Research Methods: Revision Sheet (December 09)
Quantitative & Qualitative data
Qualitative – non-numerical data information that cannot be counted, e.g. about how you think or feel. Usually focuses on a participant’s thoughts and feelings.
• Represents the true complexities of human behaviour, gains access to thoughts and feelings.
• Tends to give richly detailed data.
• However more difficult to analyse.
• They can be summarised, and they can be converted into quantitative by putting in categories and then counted.
• What type of research method will give qualitative data?
• Self Report – open questions (questionnaires & interviews)
• Observations (without structure, pure description)
• Case Studies
Quantitative – involves measuring behaviour in a numerical way. Quantitative refers to QUANtities. Behaviour is measured in numbers or quantities. For example about your age, how many hours you work a week.
• Easier to analyse.
• However reduces information about people to over-simplified statistics (“statistically significant but humanly insignificant”).
• What type of research method will give qualitative data?
• Experiments,
• Correlations,
• Self Report methods (closed questions)
• Observations (structured, checklists
• Content Analysis
Aims & Hypothesis
• Aim: The stated intentions of a study. So you are clear about the purpose of your study.
• Hypothesis: A clear unambiguous statement of what you believe.
• This is also known as the research or alternative hypothesis.
• If you conduct an experiment – you call it an experimental hypothesis.
• The null hypothesis predicts that the expected effect will not occur.
Sampling
– opportunity, random, stratified and systematic.
• The sample must be as representative as possible of the population to be studied, because the results can then be generalised to that population.
• In real life research, it is impossible to get a truly representative and nonbiased sample. In fact a considerable portion of the research on adults has used only participants who were psychology students!
• All samples aim to produce a representative sample but are inevitably biased.
Random sample
• Random means that every possible participant (known as the target population) has an equal chance of being selected.
• Participants are chosen mathematically using chance.
• Every person in the target population has an equal chance of being selected. e.g. names from a hat, computer generated list.
• Often, it is not possible for researchers to use random sampling. Studies using undergraduates, or people who answered an advertisement, or people sitting on a beach are not using random sampling. However, where it is possible, random sampling should be used.
Strengths
• Avoids bias as researcher has no control over who is picked.
• Chance is that it will be representative.
Limitations
• Can get unrepresentative (“freak”) samples.
• Difficult to achieve and time consuming
Opportunity sampling.
• You ask anyone who fits your requirements.
• It consists of people who are available to take part and are accessible to the researcher.
• Opportunity sampling is also used where the researcher advertises for participants.
Strength
• Less time consuming – easy to do.
Limitation
• The sample is likely to be biased. It is not a good way to obtain your sample because there may be a bias; for example you might only ask people to take part who look approachable and cooperative.
Stratified
• Identify the different sub groups (strata) of people in the target population.
• Draw participants from each of these subgroups.
• Participants selected at random from each sub group.
Strengths
• Increases chance of getting a representative sample.
Limitations
• Time consuming to select all participants – as all participants need to be assessed and categorised.
• Can be difficult to identify all the key elements in the strata
Systematic
• Every nth member of the target population is chosen, e.g. every 5th house on a street.
• Similar to random – but don’t confuse them.
Strengths
• Avoids bias as the researcher does not actually choose who to use.
Limitations
• By chance could get a “freak” sample that is not representative.
Experimental design
• The process by which a researcher decides whether the different conditions in an experiment are to be experienced by different groups of participants or the same participants.
• Do not confuse experimental design with research method.
• Independent groups
– Different people take part in each condition
• Repeated measures
– The same people are used in both conditions of the experiment
• Matched pairs
– Different people take part in each condition but they are matched in ways that matter for the experiment.
Independent groups
• Advantages
– Participants do only one condition – so no order effects (when performance in one condition can be affected by performance in the other conditions- such as practice or fatigue)
– Less likely to guess aim of experiment so reduces demand characteristics
– Less materials needed.
• Disadvantages
– Participant variables
– More participants needed
Repeated measures
• Advantages
– Participant variables are eliminated
– Fewer Ps needed
• Disadvantages
– Order effects
– Ps may figure out the aim
– May need more materials
• Counterbalancing can be used to control for order effects. Half the ps do condition A followed by condition B, the other half start with condition B then do condition A. This means the effects of order are balanced out.
Matched Pairs
• Advantage
– Participant variables minimised
– No order effects (ps only do one condition)
– Only one set of materials needed
• Disadvantage
– Matching is difficult and time consuming
– Participant variables are not eliminated as 2 people are never entirely alike.
Variables
• All research revolves around variables—things that vary
– e.g. height, anxiety, time required to recognize a word on a computer screen as a real word, helping or not, refusing a beer or not
Independent and Dependent Variables
• Note: only occur in experiments.
• IV: what the researcher changes
• DV: what the researcher measures
Extraneous variables
• Extra factors that might affect what you are trying to measure.
• These should be controlled in an experiment.
Confounding variables
• Extraneous variables that have actually had an effect.
Pilot Study
• Small scale study conducted before an investigation to check for any problems in the design.
• Also allows the researcher to become practiced with the techniques of the research.
• Questionnaires: Check the clarity of the questions. It determines whether the questions asked are clear and unambiguous and whether members from the target population can complete the questionnaire survey in the given time. Also indicates whether the questionnaire will provide the type of data required. It allows changes / revisions to be made before the real questionnaire is administered. Check the layout of the questionnaire.
• Observation: Allows researchers to see if the classification of behaviours is clear and useable. It would allow them to practise observing the actual behaviours before the main study.
