psych research exam 3

random selection

randomly select certain people to collect data from

random assignment

randomly assign group of people to a group

matching

using a technique to equate participants based on specific variables

constancy

holding variables constant

precision control

assisting people with same characteristics to different groups, one in treatment and one in control group

counterbalancing

used to control for sequencing effects

order effects

any differential effect resulting from the order of treatments/tasks

carry-over effects

performance in one condition is dependent on conditions before it

demand characteristics

cues that give away the hypothesis to the participants

double-blind model

experimenter and participant DO NOT know which condition the treatment is in

deception

provide participants with a cover story not related to the actual purpose of the experiment

control of recording errors

multiple data recorders

participants respond via electronic device

control of attribute error

use same experimenter in all treatment conditions

standardize experimenter look/demeanor

blind technique

correspond to experimenters half of double blind model

partial blind technique

experimenter kept blind until actual presentation of IV

automation

remove experimenter from data collection process

chi-square

categorical data

frequencies/counts

one or more IVs

power

likelihood that a statistical test will be significant

pilot study

trial run to determine if everything for a research procedure is working properly

informed consent form

given before a person participates in research so they know where their information is going and what the study will consist of

debriefing

happens after research to tell what the study is for

especially important when there is deception so they know what was deceiving and are fully informed on what happened

parametric tests

estimating specific parameters

assumptions about population distribution

non-parametric tests

no assumptions about parameters

used for ordinal or nominal data

distribution free tests

randomization tests

determine if something is significant because if you shuffle treatment conditions things should not change

compare orignal test to new distribution