Week 4 Flashcards
Effect Size
A quantitative description of the strength of a phenomenon expressed as a number on a scale
Why Researchers should Report on Effect Sizes
- Allows one to present the magnitude of an effect, enabling one to recognise the practical significance as well as the statistical significance
- Draw meta analysis conclusions by comparing standardised effect sizes across studies
- Previous effect sizes can be used when planning a new study in a priori power analysis
Unstandardised Effect Size
The effect size is expressed on the scale that the measure was collected on
Difference between P value and Effect Size
P Value; Used to make a claim about whether there is an effect
Effect Size; Used to decipher how large the effect is
Hypothetico Deductive Model
- Form a hypothesis
- Deduce predictions based on the hypothesis
- Test predictions through experiments/ observations
- Analyse the results
Standardised Effect Sizes
- Facilitates a comparison of effect sizes across situations where different measurement scales are used
Cohen’s D
-A type of standardised effect size
- Ranges from - infinity to + infinity with 0 indicating that there is no effect
D Family
- Consisting of standardised mean differences
R Family
- Consisting of measures of strength of association
- Describe the portion of variance that is explained by group membership
Point Estimates
- Tells us the effect size in our sample i.e. a single value
Confidence Intervals
Tells us the range of effect sizes that is 95% likely to contain the true population effect size
Facebook Experiment
- Carried out in 2014 to measure “emotional mood contagion” of users
- Concerns around ethics i.e. no informed consent and didn’t seek permission from IRB
- Manipulated the feeds of Facebook friends with some having increased positive content while other had increased negative consent
The Hungry Judges Study
- % of grated patrol sentences went from 65% on average to 0% at lunchtime or other emotionally strained times of day
Cohen’s D=1
- The standardised difference between 2 groups equals one standard deviation
Probability of Superiority
Expresses the probability that a randomly picked observation from one group will have a larger score than a randomly picked observation from the other group
Ds
The standardised mean difference between 2 groups of independent observations
Difference between t statistic and Cohen’s D
- The sample size for each group is needed for the T statistic formula but not for Cohen’s D
- T value is a function of the sample size but Cohen’s D is independent to it
True Effect
A non- zero effect size in the population
Common Language Effect Size
- The probability of obtaining a certain value on one variable, given the value of the other variable
Stroop Effect
The delay in reaction time between neutral and incongruent stimuli
Small Effect Size
d= 0.2
Medium Effect Size
d=0.5
Large Effect Size
d= 0.8
Registered Reports
Scientific publications which have been reviewed before the data has been collected based o the introduction method and proposed statistical analysis plan and published regardless of whether the results are statistically significant or not
