Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
57 Cards in this Set
- Front
- Back
|
4 ways ANOVA is better than a t-test
|
more powerful; can compare 2 or more groups
use multiple independent variables control for effects of other variables that might affect DV |
|
|
How is ANOVA used?
|
determines if there are mean differences between 2 or more experimental groups. An inferential procedure that applies sample data to population
|
|
|
3 main assumptions of ANOVA
|
normality of distribution
homogeneity of variance independence of observations (DV scores are independent of each other) |
|
|
how variance is analysed using ANOVA
|
compares amount of variation in DV due to IV to the amount due to random error or individual differences
|
|
|
Why use a Post-hoc test
|
ANOVA tells us if there's a diff, but not what is different or why. Post-hoc does this
|
|
|
When to use a post-hoc test
|
when ANOVA F-values are significant and there are more than two groups
|
|
|
Types of post-hoc tests
|
Scheffe, Newman Keuls, Tukey's HSD
|
|
|
how to analyze ANOVA in SPSS
|
GLM; it can handle unequal sample sizes, has common analytic framework
|
|
|
differences between:
Between Subjects ANOVA Within Subjects ANOVA |
within subjects ANOVA reduces error associated with individual differences, but can introduce carry-over effects
|
|
|
reasons to use within-subjects ANOVA
|
more likely to get results
|
|
|
How is variability partitioned differently in within subjects ANOVA versus between-subjects ANOVA
|
refers to conditions because there aren't distinct groups
no individual differences error Divides variability into individual diff.'s and experimental error, and only takes out experimental error |
|
|
how ANOVA is analyzed in SPSS
|
repeated measures GLM
|
|
|
When is factorial ANOVA used
|
allows you to analyze if IV's interact with each other. Can handle a large number of IV's
|
|
|
factor (in factorial ANOVA)
|
independent variable
|
|
|
3 possible ourcomes when examining interaction in Factorial ANOVA
|
both main effects and interaction effects found (looks like <)
no main effect but an interaction effect (looks like X) main effects but no interaction effect (looks like =) |
|
|
Why factorial ANOVA over simple ANOVA
|
more economical, provides control, and more generalizable to population of interest w/ better interaction with IVs
|
|
|
how to do factorial ANOVA on SPSS
|
GLM with multiples in "fixed factors" column
|
|
|
Why to use ANCOVA
|
equalize groups
examine covariates |
|
|
what is a covariate
|
variable related to DV but not directly addressed in IV
|
|
|
how to decide which covariates to use
|
use 2
correlate highest with DV and least with each other |
|
|
Why use ANCOVA
|
increases power of F test
removes predictable variance associated with covariates from error term |
|
|
SPSS procedure for ANCOVA
|
GLM with a covariate
|
|
|
4 sections of Results section
|
Analytic Strategy
Data Quality/MC Analysis of Main Hypothesis Analysis of Competing Hypothesis |
|
|
Analytic Strategy section
|
describe design and type of data
describe procedures used describe statistical analyses |
|
|
Data Quality/Manipulation Check
|
Show that experiment successfully set up conditions for testing hypothesis
Give means and SD's to show effectiveness of randomization and MC |
|
|
Analysis of Main Hypothesis (same structure for Analysis of Alternative Hypothesis)
|
Remind reader what you're studying and how
give main finding w/ numbers elaborate or qualify overall conclusion |
|
|
Tables and Figures
|
Any significant findings need tables/figures
Should stand alone |
|
|
4 sections of Discussion Section
|
Overall Conclusion, Implications, Limitations/Alternative Explanations, Directions for Future Research
|
|
|
Overall Conclusions
|
statement of support or nonsupport of hypotheses
main takehome message or implication of study |
|
|
Implications
|
Theoretical, practical, political implications of results
How results compare with previous research |
|
|
Limitations/Alternative Explanations
|
Char.'s of population sample
specific characteristics of method that may have influenced outcome Potential alternative explanations |
|
|
Future Directions
|
Implications for future research
summary statement of major findings |
|
|
Revising and editing paragraphs
|
Topic sentence content
underline claims for content check concluding sentences for content |
|
|
Revising and editing sentences
|
Use present tense
Avoid "significant" and synonyms for key terms eliminate extra adj. and adv. |
|
|
6 problems to avoid in discussion section
|
Failure to write with specifics
Claims without support Opening paragraph does not have the right info Failure to discuss confounds (alternative explanations) Improper discussion of future research Questionable conclusions |
|
|
why conduct experiments
|
examine effect of IV on DV when everything else is constant
|
|
|
role of inferential statistics
|
make inferences about population
bridge between sample and population |
|
|
major steps of hypothesis testing
|
state hypothesis
randomization manipulation compute test statistic use tables to find critical value make a decision |
|
|
important issues when collecting a sample
|
Representative Random Sample
Random Assignment |
|
|
Test Statistics
|
quantity calculated from sample data. Depends on assumed probability model and hypothesis under question (z,t,F, and X^2 are common)
|
|
|
type II error
|
Ho is false, but you don't reject it (occurs when experiment lacks sufficient power)
|
|
|
power
|
probability that you will correctly reject null hypo (Ho)
|
|
|
3 factors to increase power
|
increase sample size
increase reliability Increase alpha level |
|
|
what we learned from Francis Galton
|
don't rest on your laurels
take a forthright, non wishy-washy position don't be afraid to make mistakes never too late to apply first 3 lessons |
|
|
create a variable in SPSS
|
click "Variable" view tab
Select numeric or string Values (#'s represent categories) measures: data is ratio, interval, ordinal, or nomial |
|
|
compute a new variable in SPSS
|
transform
compute "target variable" box "numeric expression box" |
|
|
reverse score a variable in SPSS
|
under variable click "Change"
in "old value" box put original value in "new value box" put reversed score click "add" repeat for all possible combos |
|
|
how to check if data was entered correctly
|
look for out of range items, them for missing values
|
|
|
measures of central tendency
|
mean: interval or ratio (normally distributed)
median: ordinal (or inter. or rat. data NOT normally distributed) mode: nominal data |
|
|
measures of correlation
|
Kendall: nominal data
Spearman: ordinal (and inter. and rat. NOT normally distributed) Pearson: interval or ratio data (normally distributed) |
|
|
kurtosis
|
peakedness of the data
|
|
|
platykurtic distribution
|
flattened--less scores around mean and more around tails
|
|
|
leptokurtic
|
peaked--more scores around mean and less around tails
|
|
|
mesokurtic
|
normally distributed
|
|
|
SPSS to evaluate quality of DV
|
Frequency Procedure under descriptive statistics, do a histogram with a normal curve
|
|
|
APA table
|
usually has 4+ columns
no vertical lines |
|
|
what can we learn from Hughlings Jackson
|
"the study of causes of things must be preceded by the study of things caused"
reliability and validity measures of DV |
