Research Methods Flashcards
Types of Observation
Naturalistic- Watching, recording behaviour normal setting would usually occur in this setting
Controlled- Watching, recording structured environment some variables are controlled
Covert- Behaviour watched, recorded without knowledge/ consent
Overt- Watched, recorded with participants knowledge / consent
Participant- Researcher becomes member of group watch, record
Non-Participant- Remains outside group watch record
Variables
Independent Variable (IV)- Manipulated by researcher
Dependent Variable (DV)- Should be caused by IV, measured
Operationalisation- Makes IV, DV usable, testable
IV- condition 1 condition 2 DV- Amount of… Number of…
Extraneous Variables- Potentially effect DV if not controlled, for example Ppt is in a chilly room
Confounding Variable- Variable affecting DV, unsure what has caused changes to DV, for example Drowning caused by eating Ice Cream Temperature in Summer Confounding
Pilot Studies
Definition- Small scale version of study that takes place before the real study, fewer participants, checks for problems that are corrected for real thing
Aims- Questions / Interviews checked for issues so that they can be rewritten if needed. Observations check behavioural categories. Experiments design, procedure, instructions, materials checked
Real study should be a modified version which should save time and money in the long run
Types of Sampling
Population is the large group of individuals that a researchers is interested in studying
A sample is a smaller group that represents the target population that the researcher wants to study
Sampling techniques aims to produce a representative sample that is less prone to bias making results more generalisable to the population that the sample represents
Random Sampling- Assign each person a number on a piece of paper, draw out of a hat
Systematic Sampling- Every nth member of population selected, get list of people pick every nth person
Opportunity Sampling- Select people available at the time, students in canteen
Volunteer Sampling- Self-selected participants, notified through adverts, newspaper
Stratified Sampling- Sample reflects proportion of the people in population. Identify group, workout proportion, select participants needed
Sampling Advantages and Disadvantages
Random Sampling- Unbiased but difficult / time consuming. Requires a complete list of population and cannot be repeated equal chances
Systematic Sampling- No influence over who is chosen but time consuming and requires a list, person may also refuse to take part
Opportunity Sampling- Less costly but unrepresentative (only one area), cannot be generalised, Researcher Bias, too much control
Volunteer Sampling- Easy, no researcher input, volunteer bias, demand characteristics
Stratified Sampling- Reflects population, generalised, can never be a perfect representation, differences in participants in groups not taken into account
Types of Experiments
Lab Experiment- Controlled Environment, Control of extraneous variables, Researcher manipulates IV, Records effect on DV. For example, Shoe lace experiment
Field Experiment- Natural setting, less control, less control of extraneous variables, Researcher manipulates IV, Records effect on DV. For example, Lift Experiment
Natural Experiment- IV naturally occurring, No control of extraneous variables, Records effect on DV. For example, Hair length
Quasi Experiment- IV not determined by anyone; Variables already exist. For example, Age, Personality, Type
Experimental Designs
Experimental design is the different ways in which ppts can be organised in relation to the experimental conditions
Repeated Measures Design- All participants experience all conditions. Counterbalancing attempts to control for order effects, Half do A->B Half do B->A. For Example, Hazard video with caffeine pill, without caffeine pill vice versa
Independent Groups Design- Participants experience one condition, avoids practice effects, Differences in people may affect results, more people needed, assign participants randomly, ensure similar groups as well as equal chance to be in any of the conditions (reduce participant variables)
Matched Pairs Design- Pairs of participants matched in variables affecting DV (age, gender), One of each pair randomly chosen to do each condition, very difficult to obtain match pairs, time consuming and difficult
Quantitative and Qualitative Data
Quantitative data- Numerical data number of…
Strengths- Can draw graphs, calculate averages
Weakness- Lower external validity
Qualitative data- Expressed in words written description
Strengths- More meaningful, Higher external validity, broader scope
