ANOVA > Web-form > Online Help
WRAS ANOVA Online Help
Instructions
1) Select a monitored species from
the pull-down list of quantitative Variables (eg.Parrot fish).
2) Select a primary categorical variable from the pull-down list of Factors (eg.Region).
3) Pick a secondary categorical variable from the pull-down list of Sub-Factors (eg.Country).
4) Apply any necessary filter parameters to limit the spatial
and/or temporal extent of the dataset used in the analysis:
- Click to toggle between available geographical filter types. Data can be
selected by Region, Sub-Region, Country,
or any combinations of these filter parameters. Select those entries from
the respective pop-up lists to be included in the analysis. By default, all
entries are selected (highlighted), meaning that data for all geographical
areas in the database will be used.
- To apply a time-based filter and analyze data for a specific period, click
the Years button and select the years of interest from the
list. By default, the entire time series of available Reef Check data will
be highlighted.
5) Click the COMPUTE button to execute the analysis.
Interpretation of the ANOVA results
An F-Ratio statistics result table will be returned by ANOVA that describes whether significant differences in indicator variable values are observed
with respect to selected categorical factors.
- Factor effects are deemed significant when associated P-values are less
than or equal to 0.05 (5%).
- Within Group effects consider whether selected factors in isolation account for observed variability
in the data or have a statistically discernable effect.
- Between Group effects consider whether significant interactions or synergistic
action between categorical factors occur.
In addition, for each categorical factor (eg. region & year) a table listing
sample size, average and standard deviation for all data categories analysed in the ANOVA are returned.
- This allows one to compare for example which particular regions and
years may have yielded a significant result in the ANOVA table (individual factors
are significantly different if there means and standard deviations do not overlap).
- Individual factors that are not significantly different in terms
of their observed indicator variable values will exhibit overlap in their means and standard
deviations
Further Documentation
|
