Paper 2 - Research Methods

1. Lab Environment

2. Quasi Experiment

3. Field Experiment

4. Natural Experiment

• Ecological validity: Field/Natural

• Highly controlled: Lab/Quasi

• Demand characteristics: Lab/Quasi

• Extraneous variables: Field/Natural

• Lacks mundane realism: Lab

• Deception

• Right to Withdraw

• Informed Consent

• Protection from Harm

• Privacy and Confidentiality

• Independent Groups Design

• Repeated Measures Design

• Matched Pairs Design

• Demand Characteristics: RM

• Order effects: RM

• Participant Variables: IG

• Time Consuming: IG and MP

In RM, half of the participants experience Conditions AB, and the other half experience Conditions BA.

Refers to how consistently an experiment receives its results

• Internal reliability refers to whether an experiment is consistent to its methodology

• External reliability refers to the extent to which an experiment measures over time and in others

Refers to how consistent an observation is; if there is a correlation of 0.8, an observation is reliable

• Always have more than one observer

• Have clear standardised results

• Split Half Method Assesses internal reliability by splitting a test into two halves with the same participants. If results are the same for both, then it is reliable

• Test Retest Method Assess external validity by giving participants the same test on two separate occasions so participants don't recall the questions

• Standardise methodology

  • All participants undergo the same procedure

  • Use the same conditions for each participant

  • Future research becomes more consistent with higher internal reliability

• Use the same interviewer for all participants

• Make questions clear and concise

Refers to the degree to which an experiment measures what it claims to be measuring

• Internal validity concerns whether the results are due to the IV and not confounding variables

• External validity refers to extent that results can be generalised to other settings

  • Ecological validity

  • Population validity

  • Temporal validity

• Face validity An intuitive measure on whether an experiment appears to be measuring its dependent variable

• Concurrent validity Refers to when an experiment is compared with a well-established, validated test; if results are similar then the current experiment has high validity

• Control groups

• Covert observations have high ecological validity

• Observational behavioural categories must be clear

• Single blind (blind participant) and double blind (blind researcher) reduce demand characteristics

• Standardise procedure to reduce researcher bias

• Volunteer sampling

• Random sampling

• Opportunity sampling

• Systematic sampling

• Stratified sampling

• Convenient: Opportunity/Volunteer

• Demand characteristics: Volunteer

• Participant variables: Random/Systematic

• Researcher bias: Opportunity/Volunteer

• Representative: Stratified/Systematic

• Time consuming: Random/Stratified

• Overt / Covert

• Participant / Nonparticipant

• Controlled / Naturalistic

• Demand characteristics: Controlled, Overt

• Ethics: Covert

• Extraneous Variables: Naturalistic

• Researcher bias: Participant

• Validity: Naturalistic, Overt, Non-participant

Correlations refer to the strength of interaction between two co-variables on scatter-grams

• Positive/Direct: As one increases, so does the other

• Negative: As one increases, the other decreases

• No Correlation

• Used as a starting point to assess the relationship between variables before an experimental study

• Quick and economical (secondary data)

• Correlations don't show causation

• Does not explain why co-variables are connected

• Does not consider a third co-variable

• Structured interviews involve a list of questions in a fixed order

• Semi-structured interviews involve a list of questions, but interviewers can add other questions

• Unstructured interviews involve no set questions and aim to discuss a topic

• Structured: Standardised; lacks depth and validity

• Unstructured: Flexible, interviews can follow up questions for insight; qualitative data may be difficult to analyse; replicability (open ended questions)

A list of predetermined questions that a participant responds with written answers

• Open questions have no fixed range of answers and produce qualitative data

• Closed questions have a fixed range of answers and produce quantitative data

• Open questions: Participants can expand answers and reveal new insight, qualitative data; difficult to analyse (quantitative data)

• Closed questions: Quick to analyse even in large quantities; response bias (always ticking yes, answering 3 out of 5); limited selection of answers that may not represent true thoughts

The observational analysis of content, such as television and magazine, to describe a sample through the material they consume

• High ecological validity (real communications on recent topics)

• When sources are available publicly, research can be replicated

• Observer bias

• Mean: Average (Sum of values/N)

• Median: Middle value from least to greatest

• Mode: Number occurring most often

• Range: Difference between smallest + largest value

• SD: Deviation of individual values from mean value

1. Abstract

2. Introduction

3. Method

4. Results

5. Discussion

6. References

• Design: Experimental type/design + variables

• Participants: sampling method

• Apparatus: Details of material

• Standardisation: Interpretation of variables

• Ethics: DRIPP

• Summary of results

• Relationship to background research

• Limitations/Modifications

• Further research

Article: Author + date + title + journal + volume + page

Book: Author + date + title + address + publisher

1. Research report written

2. Report received by journal editor

3. Report sent to "peers" to be heavily scrutinised for methodology, findings, and conclusions

4. Research deemed unfit is sent back to researcher with criticisms and suggestions

5. Research deemed fit is returned to editor, who decide if it is significant enough for the journal

• Publication bias; editors often decide to publish based on if the research would generate more sales for the journal

• Researcher bias; reports may contradict peers' views and the status quo, which may lead to the ground breaking research being suppressed

T - Theory Construction H - Hypothesis Testing E - Empirical F - Falsifiability O - Objectivity R - Replicability K - Kuhn (Paradigm Shifts)

• Inductive Observation, Hypothesis, Study, Conclusion, Theory

• Deductive Observation, Theory, Hypothesis, Study, Conclusion

Refers to whether the experiment can be proven wrong

• Popper argued scientific theories should be able to survive attempts to falsify them, as nothing can be proven to be completely fact

Kuhn argued scientific advancement occurred through revolutionary "paradigm shifts" that referred to a mass transformation from the status quo to the new opinion, such as in the psychological approaches and scientific models

• Descriptive statistics involve summaries of data in graphs and averages, which are used as an overview

• Central tendency (mean, median, mode)

• Measures of dispersion (range, SD)

• Inferential statistics involve judgements (inferences) to be made about whole populations based on the sample used in an experiment

• Significant refers to the comparison between the OV and CV

• N = Number of participants

• Degree of significance refers to the probability of rejecting the null hypothesis when it is true

• Tail refers to whether the hypothesis is directional

• OV = The value calculated from a statistical test

• CV = The value the OV is compared to determine whether it is significant

1. Sign Test

2. Chi Squared Test

3. Wilcoxon Test

4. Mann-Whitney Test

5. Spearman's Test

6. Related T-test

7. Unrelated T-test

8. Pearson's Test

• Difference or Correlation

• Experimental Design (IG/MP and RM)

• Level of Measure

• Sign Test

• Wilcoxon Test

• Mann-Whitney Test

• Nominal (Categorical data)

• Ordinal (Unstandardised statistical data)

• Interval (Standardised statistical data)

• Type I errors involve rejecting the null hypothesis when true (too optimistic, so false positive)

• Type II errors involve accepting the null hypothesis when false (too stringent, so false negative)