Research methods

Quantitative

Reporting data as numbers

Nominal

Categorical. Qualitative data. No ranking. Mode and range.

Debriefing

Required for any deception. Will provide full disclosure.

Quasi-experimental

Experimental research without random assignment. Limited info about causes

Falsifiability

Required for science. Always an ability for the potential of it being found false

Proof and disproof in science

Absolute proof is impossible. Absolute disproof is impractical.

Principles of Animal Research

3 Rs. Replace: Use live animals if necessary, less intelligent. Reduce: # of animals used. Refine: procedures are more accurate and careful

A confidence interval of 95% means that A. there is 95% probability that the population mean lies within this interval. B. there is a 95% probability that these results can be replicated. C. there is 95% probability that the results of the study are true. D. All of the above.

A

In experimental research, A. one variable is directly manipulated while all others are held constant. B. a relationship between two variables is investigated. C. a single event or a subject is studied in depth. D. All of the above.

A

Researcher finds p=0.04. This means that: A. Probability of Type 1 error is 4% B. Probability that experimental hypothesis is true is 96%. C. Probability of replicating results is 96% D. All of the above

A

When the scores in the group are so low that a difference between the conditions cannot be determined, a ___________ occurs. A. floor effect B. confound C. type I error D. ceiling effect

A

Type 2 error

Accepted H0, but H0 is false

Independent Variable levels

Any level of measurement (nominal, ordinal, interval, ratio).

Human dominion

Any use is permitted, they have no consciousness, does not matter how we treat them.

Animal rights

Any use is unacceptable. Have rights, must treat them like small children.

Mean

Average of the numbers

Variance

Averaged squared deviation of number from mean. Distance, divided by number of scores

The degree to which research findings can be generalized beyond the specific context of the experiment is called A. ecological validity. B. external validity. C. internal validity. D. construct validity.

B

Cost benefit analysis

Benefits must be practical. Costs must be no more than real life

Rejecting the null hypothesis when it is true is called a(n) A. significance testing. B. Type II error. C. Type I error. D. error variance.

C

The degree to which an experiment is methodologically sound and free of confounds is A. external validity. B. ecological validity. C. internal validity. D. construct validity.

C

Operational definition

Clear, concise def of a concept in terms of measurable properties

Inferential statistics compute a ratio between A. effect size and confidence interval. B. null hypothesis and experimental hypothesis. C. type I error and type II error. D. systematic variance and error variance.

D

The variance that is due to unpredictable factors not related to the variable of interest is A. systematic variance. B. confidence interval. C. effect size. D. error variance.

D

Systematic empiricism

Data trumps theories. Planned and structured

Error variance

Degree at which variability can be attributed to other factors

Law

Description of natural phenomenon that holds under specific conditions. Says what happens, not why.

Range

Difference between smallest score and largest. Easily skewed

Experimental

Direct manipulation of 1+ variables. Casual info

Hypothesis

Educated guess. Can be tested by observation or experiment

EEG

Electrical brain activity.

If a researcher does not find a statistically significant difference (p>0.05), then he can conclude that the experimental hypothesis is proven false. True False

False Nothing can be "proven"

If a researcher finds a statistically significant difference (p<0.05), then he can conclude that the null hypothesis is proven to be false. True False

False. Low probability, but it could still be true

Measures of central tendency

Finding most typical value in distribution. Mean, median, mode.

Applied research

Finding research for practical problems. Specific use

Public verification & replication

Findings should be observed, replicated, & verified by others.

fMRI

Functional magnetic resonance imaging. Measures brain activity by detecting associated changes in blood flow

Null hypothesis

H0. No difference.

Alternative hypothesis

H1. Difference

Post-hoc explanation

Hindsight bias. Not good

Basic research

How stuff works. No concern for an immediate affect. "Building blocks"

PET

Imaging test. Organs and tissues

Systematic empiricism

Knowledge gained from experience. Variables controlled, conclusions based on objective statistics, not personal beliefs.

Independent Variable

Manipulated

Effect size

Measure of difference of sizes between 2 variables. Strength of relationship measured. Cohen's d.

Dependent Variable

Measured

Physiological

Measures CNS activity.

Theory

Model. Coherent explanation for large # of facts and observations

Ethical skepticism

Most thoughts of moral rights or wrongs have evolved, so researchers decided.

Converging operations

Multiple methods of measurement and research design to test theory

Levels of measurement

NOIR. Nominal. Ordinal. Interval. Ratio.

Ad-hoc prediction

Negates hindsight bias. Made before collecting data

Median

Number halfway through the set

Ordinal

Numerical degrees. Distances between rankings are not equal. No true zero. Mode, median, range.

Scientific method

Observation -> beginning of theory/hypothesis -> generate predictions -> test predictions ->prediction confirmed, model accounts for problem -> prediction refuted, model is inaccurate

Qualitative

Reporting data as field notes.

Using terminology introduced in scientific method class, explain how scientific misconduct undermines scientific method and how procedural requirements can be used to uncover cases of scientific misconduct.

Plagiarism, data fabrication, failing to acknowledge all contributions, failing to acknowledge conflict of interest.

Statistical power

Probability of not committing type 2 error. Probability of rejecting H0 when false. Can be used to calculate minimum sample size to detect a given difference before experiment.

p-value

Probability of type 1 error.

Operational definition

Question to method. Need to simplify to measure

Self report

Questionnaires and interviews. When direct obs is impractical or impossible

Measures of variability

Range, variance, standard deviation

Interval

Ranked on a scale with equal distance. No true zero. Mode, median, mean, range, variance, standard deviation.

Ratio

Ranks are equal distance. True zero. Mode, median, mean, range, variance, standard deviation

Observational method

Recording behavior in lab or real. Qualitative. Naturalistic and lab.

Type 1 error

Rejected H0, but H0 is true

Correlational

Relationship among 2+ variable. Limited info about causes

Belmont report

Respect for persons. Minimize risks, maximize benefits. Truthfulness

Testability

Science only deals with empirically answerable questions.

Subject variable

Special IV. Cannot be manipulated, and is not a true IV.

Standard deviation

Square root of variance. Normalizes the squared variance. Describes data variability.

Informed consent

Sufficient disclosure. Rarely waived.

Descriptive statistics

Summarizes and describes group of scores.is not descriptive research. Measures of variability and measures of central tendency.

Science vs. pseudoscience

Systematic empiricism. Public verification. Replication. Falsifiability.

Deontological

Universal right and wrong

Animal welfare

Use requires cost-benefit analysis. Can feel pain, and we have a moral obligation to balance animal suffering and human use.

Utilitarian ethics

Used now. Cost-benefit analysis

Mode

Value that appears most often.

Systematic variance

Variance interested in. Degree at which variability can be attributed to investigated factor

New predictions

Very important. Good theories need this. Affected by post-hoc and ad-hoc predictions

Control group

Zero level of IV. Controls extraneous variables. Measures baseline. NOT USED: @least 2 levels of IV, don't care about baseline, no sensible baseline level.

Confound

extraneous, unwanted variable that differs between groups