Research Methods
Terms
undefined, object
copy deck
- Emergent Properties
- Information not contained within elements, but only becomes apparent when we consider multiple elements and how they interact with one another.
- Hypothetico-deductive approach
- Observation -> Theory -> Hypotheses -> Research -> Research Observations
- Paradigm
- generally accepted framework of ideas that required many challenging observations before it is rejected
- Inductive research
- Observations -> Theoretical inference -> Observations...
- Positivism
- Only what can be directly observed and measured is counts as knowledge.
- Positivism also known as...
- Empiricism
- Constructivism
- Implies there are knowledges, rather than knowledge Can describe same phenomenon differently
- Anti-positivism
- Emphasises importance of differentiating between human and natural sciences. Interpretative treatment of social and cultural events Phenomological approach - look at event through eyes of people involved
- Anti-positivism also known as...
- Interpretivism or social action approach
- Nomothetic research
- Concerned with identifying general laws of human behaviour, which allows behaviour to be predicted Use stats that average out human variation. Look for general differences. Have to be careful with samples
- Idiographic research
- Concerned with exploring uniqueness. Few cases examined in more depth. By gaining greater understanding of individuals, will gain better understanding of whole
- Hermenuetic research
- Concerned with meaning in everyday life. Investigate how people interpret their experience and symbolism and meaning in everyday life
- Sampling
- Process of collecting a set of research participants to provide data for an experiment
- A sample is...
- A small subset of a population obtained for testing purposes
- A sample is representative if...
- It is typical of the population as a whole
- Which type of research allows findings to be generalised to the population?
- Nomothetic
- 2 things that must be ensured in order for the sample to be representative:
- Size of the sample Specific sampling techniques
- Specific sampling techniques to ensure a sample is representative:
- Quota sampling Stratified sampling Opportunity sampling Snowball sample Cluster sample
- Quota sampling...
- Population sorted into categories, proportional number taken from each category, Ensures representative amount of each category present.
- Stratified sampling...
- Population divided into layers, with same number taken from each layer
- Opportunity sampling...
- Researcher uses whatever participants are available at the time the study is conducted.
- Snowball sampling...
- Selected for sample by asking key figures who they think will be useful or important to include
- Cluster sampling...
- Selected from specific geographical area as being representative of the population
- Most desirable sampling technique and why.
- Random - Most likely to contain all of the characteristics of the population
- Worst type of sampling and why.
- Opportunity - Most open to bias and distortion.
- Define validity
- Determines whether or not something measures what it is supposed to measure.
- External validity and threats
- How well findings generalise to other populations or situations. Threats: Other participants, Other times, Other settings
- Construct validity and threats
- The extent to which results support the theory behind the research Threats: Ambiguous effect of manipulation, Effects of repeat testing, Regression towards the mean, Selection bias, Mortality, Experimenter bias
- Statistical validity
- Is there a true cause and effect relationship, or are effects due to chance?
- Hermenuetic Research: Levels of meaning
- Levels of meaning: conscious, unconscious, personal, social and socio-political
- Volunteer effects
- This is when participants are more co-operative than normal because they are participating in psychological research.
- Demand characteristics
- This is when participants pick up on subtle cues which cause them to behave differently to how they would normally.
- Demand characteristic cues...
- Design of the study, the setting and the age, gender, ethnicity and size of the researcher.
- Why do self-fulfilling prophecies occur?
- Participants pick up on small non-verbal cues which let them know what the expermineter expects and allow them to respond unconsciously to these.
- How to prevent self-fulfilling prophecies...
- Standardised instructions - to ensure that all participants are given the same instructions Standardised procedures - to make sure all participants are treated the same Standardised conditions - to ensure that all groups have the same experience as far as possible Double-blind testing - Ensure that the researcher running the testing is not aware of the expected outcome of the tests, so that they cannot cue the participants.
- How to eliminate volunteer effects
- Single-blind technique (where researcher knows the predicted outcome, but participants don't)
- Control condition
- A condition where we aim for participants to behave 'normally' to give us an idea of 'baseline' behaviour.
