PSYC 300: Research Methods and Data Analysis in Psychology I
Oct 3- independent groups
● Correlational research- research designed to determine whether an association exists
between two variables
○ WATCH OUT!! Correlation does not equal causation
○ Statistical technique and research method
○
○ Positive correlation- as A increases, B increases (positive slope)
○ Negative correlation- as A increases, B decreases (and vice versa, negative slope)
○ Zero correlation- no relationship between A and B (even spread)
● Advantage of formal experiments over correlational designs
○ Randomized experiments are the “gold standard” design
○ Only difference between groups is what is the variable we hypothesize to be the
cause (X) of some outcome (Y)
○ Formal experiments can establish causality!
● Experiments to determine causality
○ -ice cream example
○ Independent groups design: Participants are non-overlapping between groups
○ Simplest independent groups design has two groups
○ Random-groups design: allows researchers to establish causality
● Types of independent groups/conditions
○ Experimental condition- intervention or treatment
○ Control condition- absence of intervention/treatment
○ Independent variables (IV)- manipulate, experimental and control conditions are
levels of the IV
○ Dependent variables (DV)- measure (no other difference)
● Random assignment and preventing confounding variables
○ Each participant has an equal chance of being assigned to a group.
○ Equates groups on potential confounding (or extraneous) variables.
○ Imperative that researchers ensure that their conditions are equivalent
○ Confounding variables- third variable that may differ between groups
■ problematic if they systematically covary with the IV and DV
■ Confounds decrease causality- the ability to make causal claims
, PSYC 300: Research Methods and Data Analysis in Psychology I
■ Differences in environment, in researcher, or in interest
■
● Sample size needed?
○ The number of participants you need depends on the effect size and desired power.
○ Power: The probability of detecting an effect if it actually exists
■ Underpowered studies may not replicate
■ Power analyses can be run prior to conducting an experiment
○ Representative sample!
■ Representative samples have characteristics representative of the broader
population
■ Important to make sure results generalize to other related groups of
individuals/the population
● Matched groups- why is this design necessary?
○ Groups are matched on some variable of importance (e.g., age, gender)
○ Ensure groups are equal to make causal claims
○ Increases internal validity
● Summary
○
Oct 5- independent groups, pt 2
● Internal vs external validity
○ Internal validity- experimental control, extent to which we can be sure the IV is the
cause of the DV
○ External validity- generalizability, extent to which we can be sure we can generalize
our results to different populations
● Threats to internal validity example
, PSYC 300: Research Methods and Data Analysis in Psychology I
○
○
-procedures, versions 1 2 and 3
● Golden rules for experiments-
○ 1. Random assignment to condition
Oct 3- independent groups
● Correlational research- research designed to determine whether an association exists
between two variables
○ WATCH OUT!! Correlation does not equal causation
○ Statistical technique and research method
○
○ Positive correlation- as A increases, B increases (positive slope)
○ Negative correlation- as A increases, B decreases (and vice versa, negative slope)
○ Zero correlation- no relationship between A and B (even spread)
● Advantage of formal experiments over correlational designs
○ Randomized experiments are the “gold standard” design
○ Only difference between groups is what is the variable we hypothesize to be the
cause (X) of some outcome (Y)
○ Formal experiments can establish causality!
● Experiments to determine causality
○ -ice cream example
○ Independent groups design: Participants are non-overlapping between groups
○ Simplest independent groups design has two groups
○ Random-groups design: allows researchers to establish causality
● Types of independent groups/conditions
○ Experimental condition- intervention or treatment
○ Control condition- absence of intervention/treatment
○ Independent variables (IV)- manipulate, experimental and control conditions are
levels of the IV
○ Dependent variables (DV)- measure (no other difference)
● Random assignment and preventing confounding variables
○ Each participant has an equal chance of being assigned to a group.
○ Equates groups on potential confounding (or extraneous) variables.
○ Imperative that researchers ensure that their conditions are equivalent
○ Confounding variables- third variable that may differ between groups
■ problematic if they systematically covary with the IV and DV
■ Confounds decrease causality- the ability to make causal claims
, PSYC 300: Research Methods and Data Analysis in Psychology I
■ Differences in environment, in researcher, or in interest
■
● Sample size needed?
○ The number of participants you need depends on the effect size and desired power.
○ Power: The probability of detecting an effect if it actually exists
■ Underpowered studies may not replicate
■ Power analyses can be run prior to conducting an experiment
○ Representative sample!
■ Representative samples have characteristics representative of the broader
population
■ Important to make sure results generalize to other related groups of
individuals/the population
● Matched groups- why is this design necessary?
○ Groups are matched on some variable of importance (e.g., age, gender)
○ Ensure groups are equal to make causal claims
○ Increases internal validity
● Summary
○
Oct 5- independent groups, pt 2
● Internal vs external validity
○ Internal validity- experimental control, extent to which we can be sure the IV is the
cause of the DV
○ External validity- generalizability, extent to which we can be sure we can generalize
our results to different populations
● Threats to internal validity example
, PSYC 300: Research Methods and Data Analysis in Psychology I
○
○
-procedures, versions 1 2 and 3
● Golden rules for experiments-
○ 1. Random assignment to condition
