Year 2

• Case studies

  Involve a detailed analysis of an individual often in unusual events. Tend to be longitudinal.

  • Provide rich, detailed insights into perhaps unstudied and atypical behaviours.
  • Contributes to our understanding e.g Phineas Gage
  • Lack of generalisability, due to small sample size. Open to researcher bias, personal accounts from participants can be flawed e.g childhood memories
• Content analysis 


  Type of observational research, studying people indirectly via the communications we produce. E.g Letters, transcribed speech, art, films…

  1. Coding - Identifying behavioural categories and assigning them a code e.g M = money. Produced quantitative data
  2. If we want qualitative data, we use thematic analysis. We code behavioural categories, but write a summary.

  • Strength - no ethical as individuals aren’t being observed. High external validity
  • Limitations - people are studied indirectly, analysed outside the context they were produced, lack of objectivity in TA
• Reliability

  How consistent findings are

  Ways of assessing reliability:

  • Test-retest -> giving the same test or questionnaire to the same person or people on different occasions. Tests need to be similar to be reliable. Time allowed mustn’t be too short that the p can recall their answers, but not long enough that their views change.
  • Inter-observer reliability -> year 1 gizmo
• Improving reliability

  Questionnaires

  • Test-retest, comparing two sets of data should produce a correlation +.80. Questions that have low test-retest reliability may need to have questions rewritten or removed. E.g replacing open questions with closed.

  Interviews

  • Use the same interviewer each time, if not possible they need to be trained to make sure one interviewer doesn’t ask a too ambiguous question.
• Improving reliability (2)

  Observations

  • Behavioural categories need to be operationalised, measurable no self-evident. Cannot overlap. Can also train observers.

  Experiments

  • Use standardised procedures, can be replicated, using the same instructions for all p’s
• Types of validity


  Internal validity

  • Whether the effects observed in an experiment due to the manipulation of the IV and not some other factor

  External validity

  • Whether the effects can be generalised

  Types of external validity

  1. Ecological validity - can we generalised to every day life situations?
  2. Population validity - can we generalise to the general population?
  3. Temporal validity - can we generalise to modern society?
• Ways to assess validity

  1. Face validity - whether attached appears to measure what it is meant to measure. Can be done by simply ‘ eyeballing’ the measuring instrument, or giving it to an expert to check.
  2. Concurrent validity - demonstrated results obtained are close to or match other data which used an established and well recognised test. E.g the IQ test and the Stanford-Binet test. Close agreement is indicated when the score exceeds +0.80
• improving validity


  Experiments.

  • Using a control group as a way of comparison. Standardised procedures. Use a single blind or double blind study.

  Questionnaires

  • incorporate distractor questions, so participants can’t guess what the researcher is trying to investigate, reduces demand characteristics

  Observations

  • Using a covert observation and operationalised behavioural care categories
• improving validity (2)

  Qualitative data

  • Needs to demonstrate interpretative validity (whether the researcher thoughts of events matches the participants opinion of events) which can be shown through coherence of the research’s narrative and the inclusion of direct quotes from participants within the report.
  • Validity is further enhanced through triangulation - using a number of different sources as evidence. E.g interviews with friends and family, content analysis of personal diaries, observations etc
• How to choose a statistical test?

  1. is it a test difference (experiment)or a correlation?
  2. What is the design of the study? Independent groups = unrelated design. Matched pairs and repeated measures = related design.
  3. What is the level of measurement? Nominal, ordinal, interval
• Levels of measurement 


  nominal data - data in categories, data is discrete which means only one item can appear in each category. E.g ‘yes’ or ‘no’

  Ordinal data - data that can be put in order, mostly quantitative e.g 1st, 2nd…

  Interval data - based on numerical scales that have a universal unit of measurement e.g hours, seconds, MPH

  NOTE -> when there are two sets of data, use the one which is worth ‘less’.
• Stats table

  Tests which use interval data are ideal tests, they are called parametric tests

  
• Probability and significance

  The null hypothesis -> states that there will be no relationship or difference. We either except the null hypothesis or rejected, according to whether the statistical test determines whether the data is significant or not.

  Levels of significance and probability -> You will be told a significant level to use, if not, use 5% or 0.05. Balances the chances of type 1 and type 11 (2) errors.
• using statistical tables


  Calculated value - the number we used to compare to the table.

