Research methods Flashcards
Quantitative research
Aim:
- nomothetic approach (derive universally applicable rules)
- these rules may be applied to the behaviour of large groups of individuals
Focus: behavioural manifestations (operationalisations)
Data: Numbers
Objectivity: more objective- the researcher is eliminated from the studied reality
Types:
- experiment
- quasi-experimental
- correlational study
Qualitative research
Aim:
- idiographic approach (in-depth understanding of a particular case or phenomenon)
- obtained knowledge isn’t a universal law, but it’s deeper in the sense that a particular case is understood more holistically
Focus: Human experiences, interpretations, meanings
Data: Texts
Objectivity: more subjective- researcher is included in the studied reality
- researcher is an integral part of the procedure and a “tool of measurement”
Types:
- observation
- interview
- focus group
- case study
Sample
a group of individuals taking part in the research study
Define sampling
process of recruiting individuals for participation
Define credibility
extent to which results of the study can be trusted to reflect the reliability
- study is credible when there are reasons to believe that its findings are true
Qualitative research study:
Credibility = trustworthiness
Quantitative research study:
Credibility = internal validity
Define bias
- characterises various distortions introduced to the findings by the researcher, research procedure, mistakes in process of measurement etc.
Define generalisability
extent to which results of the study can be applied beyond sample and setting used in the study itself
Sampling in quantitative research
Experimental studies and correlational studies
- random sampling
- stratified sampling
- self-selected sampling
- opportunity sampling
Sampling in qualitative research
- quota sampling
- purposive sampling
- theoretical sampling
- snowball sampling
- convenience sampling
Two types of quantitative research
- Experimental studies
2. Correlational studies
Generalisability in quantitative research
Experimental studies:
- External validity: ecological validity and population validity
- Construct validity
Correlational studies:
- population validity
- construct validity
Generalisability in quantitative research
- sample-to-population generalisation
- case-to-case generalisation
- theoretical generalisation
Credibility in quantitative research
Experimental studies:
- referred to as internal validity
- ways to improve this: controlling confounding variables; eliminating or keeping constant in all conditions
Correlational studies:
- referred to as credibility
- ways to improve: using reliale ways to measure the variables; avoid biases in interpreting results
Credibility in qualitative research
- referred to as credibility/trustworthiness
- ways to improve this: triangulation, establishing a rapport, iterative questioning, reflexivity, credibility checks, thick descriptions
Bias in qualitative research
Experimental studies- Threats to internal validity:
- selection, history and maturation
- testing effect and instrumentation
- regression to the mean
- experimenter mortality
- experimenter bias
- demand characteristics
Correlational studies:
- While measuring variables: depends on the method of measurement
- While interpreting findings: curvilinear relationships, third variable problem, spurious correlations
Bias in qualitative research
Participant bias:
- acquiescence
- social desirability
- dominant respondent
- sensitivity
Researcher bias:
- confirmation bias
- leading questions bias
- question order bias
- sampling bias
- biased reporting
Variable
any characteristic that is objectively registered and quantified
Construct
any theoretically defined variable eg. violence, attraction, memory, anxiety
- constructs need to be operationalised
What does it mean to operationalise a construct?
