Lecture #2: Research Methods

Question 1

What is distribution

Answer 1

• Graphing variables

- Random sampling– ensure that the highest [variables] is in the middle and the least is on the extreme sides.

Question 2

Central tendency

Answer 2

• Characteristic of distribution.

- Mean, median and mode.

Question 3

Mean

Answer 3

• Average (e.g. Mean for IQ = 100).

- Affected by outliers (> 2 SD from the mean).

Question 4

Median

Answer 4

• Middle value of a sequence.

- Used if there are many outliers.

Question 5

Mode

Answer 5

• Most occurring data point.

- Least used

Question 6

Standard Deviation

Answer 6

• Characteristic of distribution.
• Typical deviation (+/-) from the mean.
• Low SD = more [variable] around mean (lower is better: prob used random sampling).
• High SD = less [variable] concentration around the mean (higher is concerning: more non-random sampling). IQ SD = 15

Question 7

Conceptual variable

Answer 7

• Conceptual variables are non-tangible things that are difficult to isolate and directly measure.
  e. g. intelligence, happiness, sadness etc.

Question 8

How do you study conceptual variables?

Answer 8

• Measure the behaviours that are associated with the conceptual variables.
• e.g. measure intelligence by looking at performance in university.

Question 9

Operational definition

Answer 9

• A definition that states how a conceptual variable will measured.

Question 10

Construct validity

Answer 10

• How well the operational definition is at measuring the conceptual variable.
• Low Concept validity = no relationship (scrap the operational definition).
• High concept validity = relationship between conceptual variable and operational definition).

Question 11

What makes a useful test

Answer 11

1. High construct validity (must be high—have a relationship)
2. Test-retest reliability: (consistent results when completed multiple times).
3. Inter-rater reliability: (consistent results with different administrators).

Question 12

1. Descriptive research

Answer 12

• First type of research approach.
• Observations
• Cannot infer causation as there is no manipulation of the IV.
• Case studies, surveys and naturalistic observations.

Question 13

Case study

Answer 13

• Examining exceptional examples of an individual.
• Give proof of existence.
• Doesn’t always apply to the rest of the pop’n (rare cases).
• Example of descriptive research (cannot infer causality). - Phineas Gage.

Question 14

Surveys

Answer 14

• Generalized questionnaires

Limitations:
- Problems of self assessment: lying, positive impression management (exaggerating positive traits), malingering (exaggerating problems) and wording of questions could give inaccurate results.

Question 15

Naturalistic observations

Answer 15

• Collecting data by observing
• High external validity (still valid outside lab–as already out of lab).
• No intervention (no problems with self-assessment).
• No negative observer effects.

Limitations:

• Cannot be controlled
• Limited variables difficult to study (infrequent behaviours)
  e. g. in the subway.

Question 16

2. Correlational analysis

Answer 16

• Type of research approach.
• Measuring relationships between data.
• Negative correlation, no correlation, positive correlation.
  Pearson r: (0-1)
• Lower “r” value = weaker correlation.
• Higher “r” value = stronger correlation
• ”+/-“ (“+” = upward trend & “-“ = downward trend).

Question 17

Interpreting correlation

Answer 17

• Correlation does not mean causation (just suggest causation).
• May be a 3rd factor, spurious association (ridiculous correlations) or there may be a relationship but no correlation (non-linearity or U-shaped–making r-value = 0).

Question 18

3. True Experimental research

Answer 18

• Type of research approach.
• Way to determine causality.
• Manipulate Independent variable (IV) and measure the Dependent variable (DV).
• Keep all variables constant except the IV to accurately determine causality.
• Include a control group.
• Randomly assign people.

Question 19

Quasi experiment

Answer 19

• Don’t randomly assign people.

- Not a true experiment.

Question 20

Between-subject design

Answer 20

• 2 + groups are given a different treatment.

- Cheaper

Question 21

Within-subject design

Answer 21

• One subject group is studies 2 + times.
• Good for experiments involving time.
• Control for confounding variables
• Loss of subjects

Question 22

Mixed

Answer 22

• 2 + groups studies 2 + times.
• Control for confounding variables.
• More expensive and take longer
• Loss of subjects.

Question 23

Sampling

Answer 23

• Concern with experiment design
• External validity: make sure that the lab conditions match the conditions in the pop’n.
• E.g. W.E.I.R.D. oversampling (western, educated, industrialized, rich and democratic).

Question 24

Confounding variables

Answer 24

• Concern with experimental design.

- Variables that are not controlled (affecting the DV and/or IV).

Question 25

Demand characteristics

Answer 25

• Concern with experimental design
• Giving clues to subjects (allows for subjects to modify their behaviour)
• So, deceive subjects (although inform of risk).

Question 26

Observer/Hawthorn effect

Answer 26

• Subjects trying more as they know they are being watched.

Question 27

Placebo

Answer 27

• Counters pre-existing expectation about treatment (that can alter behaviour).
• If treatment works: the drug must work more than the placebo group and the control group.

Question 28

Experimenter effect

Answer 28

• Unconsciously affect experiments when they want certain results.
• Solution: double-blind (experimenters and subjects both don’t know which group the subjects are in.