BM Ch11

• one-group, pretest-posttest design


  one group, measure before and after exposure to IV, no comparison group
• threats to validity in one-group pretest/posttests

  maturation, history, regression, attrition, testing, instrumentation, combined
• maturation threats


  getting used to the conditions, a change in behaviour that emerges almost spontaneously over time
• example of maturation threats

  spontaneous remission
• including a comparison group can prevent...

  maturation, history, regression, testing threats
• history threats


  something else has happened between the two tests
• regression threats


  regressing to the mean - when a score is particularly high/low, the next score is likely to be closer to the mean, becomes a threat if a group is chosen due to particularly high/low prestest
• attrition(/mortality) threats


  a reduction in participants when they drop out of the study
• attrition is a threat when...

  the loss is systematic (one type of participant)
• preventing attrition threats


  remove their pretest score, or check it to identify if it is particularly high or low
• testing threats


  specific type of order effect, a change in the participant due to taking the test multiple times
• testing threats specific examples

  practice/fatigue effects
• preventing testing threats


  switch to posttest-only, use alternative tests for each
• instrumentation threats


  when a measuring instrument changes over time, or the tests for before and after are different but not sufficiently equivalent
• preventing instrumentation threats


  switch to posttest-only, ensure tests are equivalent, counterbalance test types
• combined threats: selection-history threat

  outside factor affects those at one level of the IV
• combined threats: selection-attrition threat

  one experimental group experiences attrition
• threats to internal validity (all studies)

  observer bias, demand characteristics, placebo effect
• observer bias


  when researchers' expectations influence their interpretation of the results
• observer bias threatens...


  internal and construct validity
• demand characteristics


  cues in an experiment that tell the participant what behaviour is expected, therefore changing their behaviour
• double-blind studies


  neither participant nor researcher knows which group a participant is in
• masked design (blind design)

  participant knows what group they are in, the observer does not
• placebo effect


  when the participant receives a treatment and does improve, but only because they believe their treatment is valid
• placebo effect may include...

  placebo side effects
• double-blind placebo control study

  a study that uses a treatment group and a placebo group and in which neither the research staff nor the participants know who is in which group
• double-blind placebo control studies may also include...

  no treatment groups
• null effects/results


  no significant covariance in IV and DV (very common)
• cause of null effects

  incorrect theory (IV does not affect DV), study poorly designed
• not enough between-groups difference

  weak manipulations, insensitive measures, ceiling and floor effects
• weak manipulations


  small, negligible differences between the levels of the IV
• insensitive measures


  operationalisation not sensitive enough
• ceiling effects


  all scores squeezed together at high end
• floor effects


  all scores squeezed together at the low end
• ceiling and floor effects are special cases of...

  weak manipulations and insensitive measures
• manipulation checks used to detect...

  weak manipulations, ceilings and floors
• design confounds in reverse


  reverse or counteract the true result
• too much within-groups variability

  measurement error, individual differences, situation noise, problem with statistical validity - disrupts detection of a true difference
• problem with statistical validity

  greater overlap = smaller effect size = less likely to be significant
• measurement error (human or instrument)

  lots of measurement error = greater spread