means expressing the construct in terms of observable behaviour
A good operationalisation will:
- capture the essence of the construct
- be clearly measurable
Independent variable
variable that is manipulated by the experimenter
Dependent variable
variable that changes as a result of the manipulation by the experimenter- the one that is measured
Confounding variables
other variables (other than IV and DV) that can interfere in the relationship between the IV and the DV - this is to ensure that it is the change in the IV that causes the change in the DV
Target population
the group of people to which the findings of the study are expected to be generalised
- sample: group of people that take part in the experiment; a sub-set of the target population
Target population and generalisability
- results of quantitative research need to be able to to be generalised from the sample to the target population
- for this to be possible, sample must be representative of the target population
- a sample is representative if it reflects all of the essential characteristics of the target population
Sampling techniques used in quantitative research
- Random sampling
- Stratified sampling
- Convenience (opportunity) sampling
- Self-selected sampling
Random sampling
- What it is
- Advantages
- Disadvantages
What it is:
- create a list of all members of the target population and randomly select a sub-set
- this way every member of the target population has an equal chance of being a part of the sample
Advantages:
If sample size is sufficient, researchers may be certain that even unexpected characteristics are fairly represented in the sample
Disadvantages:
It’s practically impossible to carry out truly random sampling
eg. target population may be geographically dispersed
Stratified sampling
- What it is
- Advantages
- Disadvantages
What it is:
- decide on the list of essential characteristics of the population that the sample has to reflect
- then study the distribution of these characteristics in the target population
- then recruit participants randomly, but in a way that keeps the same proportions in the sample as observed in target population
Advantages:
- allows researchers to control representativeness of some key characteristics without relying on chance
- useful when: researcher is certain about which characteristics are essential; sample sizes aren’t large
Disadvantages:
- requires more knowledge about the characteristics of the target population
- harder to implement
Convenience (opportunity) sampling
- What it is
- Advantages
- Disadvantages
What it is:
- recruiting participants that are easily available
Advantages:
- useful when financial sources are limited
- in some studies, there may be a reason to believe that people aren’t that different
- useful when the generalisation of findings isn’t the primary purpose of the study
Disadvantages:
- generalisation is very limited due to sampling bias
Self-selected sampling
- What it is
- Advantages
- Disadvantages
What it is:
- recruiting volunteers
- anyone that wants to participate is included in the sample
Advantages:
- a quick and easy method to recruit participants
- at the same time has wide coverage
Disadvantages:
- representativeness and generalisation are limited
- a typical volunteer is more motivated than the average participant from a bigger population
- volunteers could pursue monetary incentives for their participation
3 types of experimental design
- based on how the independent variable is manipulated
- Independent measures
- Matched pairs
- Repeated measures
Independent measures design
- advantages
- disadvantages
- how to overcome the disadvantages
- IV is manipulated by randomly allocating participants into different groups
- rationale behind random group allocation: all potential confounding variables cancel each other out
Advantages:
- can have multiple groups
- participants only take part in one condition, so no order effect and more difficult for them to figure out true aim of the study
Disadvantages:
- participant variability: different people are used, likely that participants in the groups won’t be completely equivalent at the start of the study
How to overcome the disadvantages:
- when allocation into groups is random and groups are large enough, it’s likely that pre-existing individual differences will cancel each other out and groups on average will be equivalent
Matched pairs design
- advantages
- disadvantages
- how to overcome the disadvantages
- researchers use matching to allocate participants into different groups
- participants are assessed on a matching variable
- all participants are ranked according to the matching variable and allocated randomly into groups pairwise as we move along the ranks
- the participant then allocates each pair randomly into groups, until all participants have been allocated
- this way researcher ensures that the two groups are equivalent in terms of this one variable, and all other characteristics are kept random
Advantages:
- useful when researcher is particularly careful about certain confounding variables and wants to keep them constant in all groups
- useful when sample size isn’t large and there’s a chance that random allocation will end up producing groups that aren’t equivalent
Disadvantages:
- more difficult to implement because matching variables need to be measured first
- theory-driven: researcher needs to know what variables are likely to be confounding
How to overcome disadvantages:
- keeping the experiment simple
- matching is easier to implement when there’s one matching variable and 2 groups
Repeated measures design
- advantages
- disadvantages
- how to overcome the disadvantages
- same group of participants is exposed to 2 (or more) conditions and the conditions are compared
- this way participants are compared to themselves (also called “within-subject” designs)
Advantages:
- participant variability isn’t a problem as participants are compared to themselves
- means that sample sizes can be smaller
Disadvantages:
- order effects: fatigue or practice
- participants take part in more than one condition, increases the chances that they’ll figure out the true aim of the study
How to overcome the disadvantages:
- counterbalancing
- however, this is difficult when there are many conditions
How is the quality of experiments determined?
