Flashcards
STATS FINAL
Cards in this deck (17)
- 1Directional HypothesisError is on one side of the distribution
- 2Non-directional HypothesisError is split between both tails of the distribution
- 3Type 1 ErrorReject the null hypothesis as false when actually true
- 4Type 2 ErrorAccept the null hypothesis when it is actually false.
- 5Independent SampleUse Two-Sample Test
- 6Dependent SampleUse Matched-Pairs Test
- 7Two Sample Difference of Means PurposeTo compare two independent random sample means for difference
- 8Two Sample Difference of Proportions Test PurposeExamine proportion differences between two categories
- 9Wilcoxon Rank Sum TestUse with ranked data
To compare two independent rank sums - 10Mann-Whitney U TestSimilar to Wilcoxon
t-Test equivalent. - 11Matched-Pairs Difference TestUsed in the case of dependent samples
Each observation has two values - 12ScatterplotPredictor Variable: X axis
Criterion Variable: Y axis - 13Pearson's Product Moment Correlation Coefficient Requirements/AssumptionsNormally distributed
Interval/Ratio Scale
Linear Relationship b/t two variables - 14Correlation Coefficient Potential BiasRange Restrictions may cause best fit line to look different.
If variables are non-linear, attempts at a best fit line will fail
Outliers often reduce correlation - 15Correlation Coefficient Value Range-1 - 1
0 being completely unrelated
1 being positively correlated
-1 being negatively correlated - 16Spearman's Rank Correlation CoefficientUsed in cases where the data is:
Not normally distributed
Ordinal Level
Constantly rising of falling - 17Ecological Fallacy and Correlationthe correlation between individual variables is deduced from the group's correlation.
This is a statistical fallacy
Information
- Description
- Flashcards on STATS FINAL, created by James Foster9258 on 28/04/2015.
- Created by
- James Foster9258
- Views
- 2
- Rating