Intro to Simple experiments

causal claim

boldest type of claim. replaces association verbs with causal ones

experiment

researchers manipulate at least one variable and measure another

manipulated variable

aka independent v. variable whose levels are managed by researcher

conditions

levels of an independent variable

measured variable

aka dependent. variable whose levels are observed and recorded

control

any variable the experimenter holds constant across conditions on purpose

control group

participants exposed to “no treatment” level of independent variable

treatment group

participants exposed to intervention level of independent variable

placebo group

participants exposed to an inert treatment

confounds

accidentally varies systematically with changes in IV. possible alternate explanations for changes in dependent variable.

selection effect

when the participants in one level of IV are systematically different than those at another level

matched groups design

controls for selection effect by placing participants with a similar possible confound into matched sets across experimental conditions so they cancel each other out

independent-groups design

separate groups are put into different levels of IV

posttest only design

type of independent-groups design where participants are only tested once on DV after experimentation

pretest/posttest design

type of independent-groups design where participants are tested twice on DV once before and once after IV exposure

within-groups designs

each participant is presented with all levels of IV

repeated-measures design

type of within-groups design where participants are measured on DV after each exposure to each level of IV

concurrent-measures design

type of within-groups design where participants are exposed to all levels of IV at once and then measured just once

order effects

disadvantages of within-group designs when being exposed to first condition changes how participants react to later condition(s) and is a threat to internal validity

practice effects

repeat testing causes participants to get better at a task

fatigue effects

repeat testing causes participants to get bored/tired of a task

carryover effects

contamination carries over from one test to the next

counterbalancing

different participants are given the dif levels of IV in different orders so order effects balance out when all data is combined

demand characteristics

a cue that can lead participants to guess the experiment’s hypothesis

observer bias

when observer expectations influence their interpretation of participant behaviours or outcomes of a study

observer effects

when the observer being there effects the participants performance

expectancy effects

maturation threat

when a change in experimental group could have emerged spontaneously over time

history threat

when something else specific+ systematic happens between pretest and posttest

regression threat

phenomenon in which any extreme finding is likely to be closer to its own typical, or mean level the next time it is measured (w or without intervention)

attrition threat

reduction of participants occurs when people drop out. Problematic when it is systematic (eg most rambunctious camper leaves early, saddest people drop out of therapy)

prevention of attrition threat

when participant drops out you remove their score from pretest too

prevention of history/regression/maturation threat

comparison group

testing threat

specific kind of order effect. change in participants as a result of taking a dependent measure more than once. includes practice effect.

instrumentation/item effects

occurs when measuring instrument changes over time. (eg changing standards of judgement, different forms of pretest and posttest)

selection-history threat

outside event/factor affects only those at one level of experiment condition

selection-attrition threat

only one of the experimental groups experiences attrition

placebo effects

when people in an experimental treatment experience a change only because they believe in the treatment

placebo-controlled double blind study

study that uses a treatment and placebo group but neither researchers nor participants know which is which

cohen’s D

index of how much 2 sample means differ. expressed in SD units

statistical power

ability to detect an effect if there is one. probability of rejecting the null. low power = less reliable study

manipulation check

separate dependant variable that experimenters include in a study to check if operationalization worked

ceiling effect

obscuring factor to explain null result. all scores clustered together at the high end leading to not enough variability between levels

floor effect

obscuring factor to explain null result. all scores squeezed together at the low end. leads to not much variability between levels

situation noise

obscuring factor to explain null result. external distractions. leads to too much variability within levels

measurement error

obscuring factor to explain null result. human or instrument factor randomly changes result