Methods in Psychology

Methods in Psychology

1. Probability (5% level)

   • Type I Error
     • If HE is accepted and later turns out not to be true
     • More likely if a lenient probability is used e.g. 10% level.
   • Type II Error
     • If HE is rejected and it later turns out to be true
     • More likely if a strict probability is used e.g. 1%

2. Parametric and Non-Parametric Tests

   • Non-Parametric
     • Use nominal + ordinal data, simple, often subjective scales, tests for significance, convert scores to ranks
     • sign test, Wilcoxon test, Mann-Whitney, chi-squared + Spearman's Rank Order Correlation test
   • Parametric
     • Use interval/ratio data, contains info. about size of difference between scores, more powerful test of significance
     • Related t-test, independent t-test, Pearson's Product Moment Correlation test

3. Correlations

   • Test for a link/ association between 2 key variables
     • Research Hypothesis (instead of HE) used to predict relationship e.g. 'there will be a significant correlation between variable A and variable B'
   • Positive Correlation
     • Variable A increases as variable B increases
   • Negative Correlation
     • Variable A increases as variable B decreases
   • Zero Correlation
     • No pattern/ relationship

4. Directional Hypotheses & Tails of Test

   • Non-directional Hypotheses
     • There will be a significant difference in the DV when IV is changed
     • Used by researchers to reduce experimenter bias (e.g. when rejecting anomalies
     • 2-tailed statistical test
   • Directional Hypotheses
     • Predict a significant increase or significant decrease in DV when IV is changed
     • Only used when there is strong previous evidence about the direction of the results
     • 1-tailed Statistical test

5. Reliability

   • External Reliability (test re-test)
     • How consistent a method measures over time when repeated (same people under same conditions e.g. IQ test
     • Correlate the results of a test conducted on one occaision with the results of a test conducted on a later occaision (with same subjects). High positive correlation = reliable method
   • Inter rater/ Observer Reliability
     • content analysis/ observational studies
     • 1.) Clearly describe behaviour to record (categorise)
       • 2.) Observe same ppts. for same amount of time, independently
         • 3.) Compare results, if 90% similar = reliable, consistent method

6. Validity

   • To what extent have you measured the factor you set out to measure; whether other factors have affected your results
   • e.g. Bigelow set out to measure children's understanding of friendship, but instead may have measured their language ability
   • Methods of establishing Validity::
     • 1.) Face Validity: look at the method to see if it appears to measure what it is supposed to measure; open to bias
       • 2.) Concurrent Validity: give same ppts. the new test, and an established test which measures the same thing (or questionnaire). Test scores for correlation - significant positive = valid
         • 3.) Ecological Validity:where the conditions of a test are so controlled that they do not represent how a situation would occur naturally

7. Sampling

   • Target Population = people that your hypothesis relates to and from whom the sample is taken
   • Generalise = predict future behaviour of TP on basis of the result of sample
     • Q-Sort Cards: used to measure self esteem, select 4 cards, convert into score of mood
       • Reliability: test re-test = do test again on same ppts. after two weeks, same results = reliable
         • Validity: Compare result's correlation with an established study (questionnaire), significant +ve correlation = valid
   • Random/ Systematic Sample: everyone in TP has an equal chance of being selected, assigned numbers
     • Opportunity Sample: people who are willing and convenient, likely to be unrepresentative
       • Stratified Sample: Divided into groups (e.g. gender, nationality, age) then random representative sample from each group

8. Ethical Guidelines

   • Informed Consent
     • Verbal or written consent after information about task is explained, no deception
   • Consent for studies involving children
     • Consent from parents (if under 16) and child wherever possible
   • Right of Withdrawal
     • Informed of right to leave study at any time in brief, informed of right to withdraw information in debrief
   • Deceiving Participant
     • Explained in debrief, consent to be deceived, knowing aim may affect behaviour
   • Protection from Physical and Psychological Harm
     • Must be aware of consequences of investigation, should feel no worse about themselves after investigation, normalise by informing of other ppts., counselling

9. Hypotheses

   • Experimental (HE)
   • Null (Ho)
   • Alternative (H1)
   • Research (HR)

10. Descriptive Statistics

    • Averages
      • summarise set of data into 1 value
        • Compare 2 sets of scores
          • Central point of 2 sets of scores
      • Mean: interval/ratio data
      • Median: Ordinal data
      • Mode: Nominal data
    • Measures of Dispersal
      • How spread out a set of scores are
      • Tells yous if your ppts. scores are consistent or variable
      • Range: largest - smallest value, affected by extreme scores
      • Standard Deviation: uses all scores, use for interval/ ratio data, measures the deviation of each score from the mean

11. Quantitative & Qualitative Methods

    • Quantitative
      • +Controls extraneous variables, +Establish C&E. +Can test hypothesis, +Can carry out statistical tests
    • Qualitative
      • +Generates new theories, +High face validity, +High ecological validity, +High level of detail; understand behaviour better