psych analysis chapter 4 (10) - experimental design

extraneous variable

any variable other than the independent variable that could affect the dependent variable in a study.

experimental research

a type of research where the researcher manipulates an independent variable to see whether it causes a change in a dependent variable.

establishing if one variable actually causes another

1. temporal precedence/sequencing - (independent variable must happen BEFORE dependent variable)

2. covariance - two variables are related—they change or vary together.

3. rule out extraneous variables

temporal precedence

independent variable happens BEFORE dependent variable

covariance

two variables vary together

nuisance variable (extraneous variable)

a variable that can affect the dependent variable but is not the main variable the researcher is interested in.

• unsystematic error - random variation in your results caused by nuisance variables that affect participants differently and unpredictably.

• random occurrences - unexpected events or conditions that affect participants by chance and can influence the results of a study.

confound variable (extraneous variable)

a variable that changes along with the independent variable and provides an alternative explanation for the results.

• systematic error - occurs when an unwanted factor affects the results in a consistent, predictable way rather than randomly.

• selection effects - occurs when people select themselves into different groups, causing the groups to differ in important ways before the experiment starts.

• IV also manipulates alternative explanations (design

  confounds) - the researcher accidentally changes more than just the independent variable (IV) between the groups.

experimental design

is the specific blueprint or plan for how you will run that test, including how you group people, assign tasks, and measure results

minimizing extraneous variables

best - design the study to rule out confounds

• thoughtful study design

• standardization/holdiong extraneous variables constant

• repeat testing

• randomization

• matching

good - measure and test confounds

• testing effects of confounds as another independent variable

• measuring potential confounds to control statistically

between subjects design/independent groups design

• an experimental setup where each participant is assigned to only one condition or treatment group

• different participants in each IV group

• comparisons through groups rather than the same people doing different things

• why do we like this design?

  • time, feasibility

  • limits validity threats stemming from repeated measurement

  • 
    

validity threats to subject design

• selection effects

  • how do we decide who goes in what group

• minimizing threats to validity

  • matching

  • randomization

    • good idea to check group differences

• attrition/mortality

• order effects (carryover)

  • learning

  • fatigue

  • contrast

  • habituation

  • sensitization

• demand characteristics

attrition/mortality

• attrition - participant dropout over time

• mortality - participant died

minimizing threats to validity

• attrition/mortality - keep it short and simple, incentives

• demand characteristics - mask purpose/hypothesis

• minimize carryover

  • learning - pretraining and practice sessions, space out conditions

  • contrast - space out conditions

  • habituation, sensitization, fatigue - rest periods, keep it short and simple

  • counterbalancing

experimenter bias

occurs when a researcher’s expectations of beliefs influence the participants, procedures, or interpretation of results

• expectancy effects - researcher’s expectations about what should happen influence participants or the study results

• treating groups differently

expectancy effects (experimenter bias)

occurs when a researcher’s expectations of beliefs

solutions to experimenter bias

• single blind - the participants don’t know which condition they’re in, but the researcher usually does

• double blind - both participant and researchers don’t know

single blind

the participants don’t know which condition they’re in, but the researcher usually does

double blind

both participants and researchers don’t know