Confounding Variables

confounding variable

-a type of extraneous variable that systematically affects one (or more) level(s) of the IV differently -differentially affects one level/multiple levels more than others -obscure different groups or make differences appear that are not really there

matched random assignment

-already have confounding variable in mind, collect info based on the confounding variable then pair participants up based on the info to match subjects (each pair has one participant in other group), based on number generations, same representation in each group, helpful with small samples (greater chance of being uneven) -variables should be held constant, reduces external validity -build second variable in the study

experimenter effect

-arise from the researchers knowledge of the hypothesis from the nature of the experimental and control conditions, and knowledge for which each participant is in, researchers suddenly/inadvertently affect the behaviors of participants -EX: behavior of researcher towards the experimental group is different compared to the control group

free random assignment

-assign participants in groups in a random matter, assignment of one participant doesn't affect any other participant Ex: random number generator, problem: unlikely will have even number of participants

random assignment

-assign to different variables -unlikely with one characteristic in one group, distribute confounding variables across conditions equally, only can control unknown factors

objective measures

-based on empirical, observable, and clearly specified events in which two or more people could easily agree, -precisely define behaviors, require minimal judgment calls, produce high rates of interrater reliability

Control for experimenter and participant effects

-blind studies, automate studies, objective measures

automate studies

-control experiment effects -ex: computer tasks: same procedures, randomly assign participants to conditions, leave experimenter blind to conditions, precise and unbiased for data collection, reduces contact with participants to reduce experimenter effects

cover story

-deception to control groups, also debrief

randomizing within blocks

-each participant given a number, then randomize list until everyone is assigned to a condition, ensure equal number in each condition

unbiased assignment

-experimental and control were equal

To prevent confounding variables--controls

-experimental controls, preparation of setting, participant selection and assignment

blind study

-experimenter, researcher, or both are unaware of the participants treatment

extraneous variable

-factor other than your treatment that affects the outcome of your study Ex: drawing attention to hair

ethical situations

-good thoughtful considerations for control groups

preparation of the setting

-labs: constant for each participant in each condition (reduce variables that could influence performance), artificial may reduce external validity

double blind

-no one is aware of the participants conditions -use assistant to make study blind (assistant not aware of the hypothesis or blind to the condition the participant is in)

participant effects

-participant behavior changes since they are in an experiment -hawthorne effect: differences in participants behavior that arise from knowledge that participants are being observed -demand characteristics: type of hawthorne effect in which participant pick up cues about the purpose of the experiment EX. may act "helpful" when understanding the purpose of the experiment, conform to researchers hypothesis

single blind

-participants or researcher don't know if participants are in the experimental or control group

experimental controls

-procedures used to counteract potential threats to the internal -validity of research

experimenter and partcipant effects

-should be blind during data collection and during scoring of data--experimenter -control conditions are believable (drugs and placebos)--participants

interactions between the experimenter and participants

-social interactions can affect the results -biosocial effects: biological characteristics of the researcher (age, sex, race, etc.) that affect behavior of the participant -psychosocial factors: psychological characteristics of the researcher that can affect the behavior of the participant

Types of confounding variables

experimenter effect, participant effects, interactions between the experimenter and participants