lesson 10 - experiments cont'

demand characteristics

participants change their behaviour based on what they think the hypothesis is

placebo effect

participants improve only because they believe they’re getting treatment

observer bias

the researcher’s expectations interfere with their observations/results

maturation

people naturally change over time

history

something external happens during the study that affects the results

regression to the mean

extreme scores move towards average over time

testing

taking the same test more than once affects performance

instrumentation

measurement changes over time

attrition

participants leave the study

comparison groups

controls for:

• maturation

• regression

• history

• testing

both groups experience those things → allows you to see if change is due to IV or confounds

placebo control groups

control group gets a fake treatment → controls for placebo effect

double blind study

participants dont know their condition—nor do the researchers

controls for:

• demand characteristic

• placebo effects

• observer bias

null effect

the study finds no difference between groups (IV didnt affect DV)

→ BUT does not mean the IV has no effect in reality

3 possible explanations for a null effect

1. not enough difference between groups results → IV didn’t create a strong enough change

2. too much variance within groups → people within each group are very different from each other

3. true null effect → the IV doesn’t affect the DV at all

systematic variability

differences in the DV that are consistently associated with the IV

• comes directly from the IV

  • good because it reflects the true effects of your IV—the “signal” you’re trying to detect

unsystematic variability

differences in the DV that are random or unrelated to the IV

• comes from individual differences, measurement errors, situational factors

  • bad for detecting effects—obscures the IVs affect

effect size

a measure of how strong the effects fo the IV is on the DV

• high systematic variability → larger effect size

• high unsystematic variability → smaller effect size

statistical significance

a measure of how likely the observed effect was due to chance

• high systematic variability → easier to reach significance

• high unsystematic variability → harder to reach significance

reasons for inadequate difference between group’s results

• weak manipulation of the IV

• ceiling or floor effects → DV scale limits the ability to see differences

• insensitive measure → the DV isnt precise enough to detect changes

reasons for large within groups variance

• individual differences

• measurement error

• uncontrolled extraneous variable(s)