Karteikarten
BE428 Lecture 7: Parametric vs Non-Parametric Statistics
Karten in diesem Stapel (12)
- 1what are the three ways to determine normality?1. sketch the data/produce a histogram
2. make a normal probability plot to determine if your data is normally distributed
3. carry out a shapiro-wilk test - 2how can a normal probability plot be used to determine normality?- straight line represents the normal distribution
- a straight trend in the data is observed for a normal distribution
- curved trend may indicate non-normality - 3how can a shapiro-wilk test be used to determine normality?p-value less than alpha means the sample is not normally distributed
- 4what are non-parametric methods?those data analysis techniques that don't require the data to have a specific distribution
- 5what can non-parametric methods require in some cases?1. the data have a symmetric distribution
2. data from two populations must have the same distributions - 6what is the definition of sign?positive, negative or zero
- 7what is the definition of rank?- assignment of ranks to data
- might not be interested in absolute difference from a specified value, but just that a number is different from the value
- can be combined with signs - 8why are medians used?many non-parametric tests examine the relationship between medians
less influenced by outliers
measures centre of distribution - 9give four examples of equivalent parametric and non-parametric testsindependent sample t test = mann-whitney test
paired samples t test = wilcoxon signed rank test
one way ANOVA = kruskal wallis test
one way repeated measures ANOVA = friedman's ANOVA - 10what is a Mann-Whitney test?- compares two independent samples
- does this by rank
- calculates the u statistic
- look up in tables (Us < Uc = rejection of Ho) - 11what is a Wilcoxon test?- compares paired samples
- does this by rank and sign
- calculates the W statistic
- look up in tables (Ws < Wc = rejection of Ho) - 12what are the advantages and disadvantages of non-parametric methods?adv: open up possibility of analysing data sets when the normality conditions are not met
dis: not as efficient as parametric tests, do not find statistical significance as often as they should
Informationen
- Beschreibung
- 4 BE428 Karteikarten am BE428 Lecture 7: Parametric vs Non-Parametric Statistics, erstellt von Caitlin M am 10/10/2018.
- Erstellt von
- Caitlin M
- Aufrufe
- 2
- Bewertung
Mehr zu Parametric and Non-parametric Tests