Research Method / Description / Advantages / DisadvantagesExperiments / • All experiments have an independent variable (IV), and a dependent variable (DV)
– IV: what the researcher changes
– DV: what the researcher measures
Experiment, isolates and manipulates the independent variable to observe its effect on the dependent variable, controls the environment in order that extraneous variables may be eliminated, predicts cause and effect. / · Enables inference about cause and effect.
· Can control many extraneous variables so that reliability is high and alternative explanations of events is eliminated. / Can produce quite artificial conditions and measures of variables.
Laboratory Experiment / • Does not necessarily take place in a laboratory.
• Takes place in a carefully controlled environment, e.g. a quiet room, where factors such as noise levels, temperature, light levels can be controlled, so are the same for all participants (ps) / • A cause and effect relationship can be demonstrated between the IV and DV.
• The environment can be controlled.
• Replication is possible because all the variables have been identified.
• Ps know they are taking part so less ethical issues. / • Artificial situation - low ecological validity
• Demand characteristics – participants act in a way they would not normally because they know they are taking part in an experiment.
Field Experiment / • Conducted in a more natural setting (i.e. “In the field”)
• Participants often not aware they are taking part. / • Greater ecological validity.
• Avoids participant effects as they are unaware they are in an experiment. / • Harder to control all the variables.
• More time consuming.
Ethical issues
• It is not possible to get informed consent from the participants
Natural Experiment / • Also known as a “quasi experiment” (quasi means not quite the real thing)
• The IV is not directly manipulated.
• e.g. seeing how a volcano eruption affects stress levels. (what is the IV and DV?) The experimenter could not change the IV. / • It is the only way to study cause and effect in certain situations.
• Can study “real” problems.
• / • Not easy to replicate
• Many confounding variables
Content analysis / • Used to quantify the content of any form of communication, usually media.
• Detailed and systematic examination of material to identify patterns, themes, or biases.
• Used on books, magazines, TV programs, transcripts of conversations, e.t.c.
• Units analysed can be individual words, ideas, characters, images, or amount of space/time dedicated to each issue.
• Can also analyse what is not there.
• Usually record the frequency each unit occurs. (Similar to another research method: observation).
– a word-frequency count. (The assumption made is that the words that are mentioned most often are the words that reflect the greatest concerns – however issue with use of synonyms.)
– e.g. number of times Obama uses the word “change” in his speech
• Might also use a rating system
– e.g. Rate confidence level of female characters in Bond films on a scale of 1 à 10 / • Allows analysis of a wide range of materials
• Few ethical problems as not dealing directly with people – it is unobtrusive
• Can usefully summarise key features of a large body of data,
• Can highlight similarities and differences across the data set.
• Can generate unanticipated insights.
• Relatively easy and quick method to learn, and do. / • Most disadvantages are concerned with not conducting it properly
• e.g. Researcher bias:
– Interpretation of data is subjective
– Categories decided by researcher
• often disregards the context that produced the text, (e.g. doing a word count does not inform us about the context in which the word was used)
• / •
Observation / Observation requires carefully watching and recording of the behaviour that is of interest.
· The observational method is a non-experimental design.
· Key feature is a standardised, planned, and systematic approach to objectively observe and record behaviour.
· Aims to generate data upon which to base any conclusions.
· The absence of an independent variable does not allow any cause-effect conclusions to be drawn from observational research.
· Observers may make written records of behaviour as it happens
· May be video taped first and analysed later.
· Sometimes, observers note all behaviour in a fixed period of time.
· Sometimes they may sample particular types of behaviour.
· Sometimes they note behaviour which occurs at particular times.
· Construct an observation schedule.
· List the behaviours that are of interest.
· Note each time they occur, as well as how long they last, who shows the behaviour, who else is involved and so on.
· Due to complexity of behaviour, and the possibility of bias in the observer, it is useful to have more than one observer.
· Each is trained so that they all interpret the behaviour in the same way.
· This is called interobserver reliability.
· Practice observing the type of behaviour interested in, / • They provide a lot of detailed information.
• They enable psychologists to study people when other methods might be unethical.
• Less artificial than experimental method
• Provide hypotheses for future research. / • The observer effect. If participants are aware they are being observed they may not behave naturally / change their behaviour. (This is a confounding variable in observational research.)
• Observations can be expensive, because they require several specially trained people and.
• Analysis can take a long time.
• Cannot imply causality.
• The absence of an independent variable does not allow any cause-effect conclusions to be drawn from observational research.
• It is difficult to record all behaviour, so something important may be lost.
• Observer bias. The psychologist may look for a particular behaviour which fits with his expectations therefore reducing validity. It is difficult for observers to be completely objective. Can occur in that the observer, and thus their results, may be influenced by prior knowledge, or experience of the situation / subject / participants under investigation.
Ethical issues
• Gaining consent is difficult.
• Should only observe people in situations where they would expect to be observed.
Naturalistic Observation / • Naturalistic observation is the planned watching and recording of behaviours as they occur within a natural environment. An example would be naturalistic observation of animals in their natural habitat. / • High ecological validity.
• Enable researchers to see people's natural behaviour in an everyday setting.
· Particularly good for observing specific subjects.
• Spontaneous behaviours are more likely to happen. / • Confounding variables: Researchers are not able to control the variables in a natural setting and something that the observer is not aware of may affect the participant's behaviour.
• Ethics: Where research is undisclosed consent will not be obtained, where consent is not obtained - details may be used which infringe confidentiality.