Quality of an experiment is characterised by its:
- construct: characterises generalisability of results
- internal validity: credibility of the experiment
- external validity: characterises generalisability of results
Construct validity
a characteristic of the quality of operationalisations
- operationalisations express constructs in terms of observable behaviour
- is high if operationalisation provides sufficient coverage in the construct
- relates to the overarching concept of generalisability; it characterises generalisability of findings to the theory
External validity
- a characteristic of generalisability of findings to other people and other situations
Two types:
- population validity
- ecological validity
Population validity
- the extent to which findings can be generalised from the sample to the target population
- it depends on how representative the sample is
Ecological validity
- the extent to which the findings can be generalised from the experiment to other settings or situations
- depends on how artificial the experimental procedure is
Lab experiments:
- participants find themselves in situations that don’t normally occur in their daily lives
- this can change their behaviour, making it less natural
The more closely the experimental procedure approximates real-life situations, higher the ecological validity of the experiment
Internal validity
- a characteristic of the methodological quality of an experiment
- it relates to credibility of the research study
- it’s high when confounding variables have been controlled and are certain that it was the change in the IV that caused the change in the DV
- it links directly to bias
- the less bias there is, higher the internal validity of teh experiment
Internal validity and ecological validity
Usually there is an inverse relationship between internal validity and ecological validity
- when internal validity is high, ecological validity is low
- when ecological validity is high, internal validity is low
Threats to internal validity (RED TIME)
- Regression to the mean
- becomes a threat when initial score on DV is extreme (v. low or v. high)
- Counteracted: a control group w/ same starting score on the DV, but no experimental manipulation - Experimental mortality
- occurs when some participants drop out of the experiment; is only a problem when rate of dropping out isn’t the same in every experimental condition
- Counteracted: Design experimental condition so participants don’t feel discomfort causing them to withdraw - Demand characteristics
- occurs when participants understand true aim of the experiment and alter their behaviour (unintentional/ intentional) as a result
- a bigger problem in repeated measures design as participants take part in more than one condition
- Counteracted: deception to conceal true aim of the study (but ethical considerations arise); post-experimental questionnaires to investigate extent to which participants could guess true aim of the study - Testing effect
- first measurement of DV may affect subsequent measurements; sometimes in independent measures, DV is measured twice eg. before and after experiment
- in repeated measures designs testing effect is a special case of order effects
- Counteracted: in independent measures designs there must be a control group, the same test and retest, but no experimental manipulation; in repeated measures, counterbalancing must be used - Instrumentation
- occurs when instrument measuring DV changes slightly between measurements, compromising standardisation of measurement process
- Counteracted: standardise measurement conditions as much as possible across all comparison groups and all observers - Maturation
- natural changes that participants go through in the course of the experiment eg. fatigue or growth (if procedure is extended in time)
- Counteracted: have a control group; if we can assume that rates of maturation are the same in both groups, comparison won’t be affected - Experimenter bias
- occurs when researcher unintentionally influences participants behaviour and results of the study
- Counteracted: use a double-blind design; neither participants nor experimenter knows who has been assigned to what condition
What are the different types of experiments?
