Research Methods

• Variables


  DV - What's measured

  IV - What's changed, in order to measure the DV

  Extraneous Variable - A variable, other than the IV, which may affect the DV

  Confounding Variable - A variable, other than the IV which does affect the DV
• Hypothesis


  Alternative Hypothesis - States that there will be a relationship/effect between the IV and the DV

  Null Hypothesis - States that there will be no relationship/effect between the IV and the DV

  Directional/One tailed hypothesis - States that there will be a relationship/effect between the IV and the DV and states the direction

  Non directional/two tailed hypothesis - States that there will be a relationship/effect between the IV and the DV, but doesn't state which direction
• Types of Experiments


  Lab - Controlled, usually artificial setting, researcher manipulates IV and measures DV

  Field - Natural setting, researcher manipulates IV and measures DV

  Natural - Researcher does not manipulate the IV (the change would be occurring regardless of whether or not the experiment was taking place) and measures the DV

  Quasi - Researcher does not manipulate the IV but measures the DV, the IV is not present in everybody and therefore the researcher cannot use a random allocation of PP's
• Evaluation of Experiments


  Lab - (S) Highly scientific, high control over variables, easily to replicate

  (W) Low mundane realism/ecological validity, risk of demand characteristics

  Field - (S) Higher in mundane realism/ecological validity, lower risk of demand characteristics (W) Less control of extraneous variables so lower internal validity, potentially more expensive

  Natural/Quasi - (S) Allows research into things where the IV cannot be manipulated for ethical or practical reasons, can study real life issues (W) Cannot establish cause and effect as the IV is not being directly manipulated
• Evaluation of Experimental Design

  Repeated Measures - (S) Fewer PP's are needed, (W) order effects, increased risk of demand characteristics

  Independent Groups - (S) No order effects so reduced chance of demand characteristics, (W) requires more PP's and can't control PP variables

  Matched Pairs - (S) Minimises the risk of extraneous variables, (W) Time consuming to match every PP's on variables,m
• Counterbalancing and Random Allocation

  Counterbalancing - Used to overcome order effects (RM) e.g ABBA Method - Every participant experiences both conditions but half do A then B and the other half do B then A

  Random Allocation - Used to overcome issues with extraneous variables – put PP names in a hat and draw out the names so every other person goes in condition A 
• Sampling


  Opportunity - Selecting the PP's who are the most convenient to you

  Random - Every person in the same population that has the same chance of being selected

  Stratified - The sample is representative of sub groups of the population

  Systematic - Selecting every nth PP from sampling frame

  Volunteer - Advertise for PP's in newspaper etc.
• Evaluation of Sampling


  Opportunity - (S) Less time consuming, (W) bias as PP's drawn from small % of the population

  Random - (S) Unbiased, (W) Time consuming as you need a list of the whole population

  Stratified - (S) Generalisable as it's proportionate to the population, (W) Time consuming

  Systematic - (S) Unbiased due to objective system, (W) Time consuming

  Volunteer - (S) No coercion, (W) Volunteer Bias
• Ethical Issues


  Informed Consent

  Deception

  Right to Withdraw

  Protection from Harm

  Confidentiality

  Privacy
• Observational Technique


  Naturalistic - Carried out in an everyday setting where the investigator does not interfere in anyway

  Controlled - Behaviour is observed under conditions where certain variables have been organised by the researcher

  Overt - PP's are aware that their behaviour is being studied

  Covert - PP's are unaware that their behaviour is being studied

  Participant - Observations made by somebody who is participating in the activity being observed

  Non - PP - Observations made by somebody who isn't participating in the activity being observed
• Evaluation of Observational Techniques

  Naturalistic - (S) High in ecological validity, (W) less control makes it harder to account for the reason for behaviour

  Controlled - (S) Can focus on specific aspects of behaviour in replicable environments, (W) Lower in ecological validity

  Overt - (S) Ethically sound, (W) Hawthorne Effect as PP's know that they're being observed

  Covert - Behaviour is more likely to be genuine as PP's don't know that they're being observed, higher in external validity, (W) Ethical issues such as privacy/deception
• Evaluation of Observational Techniques pt.2

  PP - (S) Researchers can experience the phenomenon first hand, (W) Might not remember everything as it will be reported at a later date

  Non PP - (S) More objective as the researcher is separate from the PP's, (W) Might not truly understand the the behaviour being studied
• Evaluation of Self Report Techniques

  Questionnaires - (S) Easier to distribute to large numbers, respondents might be more willing to disclose personal info, (W) Can only be done by those who can read and write, answers can't be clarified if not understood, acquiescence bias.

  Structured - (S) Easily replicated, standardised - increased reliability, (W) No flexibility, can't follow up with PP if they say something interesting.