Weakness- Harder to identify patterns, comparison, subjective to interpretation, researcher bias
Both can be obtained from questionnaires, interviews, observational studies etc
Observation Design
Behavioural Categories- Behaviour operationalised into observable, measurable components. For example, affection (emotion) would be operationalised to Kissing, hugging (Observable action)
Sampling behaviour- Event Sampling and Time Sampling used, easier than continuous observation
Event Sampling- Counting how many times a behaviour occurs “Laughed 6 times”
Time sampling- Recording behaviour in particular time frame “5 minutes”
Interviews
Interviews- Can be structured or unstructured
Structured- Made up of pre-determined set of questions. Conducted face to face
Advantages- Straight forward to replicate, reduce interviewer differences
Disadvantages- Limited richness of data, unexpected info received
Unstructured- No set questions, interaction free flowing, expand elaborate answers
Advantages- More flexible and allows for follow up points
Disadvantages- Interviewer bias, interviewee may lie (social desirability bias)
Questionnaires < Interviews- Less Demand Characteristics, More information can be obtained
Questionnaires
Questionnaires- Self-report technique, pre et list of written questions, researcher assesses responses given
Open Questions- Does not have fixed answer, Qualitative data
Closed Questions- Fixed number of responses, Quantitative data
Advantages- Cost effective, large amounts of data quickly, researcher not required to be present, straight forward to analyse
Disadvantages- Responses may not be truthful, Demand Characteristics
Questionnaires > Interviews- Cost effective, no researcher needed
Primary and Secondary Data
Primary Data- Original data collected by researcher through experiment, questionnaire, observation
Strengths- Extract only data needed, Relevant to research aims
Weakness- Takes time, expensive, secondary accessed faster
Secondary Data- Collected by someone other than researcher such as Government statistics
Strengths- Desired info already exists, minimal effort, inexpensive
Weakness- Info may be outdated, incomplete, challenges validity of any conclusions
Peer Review
Definition- Before publication, all aspects of investigation scrutinized by experts in the field. Objective, unknown to researcher
Aims- Allocate research funding, Validation of quality and relevance of research, Improvements and amendments suggested Reviewers unpaid, single-blind
Advantages- Minimises possibility of fraudulent research, published research of highest quality (protected), increases credibility and status of psychology
Disadvantages- Competition for limited research funding, Publication bias (headline grabbing more favourable), Ground breaking research could be buried if goes against reviewers view, anonymity used to criticise rival research
Case Studies
Case Study- In depth investigation of a single individual, group or institution
Strengths- Offers rich detailed insights, gives better idea of unusual behaviour, may generate hypothesis for future study, contribute to understanding of typical functioning
Weaknesses- Generalisation difficult small sample size, Information subjective to researcher, Personal accounts from ppt prone to inaccuracy If childhood story told, lower validity
Correlations
Correlation- Relationship between co variables
+ve- As one increases other increases -ve- As one increases other decreases none- no relationship
Directional – There will be a positive / negative correlation between scores on a happiness questionnaire and scores on a health questionnaire
Non-Directional – There will be a correlation between scores on a happiness questionnaire and scores on a health questionnaire
Hypotheses
- D|ND
- What are the two types of hypotheses?
- Describe how the two differ with an example hypothesis
- Statement made at start of study that describes relationship between variables
- Investigation aim may be the statement below
- Drinking SpeedUpp causes people to become more talkative
- Either directional or non-directional
- Directional includes words like more, less, higher or lower
- People who drink SpeedUpp become more talkative than people who don’t
- People who drink water are less talkative than people who drink SpeedUpp
- Non-directional hypothesis just states there’s a difference, nature not specified (neutral)
- People who drink SpeedUpp differ in terms of talkativeness compared with people who don’t drink SpeedUpp
Independent and dependent variables
- What is an independent variable?
- What is a dependent variable?
- How do we test the effect of an independent variable?
