INVESTIGATIONS

INVESTIGATIONS

1. SELF-REPORT

   • Questionnaires & Interviews
     • Two types: Non-directive & Fully structured
       • Non-directive is good for counselling and therapy
       • Fully structured gives the experimenter the answers they want quickly
     • Different types of question: Closed questions, open questions & rating scales
   • RELIABILITY ISSUES - Based on personal opinion - questions often open to interpretation
     • IMPROVED - use more closed questions with categorised answers - repeat experiment
   • VALIDITY ISSUES - Participants may lie or give desired answers - often qual data, considered unvalid
     • IMPROVED - avoid leading questions - inforce confidentiality - use a pilot
   • ETHICS - Doesn't always fully reveal what investigation is investigating - confidentiality - can ask personal question
   • STRENGTHS - source of primary data (valid) - large amount of data quickly
   • WEAKNESSES - open questions often subjective - response bias & social desirability

2. OBSERVATION

   • Participant and non-participant
     • Better understanding - naturalistic - investigator bias - deceptive - highly subjective
     • Ethically good - no investigator bias - generally objective - May misinterpret - unnaturalistic
     • Time and event sampling
   • ETHICS - Obtaining consent and keeping it covert difficult - whether to include results of those unhappy with being part
   • RELIABILITY ISSUES - Dependent on the opinions of observer (subjective) - human error e.g. miss things
     • IMPROVED - Inter-rater reliability - watch through one-sided mirror
   • VALIDITY ISSUES - Behaviours may be misinterpreted
     • IMPROVED - Use a pilot - then decide whether test is valid
   • STRENGTHS - First hand knowledge, both types of data, likely to get natural behaviour
   • WEAKNESSES - Observer effects - Observer bias - Ethical issues

3. EXPERIMENT

   • STRENGTHS - First hand knowledge - easy to establish casual relationship - large samples
   • WEAKNESSES - Investigator bias - extraneous variables - ecological validity (mundane realism)
   • Laboratory, Field & Natural
     • Takes place in an environment controlled by the experimenter
     • Takes place in the subject's natural environment
     • IV is naturally occuring (also called Quasi-Experiment)
   • RELIABILITY ISSUES - Analysing the data from experiment could be subjective or human error - participants have individual differences so complete reliability hard to achieve
     • IMPROVED - inter-rater analysis - use matched pairs design
   • VALIDITY ISSUES - Does the test actually test the dependent variable in relation to the independent variable - lab experiments may not be ecologically valid - are extraneous variables eliminated
     • IMPROVED - conduct a pilot study - use field experiments or Quasi-experiments
   • DESIGNS
     • Independent measures - groups include different participants in each one which are not made to resemble each other
       • Lacks reliability
     • Repeated measures - groups are carried across between conditions so the participants are exactly the same
       • Can result in order effects
     • Matched Pairs - groups include different participants in each one but are made to resemble each other in their characteristics
       • Difficult to make them fully resemble each other, time consuming
   • Null and alternative hypotheses
     • Null - saying there is will be no difference between two conditions
     • Alternative - there will be a difference between two conditions
       • OPERATIONALISE!
   • Displayed by frequency polygons & histograms

4. CORRELATION

   • STRENGTHS - Shows the relationship between two co-variables and strength - not unethical at all - starting point for research
   • WEAKNESSES - Cannot assume cause and effect - secondary data so risks with validity
   • Positive or negative correlation
   • ETHICS - good as all data is secondary and no participants are involved

5. TYPES OF SAMPLE

   • Snowball - Investigator approaches one participant who in turn approaches several others and the sample 'snowballs' to become a large sample
   • Opportunity - Not really a 'technique', researcher uses people that are readily available, usually known to investigator prior to experiment
   • Random - All target population have an equal chance of being selected, usually using a computer to ensure complete randomness, very rarely used
   • Systematic - Finding a starting point within a target population and then obtaining subsequent participants by using a constant interval e.g. every 10th person
   • Stratified - Target population is divided into subgroups and then random samples are taken from the subgroups

Bilder und Dateien der Mindmap