Module 2 Research Methods

Scientific method

A cycle in which a theory generates a testable hypothesis, the hypothesis is tested with empirical observation or experiments, and the results support or modify the theory

Steps of the scientific method

(1) Ask a question/observe, (2) form a hypothesis, (3) design a study and collect data, (4) analyze data and draw conclusions, (5) report results so others can replicate them and refine the theory

Deductive reasoning

Starts with a general idea (hypothesis) and tests it against real-world observations; emphasized in experiments

Inductive reasoning

Uses specific observations to build broad generalizations or theories; emphasized in case studies

Variable

Anything that can change or vary between people or conditions

Independent variable (IV)

The variable the experimenter manipulates or controls; the presumed cause

Dependent variable (DV)

The variable the researcher measures to see the effect of the independent variable; the presumed effect

Example of IV and DV from the textbook

IV = type of algebra instruction (computer program vs. in-person teacher); DV = learning (test score)

Theory

A well-developed set of ideas that proposes an explanation for observed phenomena and is consistently supported by evidence

Hypothesis

A testable prediction about how the world will behave if an idea is correct, often written as an if-then statement

Where hypotheses come from

Theories (deductive reasoning), direct observation of the world (inductive reasoning), or review of previous research

James-Lange theory of emotion

Theory that emotion comes from physiological arousal; example of a falsifiable theory

Falsifiable

Capable of being shown wrong by evidence

Why hypotheses must be falsifiable

If no observation could disprove an idea, it can't be scientifically tested; surviving attempts to disprove it gives us confidence in it

Example of a non-falsifiable idea

Freud's id, ego, and superego

Operationalize

To define exactly how a variable will be measured or manipulated

Operational definition

A precise description of how variables will be measured or manipulated (e.g., "learning" = score on a test of the material)

Why operational definitions matter

They let others understand exactly what was measured and replicate the study

Descriptive research designs

Methods that describe behavior without manipulating variables (case studies, naturalistic observation, surveys, archival research); correlational in nature, so they can't show cause and effect

Case study (clinical study)

An in-depth study of one or a few individuals, often with a rare condition

Strength of case studies

Unmatched depth and richness of information

Weakness of case studies

Hard to generalize because the cases are unusual

Generalize

Apply findings from a sample to the larger population

Naturalistic observation

Observing behavior in its natural setting without interfering (e.g., Jane Goodall's chimpanzees)

Strength of naturalistic observation

High ecological validity (realism); people behave naturally

Weaknesses of naturalistic observation

Little control, costs time and money, depends on luck, and risks observer bias

Structured observation

Observing people as they do set, specific tasks (e.g., Ainsworth's Strange Situation)

Observer bias

When observers' expectations skew what they record

Inter-rater reliability

How consistently different observers agree on what they observed

Survey

A list of questions answered by participants (paper, online, or verbal)

Strengths of surveys

Quick and cheap, allow large samples, more generalizable

Weaknesses of surveys

Less depth of information; self-report problems (people lie, misremember, or answer to look good)

Archival research

Using existing records or data sets to answer research questions

Strengths of archival research

Inexpensive, fast, no interaction with participants

Weaknesses of archival research

No control over what data was originally collected; records may be inconsistent

Correlational study

Research that measures two or more variables (without manipulating them) to see whether they are related

Correlation

A relationship between two or more variables; as one changes, so does the other

What correlational studies tell us

The strength and direction of a relationship, which allows prediction

Strengths of correlational studies

Can study things that can't ethically be manipulated, often real-world, find relationships to test later, allow prediction

Weaknesses of correlational studies

A third variable might cause both; can't tell which variable affects which; people wrongly assume causation

What correlational studies can never tell us

Cause and effect (correlation does not equal causation)

Why correlational studies are useful

Prediction (e.g., SAT predicting GPA), finding real relationships, studying questions experiments can't ethically address (e.g., smoking and cancer)

Longitudinal research

Testing the same group of people repeatedly over a long period of time

Strength of longitudinal research

Shows real change within individuals without cohort differences

Weaknesses of longitudinal research

Takes a lot of time and money; high attrition

Attrition

Loss of participants who drop out of a study over time

Cross-sectional research

Comparing different age groups (segments of the population) at one point in time

Strength of cross-sectional research

Faster and cheaper than longitudinal research

Weakness of cross-sectional research

Cohort effects, meaning differences may reflect the generation people grew up in rather than age