- Independent Variable (IV) is changed/manipulated by the researcher, it effects the DV
- Dependent Variable (DV) is the variable being measured by the researcher (change should be the effect of the IV)
- We test effect of IV using experimental conditions
- The control condition and the experimental condition
- If we were to give ppts some SpeedUpp, how would we know how talkative they were?
- Compare ppts talkativeness before and after drinking SpeedUpp (Experimental condition)
- Compare this to a control group who only had water (Control condition)
Operationalisation of variables
- What does this process aim to do?
- Describe the process
- This is a process of making variables in a hypothesis testable/measurable objectively
- IV can be a concept (subjective), we need to operationalize this so that we can observe this
- The IV is broken down into two conditions (experimental condition, controlled condition)
- The DV is phrased in the following way “Amount off SpeedUpp consumed”, “Number of words said in a given amount of time”
Research issues
- EV|CV|DC|IE
- What are the four research issues
- Describe each with examples
- Extraneous variables (EV), any variable other than the IV that may affect the DV if it is not controlled, “nuisance variables” (Weather, type of room etc)
- EV can be subdivided into Participant variables and Situational variables
- Confounding variables (CV), type of EV, varies systematically with the IV, we cannot tell if any change in the DV is due to the IV or CV (Trauma, Individual differences, Mental state etc)
- Demand Characteristics, any cue from researcher or situation that may be interpreted by ppts as revealing purpose of investigation, may lead to them changing behv within research situation
- Investigator effects, effect of investigator behv (conscious or unconscious) on the DV, may include design of study, selection and interaction with ppts etc
Tackling research issues
- R|S
- What two processes can be used to deal with research issues
- What specific issues do they counter?
- Randomisation, randomise ppts to control the effects of bias when designing materials and deciding order of experimental conditions (names in a hat)
- Standardisation, using same formalised procedures and instructions for all ppts in a research study
Ethical issues
- I|D|P|C|R
- What are the ethical issues that must be considered?
- Informed consent
- Deception
- Protection from Harm
- Confidentiality
- Right to Withdraw
Single-blind and double-blind procedures
- What is a single-blind procedure?
- What is a double-blind procedure?
- What is the difference between the two?
- The ppt is made unaware of the condition or the experiment that they are in in a single blind test, only the researcher knows what condition ppts are apart of
- The ppt as well as the researcher are made unaware of the condition of the experiment that is taking place in a double-blind test
- Researcher usually a third party that conducts investigation without knowing its purpose, this avoids investigator bias
Structured vs unstructured observation
- What is a structured observation
- What is an unstructured observation
- What is the difference between the two?
- Structured observation is when target behvs are simplified and measured using behv categories
- Unstructured observation is when a researcher writes down every detail that they see, tends to produce accounts off behv that are rich in detail
- Appropriate when observations are small in scale and involve few ppts, for example observing interactions between a couple and a therapist in a counselling session
- Structured observations are used when there is too much going on in a single observation for the researcher to record all of it, may be less rich in detail but is too the point (filtered)
Behavioural categories
- What are behavioural categories?
- How do we create behavioural categories?
- What criteria should behavioural categories meet?
- Used in structured observations, DV operationalised into behv categories
- The target behaviour that we want to assess is broken down into a set of behavioural categories
- Target behaviours should be precisely defined (should not overlap) and made observable and measurable
- For example, the target behv “affection” can be broken down into observable behv categories, hugging, kissing, holding hands etc
- Inferences should not need to be made (clearly observable)
- Behv categories form a behavioural checklist (record sheet) to record frequency of observations
Sampling methods
- ES|TS
- What are the two sampling methods used in a structured observation?
- Describe the two and when they would be used
- Continuous recoding of behv is a key feature of unstructured observations, all instances of target behv recorded (not practical when complex behvs are being measured)
- In structured observations, a systematic way off sampling observations must be used
- Event sampling involves counting number of times a particular behv (event) occurs in a target individual or group
- For example, event sampling of dissent in a football match would mean counting number of times players disagree with the referee
- Time sampling involves recording behv within a pre-established time frame
- For example, in a football match we may only be concerned with one player so we use behv categories to see what the individual does every 30 seconds (whatever time frame)
Conducting an experiment
- Describe what must be considered when conducting an experiment
- What are the aims of the study?
