Study Guide: Research Methods

ways of knowing (I, A, R, E, SM)

intuition, authority, rationalism, empiricism, scientific method (systematic empiricism plus logic to test our ideas)

Authority

Accepting knowledge because a trusted source says it is true.

Intuition

Accepting knowledge because it feels true, without evidence or reasoning.

Rationalism

Acquiring knowledge through logical reasoning.

Empiricism

Acquiring knowledge through observation and evidence.

Systematic empiricism

Careful, systematic observation and measurement, core of the scientific method

science:

Science shares an approach not subject matter

It is to describe (observe), predict (anticipate behavior), and explain (cause)

Publication makes science a social process (collaboration)

Small samples unreliable to open access shows to public and allows for recreation

Can’t answer non-empirical thoughts like “what is crime” that is values

Empirical questions ask what is; value questions ask what ought to be. Science answers the first and informs the second, but cannot settle it.

Pseudoscience

Claims presented as scientific that lack one or more features of science.

Falsifiability

A scientific claim must say what evidence would count against it

Folk psychology

Everyday beliefs about human behavior; often incorrect, which is why psychology uses science.

Confirmation Bias- notice cases that fit and forget ones that do not

Skepticism

Searching for evidence and considering alternatives before accepting a claim.

Tolerance for uncertainty

Withholding judgment when there is not enough evidence to decide.

Basic vs. applied research

Basic research learns about behavior for its own sake; applied research solves a practical problem.

Research literature

All published research, mainly journal articles and scholarly books.

Theory

A broad explanation that accounts for larger bodies of data (tested, not guessed like evolution or cognitive dissonance theory to explain feelings of discomfort)

Hypothesis

A specific prediction about the outcome of a particular study.

sees if theory holds

Hypothetico-deductive method

Deriving hypotheses from a theory, testing them, and revising the theory based on results.

testable and falsifiable (result that could disconfirm it), logical (grounded in theory not random guess), positive (predicts an effect/relationship bc a no difference result can’t confirm a theory)

Variable

Something that varies across people or situations; quantitative (e.g., age) or categorical (e.g., major).

Population and sample

A population is the whole group of interest.

A sample is the subset selected to participate.

Simple random sampling

Every member of the population has an equal chance of being selected.

Convenience sampling

Selecting people who are nearby and willing to participate; may not represent the population.

Experimental vs. non-experimental research

Experiments manipulate a variable.

non-experimental research measures variables as they naturally occur.

Measures of central tendency and dispersion

Mean (average), median (midpoint), and mode (most frequent) describe the typical score; range, standard deviation, and variance describe spread.

what researchers should do for (scientific community) also bc failing to have research ethics affects research participants, society, and scientific community

No fabrication or falsification: never invent or alter data. 

No plagiarism: credit others' ideas and words. 

Competence: work within the limits of your training, and report methods fully so others can replicate.

Its never one study its years of studies 

four moral principles (add definition)

Weigh risks against benefits

act responsibly with integrity

seek justice

respect people's rights and dignity

Informed consent

Obtaining and documenting agreement/ a contract after informing people of everything that might reasonably affect their decision.

Show expected risks/benefits for participants

Deception

Misleading participants about a study.

The APA Ethics Code allows it only under specific conditions.

Debriefing

Informing participants about the study's purpose and any deception after their participation

Institutional review board (IRB)

A committee that reviews research protocols to make sure they meet ethical standards.

5 members with diff backgrounds and one unaffiliated with institution

IRB checks risks are minimized, benefits outweigh risks, participants treated fairly, and consent procedures are adequate

IRB Three levels of review

Exempt: the lowest risk. Normal educational research, non-sensitive surveys kept confidential, and existing public data. Approved once, with no continuing review. 

Expedited: no greater than minimal risk, the level healthy people face in daily life. One IRB member or a delegated committee can approve. 

Full board: risks/anything greater than minimal risk goes to the full IRB.

APA Ethics Code

The most important ethics code for psychology researchers; Standard 8 covers research.

Written ethics codes

Documents such as the Nuremberg Code, Declaration of Helsinki, Belmont Report, and Federal Policy for the Protection of Human Subjects

Belmont Report J RP B

Justice (equality): ensuring risks and benefits are distributed equally among participants 

Respect for persons: acknowledging autonomy and emphasizes informed consent  

Beneficence: maximizing benefits and minimizing risk across entire study design

Nuremberg code (1947):

Written after trials of physicians who abused prisoners. It stressed weighing risks against benefits and informed consent

10 ethical rules for testing medical treatments on human beings

Ethical conflict

Unavoidable in research; researchers must think through issues, minimize risks, and take responsibility for decisions

Confidentiality

Protecting participants' identifying information; a responsibility that continues after IRB approval.