- Randomisation
- Randomly assigning participants to experimental groups in order to minimise the effect of any individual differences.
- Repeated-measures designs
- This is when the same experimental participants are used in all the different experimental conditions, thus minimising the effect of any individual differences.
- Repeated-measures design also known as...
- Related-measures design or correlated-subjects design.
- Order effects
- This is when the order of the conditions in an experiment affects the results obtained for each. To counteract this, we must counterbalance the experimental conditions, so that each condition follows every other one equally often.
- Practice effects
- This is when participants get better at a task, thus affecting experimental results. The solution to this is to have a second group of participants experience the trials the opposite way round.
- Independent-measures design
- This involves different people doing different conditions of the study. This can cause problems if there are systematic differences between the two groups.
- Matched-participant designs
- This is an experimental design that involves having different people in the different conditions. This can cause problems if there are systematic differences between the two groups of of people, but there are ways of minimising this. In practice matching participants is difficult, expensive and time consuming.
- Comparison group
- One of the control groups used when researchers are looking at multiple aspects of a problem.
- Operational definition
- A way of defining something that isn't quite perfect, but will serve as a workable definition for, for example, a research project.
- The 4 types of variables:
- Independent variable Dependent variable Confounding variable Contaminating variable
- Independent variable
- This is the variable that is controlled by the experimenter
- Dependent variable
- the variable that is measured to discover the effect of changing the independent variable.
- Confounding variable
- a variable which affects one condition of the independent variable (i.e. experimental or control group) but not the other, or not to the same extent.
- Contaminating variable
- – this is a variable that randomly effects the experimental results, making them less reliable.
- 5 types of experiments
- Laboratory Quasi-experiments Field experiments Natural experiments Single-case experiments
- Laboratory experiments
- The experimenter manipulates variables under controlled conditions, in the psychological laboratory, using randomised allocation of the participants to the experimental conditions
- Quasi-experiments
- The experimenter manipulates variables under controlled conditions, in the psychological laboratory, but randomised allocation of participants to conditions is not possible.
- Field experiments
- The experimenter manipulates variables in a natural setting, and is therefore not able to control all possible contaminating or confounding variables.
- Natural experiments
- The experimenter does not manipulate the variables, but studies a situation in which a single variable has changed through natural or socio-political causes.
- Single-case experiments
- The experimenter manipulates variables under controlled conditions, but draws the data from a single case rather than a group of participants.
- Methodology
- A theoretical analysis defining a problem and detailing how research should proceed
- Method
- The technique/strategy actually adopted in the research.
- Reliability
- Consistency and stability of measurements and findings
- Statistical tests to assess reliability of an instrument:
- Test-retest - Comparison scores on two occassions Split-half reliability - Halfof items compared to other half Alternative-forms - Two versions of test compared
- Reliability of coding in analyses:
- Inter-rater - comparison of two or more coders Itra-rater - Consistency of a single coder
- Reliability between observers:
- Inter-observer reliability
- Non-experimental research:
- Research in which the experimenter does not exert any manipulation over the conditions of the study.
- Four types of non-experimental research:
- Observational - Record behaviour 'as is' Archival - Study existing data Case study Survey/Interview
- Uses of non-experimental research:
- Can describe behaviour Can predict behaviour Can explain behaviour (but not to the extent of establishing cause and effect due to lack of control)
- Unstructured observation research:
- Main uses: for pilot studies, gor grounded theory epistemology Start with limited preconceptions - record whatever is relevant/of interest Covers a wide range of behaviour within a setting Observer often a participant
- Structured observation research:
- Greater focus - specific behaviours observed Measurement - more likely to be categorical than verbatim reports Observer unlikely to be a participant - more likely to be covert
- Sampling by time:
- Set periods Allows intense focus of concentration and breaks May miss infrequent behaviour and sequences of behaviour
- Sampling by event:
- Set periods, but defined by event. Picks up infrequent behaviour Less loss of sequences Can be demanding
- Archival research:
- Purpose - attempting to answer research questions from archival data Archival data - factual information in existing records or archives. Limitations of Archival research: *Data collected for some other purpose Ruling out alternative hypotheses may be diffcult No real possibility of informed consent.