- What is the hypothesis, is it directional or non-directional?
- What is the setting? (Controlled or naturalistic)
- What is the observer’s status? (Covert or overt)
- What is the observer’s involvement? (Ppt or non-ppt)
- What sampling method is being used? (Continuous, time sampling or event sampling)
- Is it a structured or unstructured observation?
- Has the DV been fully operationalised into behavioural categories?
- Has a behavioural checklist been created?
- Have ethical issues been considered?
- How will results be presented?
Meta-analysis
- What is a meta-analysis?
- Describe its advantages and disadvantages
- The process of combining the findings from a number of studies with the same research aim
- This produces an overall statistical conclusion (effect size) based on a range of studies
- Larger and more varied sample created, results can be generalised across much larger populations increasing external validity
- Meta-analysis is prone to publication bias (the file drawer problem)
- Cherry pick studies in favour of what researcher believes, neglects studies with negative or non-significant results
- Conclusions made therefore prone to bias
Measures of central tendency
- M|M|M
- What is the measure of central tendency based on?
- What are the three types of central tendency?
- Describe the advantages and disadvantages of the three
- General term for any measure of the average value in a set of data
- Methods of measuring central tendency include the mean, median and mode (descriptive statistics)
- The mean is the most representative due to it including all scores/values in a data set
- However, it is easily distorted by extreme values (anomalies), long to calculate
- The median is unaffected by extreme values (anomalies) and is easy to calculate if data set is in order
- However, it is less representative because it ignores higher and lower numbers, also extreme values may be important
- The mode is very easy to calculate but is not representative of the whole data set
- May not be a useful piece of info in most cases, however it may be the only method that can be used in some cases
- For example, when dealing with categories, the only way to identify the most “typical” or average value would be to select the modal group
Measures of dispersion
- R|SD
- What is the measure of dispersion based on?
- What are the two measures of dispersion we use?
- Describe the advantages and disadvantages of the two
- Based on the spread of scores, how far scores vary and differ from on another
- The range is easy to calculate but only takes into account the two most extreme values which may be unrepresentative of the data set as a whole
- The range is also heavily influence by outliers/anomalies, it also does not give an idea whether most numbers are closely grouped around the mean or spread out
- Standard deviation is a single value that tells us how far scores deviate (move away) from the mean
- Larger standard deviation means a greater dispersion or spread within a set of data
- If we are referring to a particular condition in an experiment, larger SD suggests not all ppts were affected by IV in the same way, data widely spread, suggests there may be anomalous results
- Low SD means data close to mean, implies all ppts responded in a fairly similar way
- SD more precise than range, includes all values within the final calculation, however it can still be distorted by anomalies due to the fact it takes all values into consideration
- Extreme values may also not be revealed unlike with the range
Presentation of Quantitative data
- B|H|S
- What are the three ways that quantitative data can be presented
- Describe each referring to the axes as well as when its appropriate to use them
- Scattergram, represents strength and direction of relationship between co-variables in a correlational analysis
- Does not matter what axis either co-variable is placed upon
- Bar chart, the frequency of each variable is represented by the height of the bars
- Used when data is separated into categories (discrete data) that occupy the x-axis, the y-axis contains the frequency or amount of each category
- Bars are separated to show we are dealing with separate conditions
- Histogram, shows frequency, area of the bars as well as the height represents the frequency, x-axis must start at a true zero and the scale is continuous
- Bars touch each other which shoes the x-axis data is continuous rather than discrete
- X-axis broken down into equal sized intervals representing single categories, y-axis represents the frequency within each interval
- If there is no frequency for one interval, interval remains but without a bar
Types of Distributions
- S|N
- What are the two types of distribution that we look at?
The two types of distribution are normal and skewed
