IA Flashcards
In order for results to be significant, the p-value must be:
Less than 0.05
Nominal data
- data can only be categorized
Examples: gender, city of birth
Ordinal
- data can be categorized and ranked
Examples: likert-type questions, language ability (beginner, intermediate, fluent)
Interval data
- data can be categorized, ranked and evenly spaced
(No natural zero: when there is an absolute absence of what is being measured)
Examples: temperature, test score
Ratio data
- data can be categorized, ranked, evenly spaced and has a natural zero
(Natural zero: when the data is 0, it means there is an absence, for example, 0 years old = absence of age)
Examples: age, height
Levels of measurement
- nominal
- ordinal
- interval
- ratio
How do you determine the proper descriptive statistics for the data?
(Mean median or mode)
Nominal: mode, variation ratio
Ordinal: median, IQR
Interval: median/mean, Range/IQR/Stdev/Variance??
Ratio: mean, unless there are outliers, Stdev
Inferential statistics
Normally distributed (>30)
- Independent samples t-test
Data not normally distributed (<30)
- Independent samples: Mann Whitney U test
- repeated measures: Wilcoxon Signed-rank test
How to analyze Mann-Whitney test statistic:
- If p value < 0.05, results are significant. If otherwise, (p-value > 0.05,) the results are not significant
- U statistic “reflects the difference between the two rank totals”, the smaller it is, the less likely the results are to have occurred by change
How to analyze Wilcoxon Signed-rank test:
- If p value < 0.05, results are significant. If otherwise, (p-value > 0.05,) the results are not significant
How to calculate variation ratio:
V = 1 - (fm/N)
fm is the frequency (number of cases) of the mode, and N is the total number of cases.
Calculate if there are outliers
Large outliers = Q3 + (1.5 * IQR)
Small outliers = Q1 – (1.5 * IQR)
IQR = Q3-Q1