 - 0 means negative correlation
- closer to 1 or -1, the stronger the correlation
77
Assess reliability in correlations.
- correlations using scientific scales will be highly reliable
- correlations in variables measured using self reports, observations or estimates will be less reliable
78
Assess validity in correlations.
- more reliable correlation = more valid
| - valid if both variables measure real phenomena in effective ways, so variables must be defined and measured clearly.
79
Evaluate correlations.
✅- can be conducted in variables that can be measured but not manipulated
✅- can demonstrate presence or absence of a relationship
❌- can be used only with variables that can be measured on a scale
❌- a single correlational analysis cannot indicate whether a relationship is actually casual
