Statistical tests included in ANY-maze

Section 1

Analysis of nominal data

ANY-maze can perform one-way (but not two-way) analysis of nominal data, such as the first zone entered in a test. To do this it uses the tests shown in table 1. If the result of this analysis is significant (by default, p ≤ 0.05) and you’ve requested post-hoc analysis then the system will employ the tests shown in table 2.

Number of levels of independent variable	Number of levels of dependent variable	Independent samples?	Test
2	2	Yes	Fisher exact test
		No	Cochran Q*
	> 2	Yes	Contingency table
		No	Cochran Q
> 2	2	Yes	Contingency table
		No	Cochran Q
	> 2	Yes	Contingency table
		No	Not supported

Table 1. Tests used to perform one-way analysis of nominal data in ANY-maze.
*Strictly speaking this should be analysed using the McNemar test, but the Cochran Q test yields the same answer.

 

Number of levels of dependent variable	Independent samples?	Ns are equal?	Test
2	Yes	–	Not supported
	No	Yes	Multiple comparisons of data subjected to the Cochran Q test (Marascuilo & McSweeney)
		No	Not supported
> 2	Yes	–	Not supported
	No	–	Not supported

Table 2. Tests used to perform post-hoc analysis of nominal data in ANY-maze.

 

Section 2

Analysis of continuous data

ANY-maze can perform one-way parametric or non-parametric analysis of continuous data using the tests shown in table 3. For two-way analysis the tests used are shown in table 4. If the result of this analysis is significant (by default, p ≤ 0.05) and you’ve requested post-hoc analysis then the system will employ the tests shown in table 5.

Parametric data	Number of levels of independent variable	Independent samples?	Test
Parametric	2	Yes	t-test
		No	Paired sample t-test
	> 2	Yes	Between subjects ANOVA
		No	Within subjects ANOVA
Non-parametric	2	Yes	Mann Whitney U
		No	Wilcoxon
	> 2	Yes	Kruskal Wallis
		No	Friedman

Table 3. One-way parametric and non-parametric tests used to analyse continuous data in ANY-maze.

 

Parametric data	Independent samples?	Test
Parametric	Both IVs independent	2-way between subjects ANOVA
	One IV independent	2-way between/within ANOVA
	Neither IV independent	2-way within subjects ANOVA
Non-parametric	Both IVs independent	2-way Kruskal Wallis
	Either IV not independent	Not supported

Table 4. Two-way parametric and non-parametric tests used to analyse continuous data in ANY-maze.

 

Parametric	Comparison to control?	Independent samples?	Ns are equal?	Test
Parametric	Yes	–	–	Dunnett
	No	–	–	Bonferroni Duncan Fisher LSD Scheffe Sidak Student Newman Keuls (SNK) Tukey
Non-parametric	Yes	–	–	Not supported
	No	Yes	Yes	Non-parametric analogue of the SNK test
			No	Dunn
		No	Yes	Non-parametric analogue of the Tukey test for repeated measures
			No	Not supported

Table 5. Post-hoc parametric and non-parametric tests used to analyse continuous data in ANY-maze.