Scholarly integrity

Honesty in research and in reporting it through the publication process.

Measurement

Assigning scores to individuals so the scores represent a characteristic of them

Construct

A variable that cannot be observed directly, such as intelligence, self-esteem, or depression (my example belonging)

Abstract

Summarize complex set of behaviors and internal process

Conceptual definition

A description in words of what a construct means.

Operational definition

The specific way a construct is measured. One construct can have several.

ex: depression is measured by Beck questionnaire

Converging operations

Using multiple operational definitions of the same construct. (ex: measuring fear and using self report, behavioral, and physiological)

Levels of measurement

Nominal, ordinal, interval, and ratio, each conveying more quantitative information than the last.

Nominal, ordinal, interval, and ratio

Nominal- categories with no order like neighborhood or major

Ordinal- ranked order WITHOUT equal distances like 1 to 5 rating (1,2,3rd place)

Interval- equal distances and NO true zero like temperature in Celsius (can be negative)

Ratio- equal distances and a true zero like count of interactions like age or attendance to class

level determines which statistics are defensible

Reliability (Consistency)

Consistency of a measure across time, items, and raters.

Belonging, depression are squishy/flexible bc they all feel different/change at different times 

Test-retest reliability

Consistency of scores when the same measure is given at two points in time.

Interrater reliability

consistent across observers. Use Cohen's kappa for categories and a correlation for scores

Internal consistency (type of reliability)

Consistency among the items within one measure.

Internal Validity

Our confidence (or people outside) that the treatment (experimental condition), and not something else, caused the change in the outcome.

It is strongest when the groups differ only in the treatment they received.

Threats to Internal Validity

• History: an outside event changes outcomes for everyone 

• Maturation: participants change over time on their own 

• Selection: the groups already differ before the even study begins

• Testing: taking the measure once changes how people answer it again later

• Attrition (Mortality): The loss or drop-out of study participants over time, especially if those who leave differ systematically from those who stay. 

• Instrumentation: Changes in measurement tools, devices, or human observer criteria over time.

Validity (Accuracy)

The extent to which scores actually represent the variable they are intended to measure.

1. Face, content, and criterion validity ask whether the measure looks right, covers the construct, and predicts a relevant outcome like asking questions about being sad for depression

2. Convergent validity: correlates w/ established measures of the same construct. 

3. Discriminant validity: does not correlate with unrelated constructs like happiness and depression are both high UH OH 

A measure can be reliable and still be invalid, because consistency is not accuracy. 

Experiment

An empirical study that manipulates an independent variable, measures a dependent variable, and controls extraneous variables.

Independent variable

The variable the researcher manipulates to create conditions.

Dependent variable

The variable the researcher measures to see whether it changes across conditions. (from IV)

Extraneous variable

Any variable other than the independent and dependent variables.

Confound

An extraneous variable that varies systematically with the independent variable.

Random assignment

Placing participants into conditions by chance, used in between-subjects experiments.

• It spreads traits such as motivation and prior ability evenly across groups. 

• It does not guarantee identical groups, but it removes the selection threat in expectation. 

• It helps control for and prevent selection bias if executed properly

• Minimize bias (best way is randomizing it) 

Counterbalancing

Varying the order of conditions across participants in within-subjects experiments.

Between vs Within subject experiments

Between-subjects experiments- each participant is assigned to only one condition or group

Within-subjects experiments- same participants are exposed to all conditions of the independent variable

Placebo control condition

Participants receive a treatment look-alike that lacks the active ingredient.

Internal validity

The extent to which the way a study was conducted supports a causal conclusion.

External validity

The extent to which results generalize to people and situations beyond the study.

A sample of 40 undergraduates in one lab may not describe a neighborhood, a workplace, or another country (Overgeneralization)

How representative is this study to others?

Lab studies offer tight control, so internal validity is high, but the setting can feel artificial. 

Field studies capture real settings, so external validity is often stronger. 

No single design maximizes both; adding control usually makes the setting less realistic

More control in lab = higher internal validity