research methods. research strategies descriptive—strategies for observing and describing behavior...

RESEARCH METHODS

Research Strategies

Research Strategies• Descriptive—strategies for observing and

describing behavior (answer who, what, when, where & how often)– Correlational methods– Naturalistic observation– Case studies– Surveys

• Experimental—strategies for inferring cause and effect relationships among variables

Descriptive Study• Describes a set of facts

• Does not predict what may influence the facts

• May or may not include numerical data

• Example: measure the percentage of new students from out-of-state each year since 1980

Case Study

• In depth study of one individual with the hopes of determining universal principles

• This technique is very open to bias

• Difficulty of applying data from one person to everyone

• Generally used to investigate rare, unusual, or extreme conditions

Survey Method

• Research method that relies on self-reports; uses surveys, questionnaires, interviews.

• Usually a very efficient and inexpensive method• Watch out for Framing/Wording Effects – The

way a question is worded can bias the answer.• Have to watch out what we take from polling

information: http://www.ncpp.org/node/4/#1• Also, people lie in surveys:

http://www.responsibilityproject.com/blog/post/lying-to-pollsters-bad-vote/

SamplingTerms• Population—large (potentially infinite) group represented

by the sample. Findings are generalized to this group.• Sample—selected segment of the population• Random selection—every member of larger group has

equal chance of being selected for the study sample

• Random Sample – Results from random selection, each member of the population had an equal chance of being included.– If a sample is not random it is said to be biased.

• Representative/Stratified sample—closely parallels the population on relevant characteristics

• Watch this short clip on the difference between these. (3 min)

DAILY

DOUBLE

QUESTIONExplain the difference between a

random sample and a representative sample.

ANSWER: Random sample—every member of larger

group has equal change of being selected for the study sample

Representative sample—closely parallels the population on relevant characteristics

Naturalistic Observation

• Method of observation where subjects are observed in their “natural” environment

• Subjects are not aware they are being watched

• Could use hidden cameras or two way mirrors

Meta-Analysis

• Combining the results from multiple different research studies.

• Example: Studying lung cancer from second hand smoke.

Bias

• Situation in which a factor unfairly increases the likelihood of a researcher reaching a particular conclusion

• Bias should be minimized as much as possible in research

Researcher Bias

• The tendency to notice evidence which supports one particular point of view or hypothesis

• Objectivity tends to reduce bias.

Participant Bias/ Hawthorne Effect

• Tendency of research subjects to respond in certain ways because they know they are being observed

• The subjects might try to behave in ways they believe the researcher wants them to behave

• Can be reduced by naturalistic observation

Critical Thinking

• Thinking that does not blindly accept arguments or conclusions but questions their validity

• Need to do this to avoid bias

Longitudinal and Cross-Sectional

Studies

Developmental Psychologists

• Psychologists who study how individuals change throughout their lifetime

Longitudinal Study

• Developmental study where researchers study the same group of individuals for many years

• Can be very expensive and difficult to conduct

Cross-Sectional Study

• Developmental study where researchers simultaneously study a number of subjects from different age groups and then compare the results

• Cheaper, easier than longitudinal studies, but group differences may be due to factors other than development.

Longitudinal/Cross Sectional Study

Correlational Study• Collects a set of facts organized into two or

more categories– measure parents’ disciplinary style

– measure children’s behavior

• Examine the relationship between categories

• Correlation reveals relationships among facts– e.g., more democratic parents have children who behave

better

Correlational Study

• Important NOT to imply a cause and effect relationship between the variables

• Correlational study does not determine why the two variables are related--just that they are related.

• Correlational studies are helpful in making predictions.

Correlational Study

• Correlation CANNOT prove causation– Do democratic parents produce better behaved children?

– Do better behaved children encourage parents to be democratic?

• May be an unmeasured common factor – e.g., good neighborhoods produce democratic adults and

well-behaved children

• Doesn’t mean they are not useful – Correlation in smoking (3 min)

Correlation & Causation

• There is a strong +.90 correlation in shoe size and IQ.

• Does this mean that a large shoe size is the cause for higher intelligence?

• What else could explain this?

•YOUR FEET GROW AS YOU GET OLDER & WISER

Illusory Correlation• A perception can be formed that there is a

relationship between events, actions and behaviors when, in fact, no relationship exists.

• Example: A football fan believes that every time he wears a specific jersey his team wins, so each time they play, he will only wear that jersey.

• Need More Examples?

Coefficient of Correlation

Numerical indication of magnitude and direction of the relationship between two variables

– Positive correlation—two variables vary systematically in the SAME direction

– Negative correlation—two variables vary systematically in OPPOSITE directions

How to Read a Correlation

Positive Correlation

• As the value of one variable increases (or decreases) so does the value of the other variable.

• A perfect positive correlation is +1.0.

• The closer the correlation is to +1.0, the stronger the relationship.

Negative Correlation

• As the value of one variable increases, the value of the other variable decreases.

• A perfect negative correlation is -1.0.

• The closer the correlation is to -1.0, the stronger the relationship.

Zero Correlation

• There is no relationship whatsoever between the two variables.

Let’s Review