- True experiment
- Quasi-experiment
- Laboratory experiment
- Field experiment
- Natural experiment
True experiment
- allocation into experimental groups is done randomly
- researchers can assume that IV is the only difference between the two groups
- allows researcher to interpret results of the study as a cause-effect relationship (IV influences the DV)
Quasi-experiment
- allocation into groups is done on the basis of pre-existing differences
- researchers can’t be sure that the groups are equivalent in all other characteristics
- hence, because IV isn’t manipulated by the researcher, cause-effect inference can’t be made
- in the way they’re made, quasi-experiments resemble experiments (they involve a comparison of groups)
- but, from the POV of possible inferences, they are correlational studies
Laboratory experiment
- conducted in highly controlled, artificial conditions
- since confounding variables are better controlled this way, it increases internal validity but compromises ecological validity
Field experiment
- conducted in real-life settings
- researcher manipulates IV; but, as participants are in their natural setting many confounding variables can’t be controlled
- this increases ecological validity and decreases internal validity
Natural experiment
- conducted in participants’ natural environments
- IV isn’t manipulated by the researcher but occurs naturally
- advantage is there’s ecological validity; also they can be used when it’s unethical to manipulate the IV
- disadvantage is internal validity due to a large no. of confounding variables that are impossible to control
Correlational studies
- no variable is manipulated by the researcher
- cause-effect inferences can’t be made
- 2 or more variables are measured and the relationship between them is mathematically quantified
- cause-effect relationships can’t be made
Correlation
a measure of linear relationship between 2 variables
- correlation coefficient can vary from -1 to +1
- correlation close to 0 means there’s no relationship between 2 variables
Negative correlation: there’s an inverse relationship between 2 variables, higher the A, lower the B
Positive correlation: means a direct relationship; higher the A, higher the
Characterised by 2 parameters:
- effect size
- statistical significance
Effect size
the absolute value of the correlation coefficient (no. from 0 to 1)
- shows how large the correlation is
Correlation coefficient effect size (r) and its interpretations
Less than 0.10 = negligible
- 10-0.29 = small
- 30-0.49 = medium
- 50 and larger = large
Statistical significance
shows likelihood that a correlation of this size has been obtained by chance
- if this likelihood is less than 5%, correlation is accepted as statistically significant
Probability that the result is due to random chance and its interpretation
p = n. s result is non-significant
p < 0.5 result is statistically significant
p = < 0.01 result is very significant
p = < 0.001 result is highly significant
Credibility and bias in correlational research
- bias in correlational research can occur on level of variable measurement and interpretation of findings
- less bias there is in a correlational research study, the more credible it is
- credibility in correlation research is the same idea as internal validity in experimental research
- term ‘internal validity’ is NOT used in correlational research
Bias on the level of variable measurement
- depending on method used to measure variables, bias may be inherent in measurement procedure
- ## bias isn’t specific to correlational research; will occur in any other research study using same variable measured in the same way
Sources of bias in interpretation of findings
- curvilinear relationships between variables
- the third variable problem
- spurious correlations
Curvilinear relationships between variables
- in calculating correlation between 2 variables, we assume that the relationship is linear
- formula of a correlation coefficient = formula of a straight line
- but, curvilinear relationships can’t be captured in a standard correlation coefficient
Counteracted: if suspected, curvilinear relationships should be investigated graphically
The third variable problem
- there’s always a possibility that a 3rd variable exists that correlates w/ both A and B and explains their correlation
- if you only measure A and B, you’ll observe a correlation between them; doesn’t mean they’re related directly
Counteracted: consider potential ‘3rd variables’ in advance, include them in research study to explicitly investigate links between A and B, and these ‘3rd variables’
Spurious correlations
correlations obtained by chance
- becomes an issue if research study includes multiple variables and computes multiple correlations between them
- if 100 correlations are measured, there’s a chance that a small no. will be significant, even if in reality the variables aren’t related
Counteracted:
- results of multiple correlations should be interpreted w/ caution
- effect sizes need to be considered together w/ level of statistical significance
Sampling and generalisability in correlational studies
Sampling strategies: random, stratified, convenience and self-selected
Generalisability:
- depends on how representative sample is of the target population; representativeness of sample depends on sampling strategy
- random and stratified samples are more representative than opportunity and self-selected samples; similar to idea of population validity in experimental research
- construct validity is important in considering generalisability
Credibility in qualitative research
- ‘trustworthiness’
- a measure of the extent to which the experiment test what it’s intended to test
Measures to increase credibility in qualitative research
- Triangulation
- Establishing a rapport
- Iterative questioning
- Reflexivity
- Credibility checks
- Thick descriptions (rich descriptions)
Triangulation
Combining and comparing different approaches to collecting and interpreting data
Four types:
- method triangulation: combining different methods
- data triangulation: using data from a variety of accessible sources
- researcher triangulation: combing and comparing observations of different researchers
- theory triangulation: using multiple perspectives or theories to interpret the data
Establishing a rapport
- building a relationship of trust w/ participant; emphasise necessity to be honest; right to withdraw and that there are no good or bad answers
- the above prevents participants from altering their behaviour in the presence of the researcher
Iterative questioning
- returning to the topic later in the process of interaction w/ participant and rephrasing the question
- allows a deeper investigation of sensitive topics
Reflexivity
- taking into account possibility that the researcher’s own biases may be affecting the results of the study
- doesn’t necessarily allow researchers to avoid bias, but allows them to identify findings possibly affected by bias
Two types:
- Epistemological reflexivity: taking into account strengths and limitations of the methods used to collect data
- Personal reflexivity: taking into account personal beliefs and expectations of the researcher that might have resulted in bias
Credibility checks
- checking accuracy of data by asking participants themselves to read transcripts of interview or field notes of observation
- get them to confirm that the notes/ transcript reflect correctly what the participant said or did
Thick descriptions
- describing the observed behaviour in sufficient detail so it can be understood holistically and in context
- contextual details should be sufficient to make description meaningful to an outsider who never observed this behaviour first-hand
Bias in qualitative research
- sources of bias can be associated both w/ researcher and participant
- there are 2 groups of biases: participant bias and researcher bias
Acquiescence bias
- tendency to give +ve answers whatever the questions
- may occur due to participant’s natural agreeableness or because they feel uncomfortable disagreeing w/ RQ
To overcome:
- be careful not to ask leading questions
- questions should be open-ended and neutral
- should be clear that there are no right or wrong answers
Social desirability bias
- participant’s tendency to respond in a way they believe will make them more liked/accepted
- intentionally/ unintentionally, participants may produce a certain impression instead of natural behaviour; especially true for sensitive topics
To overcome:
- questions are phrased in a non-judgemental way
- good rapport should be established
- questions can be asked about a 3rd person
Sensitivity bias
tendency of participants to answer regular questions honestly but distort their responses to questions on sensitive topics
To overcome:
- build good rapport
- create trust
- reinforce ethical considerations eg. confidentiality
- increase sensitivity of questions gradually
Types of participant bias in qualitative research
- acquiescence bias
- social desirability bias
- dominant respondent bias
- sensitivity bias
- confirmation bias
- leading questions bias
- question order bias
- biased reporting
Dominant respondent bias
- occurs in a. group interview setting when one participant influences behaviour of others
- other participants may be intimidated or feel like they will be compared to dominant respondent
To overcome:
- keep dominant respondent in check
- try to provide everyone w/ equal opportunity to speak
Confirmation bias
- occurs when researcher has a prior belief and uses research to confirm this belief (intentional/ unintentional)
- may manifest as selectivity of attention or tiny differences in non-verbal behaviour that may influence participants’ behaviour
To overcome:
- unavoidable as in qualitative research, human observer is an integral part of the process
- bias is recognised and taken into account through process of reflexivity
Leading question bias
occurs when questions in an interview are worded in a way that encourages a certain answer
To overcome:
- researchers are trained in asking open-ended, neutral questions
NB/ issue in an interview not in an observation
Question order bias
occurs when response to one question influences particpants responses to subsequent questions
To overcome:
- bias can’t be avoided but can be minimised by asking general questions before specific ones
- asking positive questions before negative ones
