ASM - exam

Post-hoc tests are used in the analysis of variance (ANOVA)?

• To determine whether there are „within group” differences

• To calculate F statistics

• To determine which groups are significantly different

• To calculate NIR (the least significant difference)

Assumptions of the ANOVA F-test

• Randomness and independence of errors; high correlation coefficient between groups; normal distribution in population

• Normal distribution in population; homogeneity of variance; randomness and independence of errors

• Randomness and independence of errors; the same number of samples in each group; homogeneity of means

• Minimum number of group equal four; homogeneity of variance; normal distribution in population

In ANOVA analysis total variation is divided into:

• Sum of squares between groups and sum of squares between means

• Variation due to treatment and variation due to random sampling

• Individual differences within groups and variation due to random sampling

• Sum of squares between groups and error means

Degrees of freedom are:

• The number of values that are free to vary once certain parameters have been established

• The number of values that cannot be change during analysis

• The number of values to be given as a constant

• The number of F-tests parameters

F-test is significant when:

• Test statistic is smaller than 1

• Test statistic is greater than 1

• Test statistic is smaller than critical value

• Test statistic is greater than critical value

When do we use post-hoc tests?

• After a significant overall F test

• After a insignificant overall F test

• In place of an overall F test

• When we want to determine the impact of different factors

Cluster analysis is used to:

• Reduce the number of variables

• Divide a given set into disjoint subsets

• Calculate the correlation between variables

• Transform input set to orthogonal factors

Homogeneity criterion in CA (cluster analysis) means:

• Objects belonging to the same group should be exactly the same

• Objects belonging to the different group should be exactly the same

• Objects belonging to the same group should be as similar as possible

• Objects belonging to the different group should be as similar as possible

Stages of cluster analysis are:

• Data nomalization, variables orthogonal transform, grouping method selection

• Measuring distance between objects, grouping method selection, performing posthoc test

• Mapping objects in space, performing agglomeration procedure, performing post-hoc test

• Grouping method selection, normalization of variables, variable selection, measuring distance between objects

/ In the complete-link method (in CA) the similarity of the two clusters is equal to

• the similarity of the pair of the least similar (furthest) observations from both groups

• the similarity of the pair of the most similar (closest) observations from both groups

• the similarity of the centre of gravity from both groups

• the similarity of the median of all distances