week 7

Cards (14)

  • what happens to a t distribution if degrees of freedom are high enough
    matches the normal distribution
  • in t tests degrees of freedom are equal to
    n - number of groups
  • when do we use one-tailed tests
    when previous research suggests a direction the data may head in
  • what is the p-value often used in psychology
    0.05
  • What does Neyman-Pearson say
    • often when looking at the null hypothesis people fail to look at the long-term effects, this is where type 1 and type 11 errors are seen
    • incorrectly accepting the null causes a type 1 error
    • incorrectly rejecting the null causes a type 11 error
    • P(Type 1 error) = alpha
    • P(Type 11 error) = beta
    • if p = 0.05, alpha = 5%
    • standard value of acceptance level of beta is 0.2 meaning that 20% of time we accept H0 incorectly
  • What is a within subjects experiment
    repeated measures design
  • evaluate a within subject design
    • fewer ppts
    • eliminates individual differences
    • time-related threats may interfere
    • order effects may interfere
    • participant attrition
  • what are the 5 types of time-related threats
    • history - personal changes during study
    • maturation - psychological or physical effects
    • instrumentation - instruments used to measure DV
    • testing effects - effects experienced as a result of prior condition (order effects)
    • statistical regression - extreme scores may become less extreme due to statistical regression
  • what is the carryover effect
    • when an experimental condition may have a lasting effect on ppts going through it
  • what is contrast effect
    • when subjective perceptions of experimental conditions may contrast to that of the last one causing an effect on the results
  • what are examples of progressive error
    • practice effect - ppts become better at the condition because they practiced
    • fatigue - become worse at the condition
  • How can we control for the effects relating to a within subjects experiment
    • controlling time - ensuring the times between each condition are not so long that time-related threats interfere
    • switching to between subject design
    • counterbalancing
  • evaluate counterbalancing
    • causes order effects in some ppts but not all so increases variance in the study
    • asymmetrical order effect
    • number of conditions - may not be possible due to number of conditions in the study
  • why are two-treatment designs less convincing than multiple-treatment designs
    are not as good at establishing a cause-and-effect relationship