  Unstructured - (S) Can follow up in detail on interesting points, (W) Harder to analyse, lacks objectivity, interviewer bias may occur.
• Observational Design


  Unstructured - Relevant behaviour is recorded but there is no system in place. Typically used in pilot studies

  Structured - Systematic methods used to organise and record behaviour

  Behavioural - Breaking up behaviour into operationalised variables that can be measured specifically, all behaviour must fit into a category and no behaviour should fit in more than one

  Event - Counting the behaviour everytime that it occurs

  Time - Recording whatever is happening e.g every two minutes
• Self Report Design 


  Questionnaires - Open Questions - Invite different responses from different PP's – tends to produce qualitative data, Closed Questions - Have a predetermined range of responses – tends to produce quantitative data. There should be no ambiguity, leading, double negatives or double barrelled questions

  Interviews - The interviewer should consider whether to write notes now or at a later date, consider non verbal communication e.g smiling or nodding, and ask appropriate questions
• Content Analysis


  An observational study in which behaviour is usually observed indirectly in visual, written or verbal material – information can be qualitative or quantitative

  Qualitative - Looking for similarities which are turned into codes – rather than counting the number of appearances, quotes are used to demonstrate the code

  Quantitative - Tallying the number of appearances of information in behavioural categories 

  Coding - The process of placing qualitative or quantitative data into categories
• Content Analysis pt.2


  Coding Evaluation - (S) High ecological validity, based on real life observations of existing content, easy to replicate / (W) different observers might interpret the information differently, risk of cultural bias in the observer
• Thematic Analysis


  Read and re-read the data thoroughly; no notes should be made

  Break the data into meaningful units 

  Assign a code to each unit

  Combine simple codes into larger categories

  Check the emerging categories against a new set of data
• Evaluation of Case Studies

  Strengths:

  Rich, in depth data / can study behaviour where ethically and practically an experiment wouldn't be possible

  Weaknesses:

  Low generalisability, problems of confidentiality and informed consent
• Features of Science


  Control - Controlling variables in an experiment in order to establish cause and effect

  Objectivity - Research should not be affecting by the expectations of the researcher or their personal opinions

  Replicability - Repeating the study to compare results can see if the results are valid; scientists must record their procedures carefully

  Empiricism - Information is gained through direct observation and experience 
• Features of Science pt.2

  Theory Construction - Deduction – theory construction occurs at the beginning of the process after making observations / Induction – starts with observations of phenomena and then hypotheses are developed and tested which eventually leads to theory 

  Falsifiability - being able to prove a hypothesis wrong makes something scientific – scientific research should be able to be disconfirmed

  Paradigm Shift - Through scientific revolution, scientific beliefs change; through disconfirming theories, everyone gradually moves to accept the same things 
• Improving Reliability


  Inter - Observer Reliability - The extent to which there is agreement between two or more observers in observations, calculated as a correlation co-efficient. (If inter- observer reliability is low – behavioural categories need to be operationalised better or observers might need more training)

  Test - Retest - Same test or interview is given to the same PP's on two occasions to see if the same results are obtained, scores are compared using correlation. (If the test-retest reliability is low – test might be ambiguous so needs rewording)
• Improving Reliability pt.2


  Experiments -  Standardisation of experiments can help to ensure reliability and that every PP gets the same experience
• Inferential Testing


  Significance Level - The probability that the results occurred by chance, typically p<0.05 (0.01 for drug trials)

  Type 1 Error - Researcher rejects a null hypothesis that is actually true

  Type 2 Error - Researcher accepts a null hypothesis that is actually false
• Evaluation of Qualitative / Quantitative Data

  Qualitative - (S) In depth, can gain greater insight into behaviour/thoughts etc. (W) Harder and more time consuming to analyse

  Quantitative - (S) Easy to analyse and draw conclusions from, (W) Can oversimplify reality and lacks depth
• Skewed Distribution


  Positive Skewed Distribution - Most of the scores are clustered to the left – there are a few extreme scores which have an effect on the mean – mean is higher than mode or median (happy whale)

  Negative Skewed Distribution - Most of the scores are clustered to the right – there are more lower scores so the mean is lower than mode or median
• Peer Review


  Practice of using independent experts to assess the quality and validity of scientific research and academic reports

  Strengths: Ensures the validity and reliability of studies

  Weaknesses: Due to the competitive field, not all reviewers will be objective, File Drawer Problem - Studies which are not significant are less likely to be peer reviewed and published (inaccurate representation of studies)
• Correlations


  Positive - As one co - variable increases, so does the other

  Negative - As one co - variable increases, the other one decreases

  Coefficient - The product of statistical tests that measure the direction and strength of correlations – it is always a number between – 1 and +1. The sign indicates the direction and the number indicates the strength – the closer to 1, the stronger the relationship. (Ideally 0.8+)
• Measures of Central Tendency

  Median - Middle number when in order

  Mean - Add all numbers up then divide by the amount of numbers

  Mode - The most common number
• Measures of Dispersion


  Range - Biggest number take away the smallest

  Standard Deviation - Shows to what extent the values in a data set deviate from the mean. It is calculated using all of the values, and so is arguably more representative than the range.
• Stats Test


  Table

  
• Sign Test


  Work out the difference between the two sets of data

  Add up the amount positive and negative differences, ignore PP's which have no difference (e.g + = 5, - = 3)

  The less frequent sign is your S Value (e.g 3)

  Work out the N Value, which is the number of PP's not including the 0 values

  Compare your calculated value (S) to the critical value, found in the table