ch 11: experiments and internal variables

one-group pretest/posttest design

problematic experiment design where 1 group of participants is tested before and after an intervention w no comparison group

maturation threat

in 1g pre-t/post-t design. occurs when an observed change in experimental group could have emerged spontaneously over time

history threat

in 1g pre-t/post-t design. occurs when something specific + systematic occurs between pretest and posttest

regression threat/reg towards the mean

in 1g pre-t/post-t design. phenomenon in which any extreme finding is likely to be closer to its typical the next time it is measured

attrition threat

in 1g pre-t/post-t design. reduction of participants occurs when people drop out. problematic when systematic (saddest people drop out etc)

testing threat

in 1g pre-t/post-t design. type of order effect. when participants change as a result of testing more than once (including practice effect)

instrumentation threat

in 1g pre-t/post-t design. occurs when measuring instrument changes over time (eg changing standards of judgement, different forms of pre/posttest)

selection-history threat

combined threat where an outside event/factor affects only those at one level of an experimental condition

selection attrition threat

combined threat where only one of the experimental groups experiences dropping out

observer bias

when a researcher’s expectations influence their own interpretation of the results

demand characteristics

when participants guess the purpose of the study and change behaviour in expected direction

placebo effects

when participants who receive treatment really do improve but only because they believe they are receiving valid treatment

null effect

when the IV makes no difference/study finds no statistically significant difference

ceiling effect

possible cause for null effect. not enough variability between levels due to all scored being clustered together at the high end of measurement

floor effect

possible cause for null effect. not enough variability between levels due to all scored being clustered together at the low end of measurement

manipulation check

separate DV measure included to ensure operationalization (manipulation) was effective. controls for ceiling and floor effects

noise

possible cause for null effect. too much unsystematic variability within levels obscures true relationship. can be caused by situation noise, individual differences or measurement error

measurement error

human instrument or factor that can randomly inflate/deflate someone’s DV score. always there but researchers try to minimize

situation noise

external distractions in the environment of testing