module 4 Flashcards
(47 cards)
1
Q
ANOVA
A
- analysis of variance
- Between subjects one way analysis of variance
- used when IVs have more than 2 levels
- compares means from 3+ independent samples
- testing mean differences but basing the test around variance between means
2
Q
factor
A
- IV in ANOVA
or - in a description of an experimental design
3
Q
levels of a factor
A
- individual conditions comprising a factor
4
Q
single factor experiment
A
- experiment with one IV
5
Q
one way ANOVA
A
- ANOVA with a single factor
- aka single factor ANOVA
6
Q
two factor experiment
A
- experiment w two IVs
7
Q
Two way ANOVA
A
- ANOVA with two factors
- aka two factor ANOVA
8
Q
t or f: a one way ANOVA tests two or more differences among means
A
false, its determining if there is at least one mean difference among levels of the IV
9
Q
ANOVA is used to test mean differences, but its calculations are based on _____
A
- variances (size of difference among scores)
- also it explains sources of the variability
10
Q
total variabilities can be divided into ______ and ____ treatments
A
- between treatment conditions variability: some variability is possibly due to differences between conditions
- within: some variability exist within each condition
11
Q
reasons for between treatment variability
A
- treatment effect: manipulation distinguishing between conditions could influence scores
- individual differences: differences in backgrounds/abilities/attributes/circumstances of individual ppl
- experimental error: chance errors that occur when measuring construct of interest, researchers try to minimize it
12
Q
reasons for within treatment variability (drawing two scores from the same condition)
A
- individual differences: differences in backgrounds/abilities/attributes/circumstances of individuals
- experimental error: chance errors that occur when measuring construct of interest
13
Q
what test statistic associated w an ANOVA
A
- F-ratio statistic (f-test)
- give estimate and their ratio
14
Q
f test conceptual formula
A
f=variance between/variance within treatments
or
f=signal/noise
15
Q
f test as a source of variance
A
f= treatment effect + individual diff + experimental error/ individual diff + experimental error
16
Q
denominator of the f test
A
- measures uncontrolled and unexplained variability in scores
- often called error term
17
Q
numerator in f test
A
- measures same variability as denominator but also variability arising from systemic influences (treatment/condition effect)
18
Q
symbols used in ANOVA calculations
A
- k: number of levels in a factor
- n: sample size for specific condition
- N: sample size for entire study
- T: sum of scores w/in a specific condition
- G: sum of all scores in experiment
- SS: sum of squares
19
Q
S^2= _____
A
SS/n-1
20
Q
∑
A
summation (addition of a serious of numbers)
21
Q
examples of effect sizes
A
- cohen’s d: effect size, difference between two means/S (standard deviation)
- pearson’s r correlation coefficient: effect size that also shows the strength and direction between 2 cont. variables. ranges from -1 to 0 to 1
- odds ratio: ratio that determines the odds of an event occurring
22
Q
bayesian statistics
A
- uses newly collected data to update a hypothesis
- looks at the probability of the hypothesis being true based on the data collected
23
Q
bayes’ theorem
A
- conditional probability of two events can be obtained from their individual probabilities and the inverse conditional probability
= p(a/b)=p(a/b)x p(a)/p(b)
24
Q
bayes factor
A
degree of which beliefs changed due to data
25
NHST vs bayesian
- NHST more limited, accepts Ha without seeing probability of H0
- Bayesian: determines probability of both H0 and Ha
26
posterior odds
if Ha is probable based on data relative to H0
27
signal to noise ratios
- aka variance explained by model and variances that model cannot explain
- aka effect to error
28
__________ is a range of values expected to include population value with a degree of confidence
confidence interval
29
computing total variability
SS total= X^2 - G^2/N
30
what two parts are involved in analysis of variability
- analysis of SS
- analysis of df
31
SS total = SS within + ____
SS between
32
SSwithin= ____
∑ SS each condition
33
SS between= ____
∑ T^2/n - G^2/N
34
df total=
N-1
35
df within =
∑ (n-1)
or
N-k
36
df between =
k-1
37
in ANOVA, the term for variance is _______
mean square (MS)
38
MS =
SS/df
39
f ratio formula
MS between/MS within
40
omnibus test
- tells you at least one difference exists but does not say the number of differences or where they occur
41
what are the two general approaches to a follow up test for a ANOVA
- posteriori: post hoc test, not based on planning/clear hypothesis
- priori: planned/theoretically driven follow up test
42
posteriori test
- only appropriate when F test is significant (5% or more)
- has no bias and explores all pairwise comparisons
- attempts to control familywise error
43
familywise error
type 1 error rate across tests conducted on same data
44
stricter familywise control comes with less _____
power
45
common post hoc tests
- Least significant difference (LSD
- Bonferroni Adjustment: redistributes alpha to keep it at 5% overall
- Tukey Honesty Significant Difference (HSD): most widely used
- REGWQ
46
orthogonal vs non orthogonal
orthogonal: slices overlap/share a common data set
nonorthogonal: variables are separate and independent of one another/slices do not overlap
47
if the sum of all products of contrast weight are equal to 0, that means it is _______
orthogonal (has overlap)