Cohort

A group of people born around the same time who share social and cultural experiences

Illusory correlation

Believing two things are related when no relationship actually exists (e.g., full moon and strange behavior)

Confirmation bias

Tendency to look for evidence that supports what we already believe and ignore evidence that contradicts it

Effect of illusory correlations on society

They can contribute to stereotypes, prejudice, and discrimination

Third variable problem

Two variables seem related only because a third, outside variable affects both

Example of the third variable problem

Ice cream sales and crime rates rise together because hot temperature causes both

Experiment

The only research method that can establish cause and effect; the researcher manipulates an IV and measures its effect on a DV

What is needed for a study to be an experiment

Manipulation of an IV, measurement of a DV, an experimental and a control group, and random assignment to groups

Strengths of experiments

Can show cause and effect; high degree of control

Weaknesses of experiments

Often artificial settings, ethical limits, and some variables can't be manipulated

Quasi-experiment

A study where the IV can't be manipulated or randomly assigned (e.g., sex), so cause-and-effect claims can't be made

Population

The whole group of people researchers are interested in

Sample

A subset of individuals selected from the population

Random sample

A sample in which every member of the population has an equal chance of being selected

Why random samples are preferred

A large enough random sample is representative of the population, so results can be generalized

Problem with using college students as participants

They tend to be younger, more educated, and less diverse, so results are hard to generalize (a convenience sample)

Experimental group

The group that receives the experimental manipulation (the treatment being tested)

Control group

The group treated the same way but without the manipulation; serves as a basis for comparison

Random assignment

Every participant has an equal chance of being placed in either the experimental or control group

Why random assignment matters

Makes it unlikely the groups differ before the study, so differences afterward can be attributed to the IV

Confound (confounding variable)

An unanticipated outside factor that affects the variables of interest and makes it look like one caused the other

How to control confounds

Random assignment and treating all groups identically except for the IV

Expectancy effects

When someone's expectations change the results of a study

Experimenter bias

When the researcher's expectations skew the results

Placebo effect

When people's expectations or beliefs alone change their experience (e.g., feeling better after a sugar pill)

Placebo

An inactive treatment (like a sugar pill) given to the control group

Single-blind study

Participants don't know which group they're in, but the researcher does

Double-blind study

Neither the participants nor the researchers know who is in which group

Safeguards against expectancy effects

Single-blind and double-blind designs, placebo control groups, blind scorers, and clear criteria for recording behavior

Descriptive statistics

Numbers that summarize and describe a set of data (typical scores and how spread out they are)

Measures of central tendency

Statistics that describe the typical score in a data set (mean, median, mode)

Mean

The arithmetic average; very sensitive to outliers

Median

The middle score when data are in order

Mode

The most frequently occurring score

Outlier

An extreme score far from the rest that can pull the mean

Measures of variability

Statistics that describe how spread out scores are (range, standard deviation)

Range

The highest score minus the lowest score

Standard deviation

The typical distance of scores from the mean; small = scores tightly clustered, large = spread out

Positive correlation

Both variables move in the same direction (e.g., height and weight)

Negative correlation

The variables move in opposite directions (e.g., hours of sleep and tiredness)

No correlation

The variables are unrelated (correlation coefficient of 0)

Correlation coefficient (r)

A number from -1 to +1 showing the strength and direction of a relationship

What the sign of a correlation coefficient tells you

The direction (positive or negative)

What the size of a correlation coefficient tells you

The strength; the closer to +1 or -1, the stronger

Scatterplot

A graph showing the relationship between two variables; points closer to a line mean a stronger correlation

Reliability

Consistency, meaning a measure produces the same result again and again

Internal consistency

How well items on a survey that measure the same thing correlate with each other

Test-retest reliability

How consistent results are when the same measure is given multiple times

Validity

Accuracy, meaning a measure actually measures what it is supposed to

Ecological validity

How well research results apply to the real world

Construct validity

How well a variable captures what it is intended to measure