Summary
Research Methods in Psychology

Main Components
Theory
Hypothesis
Operational Definition

Types of Measures
Qualitative & Quantitative
Self-Reports
Behavioural Observations
Archival Records

Research Settings & Observational Studies
Laboratory Research
    Controlled Observation
    (Experimental Studies)

Field Research
    Naturalistic Observation
    Participant Observation

Descriptive Studies
    Case Studies
    Surveys

Tests & their characteristics

Correlational Studies

Experimental Studies
    Independent Variable
    Dependent Variable
    Experimenter Effects

Statistics
    Descriptive
    Inferential

Ethics

Theory - an organised system of ideas that is used to explain a phenomenon

Hypothesis - a specific statement that describes or explains a phenomenon, Usually derived from a theory and used to test the theory.

Falisifiability is the principle that all hypotheses must be potentially falsifiable to make them scientific (we'll look again during inferential stats).

Operational Definition - a statement of how the phenomenon is to be observed and recorded. Provides a standardised way of measuring it.

Types of  Measures
Self-Reports - These include interviews, questionnaires & diaries. The people being studied provide their own answers to research questions.

-may be quick and easy, revealing 'inner states or feelings' but not always accurate, may fake good. Can be quantified through scaling and averaging.

Behavioural Observations - Counting and coding behaviour in specific settings. May be "objective" (less biased) but trivial - what does it mean? People may behave differently in such settings, what do they really think?

Archival Records - Medical, school, or government records, newspapers, etc. May be "unbiased" but not always the specific information that is desired, incomplete or missing important or crucial elements.
 

Research Settings & Observational Studies

Laboratory Research- done in the confines of the laboratory, offering a controlled situation for observation of specific behaviours or conversations. May involved reactivity or the fact that the methods change the participant's behaviour.

- Provides precision and standard setting for all participants but they may not act in a natural fashion.

Field Research
Naturalistic Observation - research done in an everyday setting where people are in their own normal environments. Does not offer the control of the lab, but may expect more natural behaviour.

-As in laboratory may have Observer Bias.

Participant Observation - the personal making the observations is also a participant in the setting. May have an impact on the results (behaviour of people), but also has the advantage of providing and "insider's" perspective on the situation.

Qualitative (Descriptive) Studies

Case Studies - Careful and detailed (in depth) analysis of specific people or groups. Offers a rich account of the lives or conditions that are under investigation.

-May not be generalizable to all persons, however, may still reveal general patterns  or qualities that people share.
 

Surveys- Telephone, internet, or paper and pencil collections of questions. May be open ended, providing qualitative results or may be coded where responses are tallied up and statistics are used to provide information about general trends.

Important considerations are the populations and samples used in the surveys.
Population- a very large number of individuals, usually not homogenous (is not the same everywhere, but diverse and clustered)

Sample - a very small number of individuals taken non-randomly from the much larger population. Problems arise as to whether or not the sample is representative of the population.
 

Tests- Instruments for assessment that are used to measure or evaluate psychological abilities, traits, interests, values, abilities, etc.

Objective tests - Inventories that have a single (numerical) score or several sub-scores and one or more composite scores. (E.g. IQ tests)Critique

Projective tests - those designed to assess unconscious feelings or motives.
(E.g., TAT, 2,  inkblots  

Qualities of Tests
Standardized: Those with a uniform procedure for giving and scoring, often with reference norms. (IQ)

Reliability - The amount of agreement between the scores that one person has on the same test taken twice.

Validity - the degree to which a test represents the categories that it intends to measure. Does it measure what it intends to measure?

Correlational Studies - examine how two or more variables or characteristics are regularly associated.

Correlation - a measure of the strength of the relationship of a collection of people's scores on two or more variables.

Positive correlation occurs when both variables tend to go together (as one is high so is the other)

Negative correlation occurs when they go against each other. One goes up as the other goes down.
 No Correlation  

Experimental Studies are done make inferences about the cause of various phenomena, offering an explanation of what made it happen.

Random Assignment is done to control for extraneous variables (influences) that might have an impact on the results. Having a sample that is both representative and evenly distributed into the treatment (experimental) and control groups is important.

Population Sample - are these people like everyone?
 
Independent Variable - the experimental conditions that are manipulated or "varied" to produce an effect (teaching style, drug, exposure to X, placebos....)

Dependent Variable - the various resultant behaviours or conditions that are produced in the experimental situation. E.g. test scores, GPA, IQ, violence...

Experimental group gets the treatment or some manipulation of the independent variable.

Control group are people who get placebos, waitlisted or given "time out" or distractors.

Placebo effects occurs when a person thinks they are being  given a real treatment, but are given something inert, like an empty pill, a sham treatment.

Experimenter Effects

Experimenter bias - usually unconscious influence on the outcome of a study due to the experimenter's actions. Eg, smile, knowing who got the treatment...

Single-blind study - is where the subject or participant does not if he or she is getting the treatment or not.

Double-blind studies are when the experimenter and the participants do not know who is getting the treatment.

Descriptive - general summary numbers that represent the pooled response of many people on the same variable, or the same person on various scores.

Mean - arithmetic average found by adding all scores and dividing by the number of scores.

Mode - is the most common (frequent) score in a data set. On a distribution (graph) it is the highest peak, sometimes there are two + modes (bi-modal).

Range - is the total distance on a scale that is covered by the scores. Found by subtracting the lowest score from the highest.
 
Standard Deviation - is a measure of dispersion that indicates how closely grouped the scores are, the average difference between any score and the mean.
 
The Bell Curve

Correlation - as above, deals with the strength and direction of relationships, not causation.
 

Inferential Statistics- are used to make "educated guesses" about the results from experiments.

Usually pertain to the probability that the results could have occurred by "chance".  e.g., binomial distribution, t

Give some estimate of the amount of confidence that an experimenter can have.
 
We test the "Null Hypothesis"  - that there is no difference between the treatment and the control groups, we either reject the null hypothesis (if there is statistical significance) or we fail to reject the null hypothesis (if there is no statistical significance).

If we reject the null hypothesis, as according to falsification, then we can turn to
 the theoretical hypothesis (alternate) and consider it to be be still viable, not proven.

There is a family of significance tests that can be performed, such as:

t-tests - are used to estimate the confidence of any observed difference between the means of 2 groups.

Statistical Significance* - is generally at the level of a 5% (alpha = .05) that the observed differences occurred by chance. This is also called the p-value or the probability of error.

ANOVA (Analysis of Variance) - a complex t-test that considers many possible influences IV &/or DV

Problems with inferential statistics in part have led to the "Replication Crisis" in psychology and other sciences.

Ethics - is the consideration of issues of value or right and wrong. In psychology ethics has become a central feature of research, particularly since certain studies appear to have been "unethical".

The question remains: are we being sensitive enough or too sensitive? Are we ethical or Politically Correct?

Stanley Milgram's Obedience Experiment
      https://www.youtube.com/watch?v=fCVlI-_4GZQ

Philip  Zimbardo's Stanford Prison Experiment & TLE
        http://www.prisonexp.org/   http://www.thelucifereffect.com/

Canadian Psychological Association Code of Ethics
CPAfor ethics in psychology are used to help people avoid compromising situations in therapy, research & teaching of psychology.

Additional Links:

 For an online Stats book
             http://onlinestatbook.com/rvls.htm

 Vassar Stats - Online Statistics program (Free))