  Critical value - the number in the table which determines if the research findings are significant or not.
• using statistical tables (2)

  1. Is it a one tailed or two tailed test?
  2. How many participants are in the study? N = number of participants. DF = special equations for certain tests.
  3. Level of significance?

  Levels of significance -> 0.05 is the standard in psychology, however, we use 0.01 when there is a human cost e.g drug trials. The studies can be repeated in future.
• Type 1 and 11 errors


  Type 1 = when the null hypothesis is rejected when it should’ve been accepted. False positive. more likely when we use a lenient significance level e.g 0.1 or 10%

  Type 11 = when the null hypothesis is accepted when it should’ve been rejected. False negative. More likely if the significance level is too stringent e.g 0.01 or 1%
• equations for DF - need to know

  Related t-test = DF = N-1

  Unrelated t-test = DF = N(a) +N(b) -2 -> a + b are number of p’s in each condition/group

  Chi squared = DF = (rows-1) x (collums-1) -> contingency tables

  Pearson’s = DF = N-2
• the sign test -> need to know method

  Put a + if the first column is greater than the other, and the opposite with - signs. If the are equal put =. Add the +’s up and the -‘s up.

  • Smallest number of + vs - is the CALCULATED VALUE.
  • N = (number of participants) - (number of = signs)
• Designing your own study -> doing now with ostrander

  COMPLETE
• Content analysis (2)

  Notes - If data is first spoken, we need to transcribe it. To identify behavioural categories we need to 1st look at the source to develop them. Using coding, we can identify emergent themes such as why people post to negative things on social media.
• Sections of a scientific report

  Abstract - short summary, includes aims and hypothesis, procedures, results, conclusions.

  Introduction - details relevant theories concepts and studies to study

  Method - design, sample, materials used and its procedure, ethics.
• Sections of a scientific report (2)

  Discussion - summarise results with qualitative data, discussed in the context of the evidence presented in the introduction. Wider implications of the research. Suggest limitations of the research.

  Referencing - Details of any source material used in the report. E.g. Skinner, B.F (1963) science and human behaviour. New York : MacMillan.
• Features of science

  Falsifiability (Kuhn and Popper agreed that this is a key criterion of a scientific theory)

  • Scientific theories should hold themselves up for hypothesis testing and the possibility of being proven false. This is why we have an alternative hypothesis and a null. E.g Freuds psychodynamic theory is unfalsifiable.

  Applicability (element of Popper’s hypothetico-deductive method.

  • Important role in determining validity and the reliability. Can assess the extent of which the findings can be generalised.
• Features of science (2)

  Objectivity and the empirical method

  • Researcher cannot influence findings. Need to keep a high degree of control. Locke argues that a theory cannot claim to be scientific unless that it has been empirically tested. Empirical testing = Emphasise importance of data collection based on direct sensory experience. E.g a lab study or observation.
• Features of science (3)

  Paradigms- Kuhn suggested what distinguishes scientific disciplines from non-scientific is a share set of assumptions and methods (paradigm). Suggests that social sciences lack a universally accepted paradigm. Psychology is a pre-science as there are too many internal disagreements e.g the different approaches. An example of a paradigm shift is from Newtonian physics to Einstein‘s theory of relativity
• Features of science (4)

  Theory construction and hypothesis testing.

  Theory construction = Gathering evidence via direct observation ( empirical method).

  hypothesis testing = Needs to be scientifically, tested and theory should suggest a number of possible hypothesis. Hypothesis attested by using systematic and objective methods to determine whether it will be supported or refuted. The process of deriving new hypothesis from an existing theory is known as deduction..
• Descriptive statistics


  Graphs, measures of dispersion and central tendency
• Empirical data


  Feature of science, when we can visibly see it, object. Eg Freud had no empirical data because you cannot measur the unconscious directly
• writing ethical forms


  Consent forms = aims of the study, what p’s are required to do, right to withdraw. Tickbox of consent and signature.

  Standardised instructions = recipe style of instructions, check people understand instructions.

  Debriefs = must tell p’s about all conditions, outline deception, data withdrawal, right to withdraw, privacy. Leave a pretend email at the bottom saying to email if have any concerns.