- Case studies:
- Purpose - observation of an ongoing situation, usually involving a single individual. Characteristed by the uniqueness of the individual. Often use more than one research method. Often provides theoretical insights which stimulate other research.
- Surveys:
- Purpose - used to identify how people feel, think and act. Generally: *Use a fixed, quantitative design *Collect a small amount of data in a standardised form from a relatively large number of respondents. Use a selection of representative samples of individuals from known population. Are used to discern patters of assocation and differences Can be used for exploratory, descriptive or explanatory purposes.
- Big Q
- Open-ended, inductive research methodologies that are concerned with theory generation and the exploration of meanings
- Small q
- Incorporation of non-numerical data collection techniques into hypothetico-deductive research designs Start with hypothesis and research defined categories and qualitative data is checked against these.
- Three strands of qualitative methods
- Reliability & Validity Generativity & Grounding Discursive & Reflexive
- Design requirements for qualitative methodologies:
- Data collection needs to be participant-led. Design needs to be 'open-ended' and flexible Design needs to accomodate general principles with respect to: *Type of data collected *Role of the participants
- 3 types of reflexivity:
- Personal Epistemological Critical language awareness
- Personal reflexivity
- Reflect how the research was influenced by the researcher (collection and analyses) Their own values, experiences, interests, beliefs, political commitments, wider aims in life, social identities
- Epistemological reflexivity
- Reflect how research was influenced by assumptions made about the knowledge How the research question limited what could be 'found', how design and analyses are 'constructed' the data and findings, could it be investigated another way - if so, how would findings differ?
- Critical language awareness
- Also need to consider the power of words - meanings vary between people Researcher's labels and categories in analysis attach meaning Questions attach meanings e.g. "how did you feel?" evokes emotional response.
- Statistics
- A set of mathematical procedures for organising, summarizing and interpreting information.
- Parameter
- A value that describes a population. A parameter may be obtained from a single measurement, or it may be derived from a set of measurements from the population.
- Descriptive statistics
- Statistical procedures used to summarize, organise and simplify data
- Inferential statistics
- Techniques that allow us to study samples, and make generalisations about the populations from which they were selected.
- Nominal scale
- This consists of a set of categories with different names but are not related to one another in any systematic way.
- Ordinal scale
- This consists of a set of categories that are organized in an ordered sequence. Measurements on an ordinal scale rank observations in terms of size or magnitude.
- Interval scale
- Consists of ordered categories that are all intervals of exactly the same size. Equal differences between numbers on the scale reflect equal differences in magnitude.
- Ratio scale
- This is the same as an interval scale with the addition on an absolute zero point. Ratios of numbers do not reflect ratios of magnitude.
- BIDMAS?
- Brackets Indices Division Multiplication Addition Subtraction
- A positively skewed distribution
- Leans to the left
- A negatively skewed distribution
- Leans to the right
- Type I error
- This is when a researcher rejects a null hypothesis that is actually true. This means that they conclude that the manipulation has an effect on the dependent variable when in fact it doesn't.
- Type II error
- This is when a researcher fails to reject a null hypothesis that is false. A type II error means that the hypothesis test has failed to detect the effect of the manipulation.
- Factors affecting statistical power
- Alpha level - Reducing alpha level will reduce power, makes it harder to reject null hypothesis Using 2 tailed tests - harder to reject null hypothesis than with 1-tailed. Sample size - Smaller sample size will reduce power
- Participant observation
- visible presence, observer part of group observed
- Complete participant
- concealment of researcher role
- Participant-as-observer
- research role made clear from the start (dual role)
- Marginal participation
- lower participation, largely passive, but prescence can influence behaviour.
- Observer-as-participant
- no part in activity, researcher role known
- Covert observation
- complete observer