- aksing behaviour-related questions before attitude-related ones
Biased reporting
occurs when some findings of the study aren’t equally reported in the research report
To overcome:
- reflexivity
- independent researchers may be asked to review the results (researcher triangulation)
Types of generalisability in qualitative research
- Sample-to-population generalisation
- Theoretical generalisation
- Case-to-case generalisation (transferability)
Sample-to-population generalisability
- applying results of the study to a wider population
- depends on how representative the sample is
- best way to ensure representativeness of sample is to sample randomly
- nature of sampling in qualitative research is non-probabilistic, this generalisation is a weak point
Equivalent in quantitative research:
- population validity
Theoretical generalisation
- generalising results of particular observations to a broader theory
- theory plays a much greater role in qualitative research
- can generalise to a broader theory if data saturation has been achieved
Equivalent in quantitative research:
- similar idea to construct validity
- as it refers to “leap” from observable operationalisations to unobservable construct
Case-to-case generalisation (transferability)
- applying findings fo a study to a different group of people or a different setting or context
- this is the responsibility of the researcher and the reader of the research report
- researcher ensures that thick descriptions are provided so reader has sufficient information
- reader decides whether new context is similar enough to one described in report for findings to be applicable
Equivalent in quantitative research:
- ecological validity (generalising from experimental settings to real-life settings)
Data saturation
a point when further data doesn’t add anything new to the already formulated conclusions and interpretations
Sampling in qualitative research
- is non-probabilistic
- doesn’t aim to ensure representativeness in relation to a. target population
- aims to ensure that participants recruited for the study have the characteristics that are of interest to RQ
Types of sampling in qualitative research
- quota sampling
- purposive sampling
- theoretical sampling
- snowball sampling
- convenience sampling
Quota sampling
- decided prior to research how many people to include in the sample and which characteristics they should have
- decision is driven by RQ
- various recruitment strategies are then sued to meet the quota; isn’t important how many people are sampled
- it’s important that people in sample have characteristics of interest to researcher
- approach is completely theory-driven; all characteristics of the sample are defined in advance based on RQ
Purposive sampling
- recruit participants that are of interest to researcher
- sample size and proportions of participants within sample aren’t defined in advance
- target characteristics of participants are defined in advance, but composition of sample is not
Theoretical sampling
- a sampling method that stops when data saturation is reached
- whether information is “new” or not is defined on the basis of the background theory
Snowball sampling
- small no. of participants is invited and asked to invite people they know also w/ these characteristics of interest
- can be used in combination w/ other sampling strategies
- convenient w/ groups of people who are difficult to reach
Convenience sampling
- using the sample that is readily available
- most cost-efficient method, but also most superficial
Observation
Reasons to choose observation:
- focus of research is on how people interact in a natural setting; most other methods would require placing participant in an artificially created environment
- meaningful knowledge in a research area can’t be easily articulated; so, observing behaviour is preferable to asking participants for their interpretations
- observation allows researcher to gain first-hand experiences w/ phenomenon under study
Limitation:
- researcher is strongly involved in generation of data through selective attention and interpretation
- but, this is the case w/ most qualitative research methods- why reflexivity is especially important
Naturalistic observations
Carried out in real-life settings that haven’t been arranged for the purposes of the study
Pros:
- sometimes is the only option eg. when it’s unethical to encourage a particular behaviour in a lab eg. violence
- participants behaviour isn’t influenced by artificiality of research procedure
Cons:
- may be time-consuming because behaviour of interest only occurs at certain times
Laboratory observations
Carried out in specially designed environments
- participants are invited to lab and most of the time know they’re participating in psychological research
Pros:
- it’s possible to recreate situations that don’t frequently emerge in real life
- it’s possible to isolate behaviour of interest more efficiently
Cons:
- artificiality of procedure may influence behaviour of participants
Interview
Reasons to choose interview:
- may be the only way to get an insight into participant’ subjective experience and interpretations; these phenomena are unobservable, only option is to rely on verbal reports
- can be used to understand participants’ opinions, attitudes and meanings they attach to certain events
- only way to understand how participants respond to past events is through self-report; can’t recreate those experiences
- in-depth individual interviews are useful when topic is too sensitive for people to discuss in a group setting
Interview data: comes in form of an audio or video recording that’s subsequently converted to an interview transcript
Covert observations
Occurs when researcher doesn’t inform members of the group about the reasons for their presence
Pros:
- participants don’t suspect they’re being observed, so they behave naturally
Cons:
- often participants don’t consent to being observed, which raises ethical issues
Structured interview
Includes a fixed list of questions that need to be asked in a fixed order
Pros:
- useful when research project involves several interviews
- it’s essential to ensure that they all conduct interview in a standardised way
Cons:
- some participants may have unique circumstances/opinions that can’t be accommodated in a structured interview
Semi-structured
Don’t specify an order or a particular list of questions
- interview guide is like a checklist; interviewer knows there are some questions that must be asked, but there’s flexibility to ask additional follow-up questions
Pros:
- fits natural flow of conversation better
- better suited for smaller research projects
- more effective in studying the unique experiences of each participant
Cons:
- less comparability across researchers and participants
Unstructured
- is participant-driven
- eery next question is determined by interviewee’s answer to the previous one
Pros:
- very effective for investigating unique cases or cases where no theoretical expectations exist that would inform wording of the questions
Cons:
- most “qualitative” of all 3 types
- more time-consuming
- results are more difficult to analyse and interpret
Focus group
a special type of semi-structured interview that is conducted simultaneously w/ a small group of people (6-10)
- encourages participants to interact w/ each other; creates group dynamics that are observed and analysed by the researcher
- interviewer acts as facilitator who keeps interaction focused on research question
Reasons to choose focus group:
- participants interact w/ each other rather than the researcher; makes behaviour more natural
- interaction between participants may reveal more aspects than would be revealed in a one-on-one conversation w/ researcher
- is easier to respond to sensitive questions when in a group
- multiple perspectives are discussed, allows researcher to obtain a more holistic understanding of the topic
Limitations of focus groups:
- dominant respondents can disrupt group dynamics; their assertiveness may affect and distort behaviour of other participants
- it’s more difficult to preserve confidentiality in a group
- especially demanding in terms of sampling and creating interview transcripts
Case study
an in-depth investigation of an individual or a group
- involve a variety of methods eg. interviews, observations, questionnaires
- deepen understanding of an individual or a group of interest
Reasons to choose a cases study:
- useful to investigate phenomena that can’t be studied otherwise
- can contradict established theories; in this way urge scientists to develop new ones
Limitations of case studies:
- researcher bias and participant base are problems; researcher interacts w/ participant for prolonged periods of time- may compromise impartiality and influence how natural participants’ behaviour is
- generalisation of findings from a single case to other setting or a wider population is particularly problematic
- is difficult to protect confidentiality of participants and their data
Alternative/addition for structured interview
Alternative/ addition: semi-structured interview
Reason:
- fixed questions in the fixed order might force some participants to respond unnaturally
- so, some aspects of their experiences were missed
Alternative/addition for a correlational study
Alternative/ addition: experiment
Reason:
- correlational studies don’t show causation, but causation in the context of a study is important
- if one the variables can be manipulated, then the study can be conducted as an experiment
Alternative/addition for an overt non- participant observation
Alternative/ addition: covert participant observation
Reason:
- the fact that participants knew they were being observed altered their behaviour, leading to participant bias
Alternative/addition for an experiment: repeated measures design
Alternative/ addition: experiment, independent measures design
Reason:
- in repeated measures, participants take part in more than one condition
- increases chance they’ll figure out true aim of the study, which could lead to demand characteristics
Unstructured observation
Observers simply register whatever behaviour they find noteworthy- there is no checklist that was prepared in advance
Pros:
- more flexible; researcher isn’t limited by prior theoretical expectations
Cons:
- less structured, means less comparable across researchers and across participants