Statistics View

The Statistics view provides the capability to perform a variety of analyses around the coding structure, the Codes each Respondent has Comments coded to, and optionally the Rating values you gave to the Codes when you created them. (If you didn't assign values and if you didn't change the Symphony default, a value of 1 is assigned to each Code.) The primary objective is to produce quantified data from the qualitative data that make up the Comments and coding structure. These quantitative data can then be analyzed and reported on.

Defining Queries

To perform an analysis, you choose the entities and content types you want to group by, by checking the appropriate checkboxes at the top of each Filter list. You choose the statistics you want to produce by checking the checkboxes next to each statistic in the Available Statistics list. Each selection becomes a column in the results grid. To change the order of the columns, drag and drop the column header to the desire position. Columns can be removed from the results grid either by removing the checkmark from its checkbox, or by right-clicking the column header.

Note that you do not have to specify both Group By columns and Statistics columns, but you do have to specify at least one.

By default, an analysis is run using the entire project data. You can run an analysis for a cross-section of data by checking appropriate filter options in the Filter Group.

Running an Analysis

Once you specify the query, you run it either by clicking the Run/Refresh button on the toolbar, or by pressing Function Key F5.

Row Sort Order

By default, the sort order of the output rows is based on the Group By columns. The output is first sorted by the left-most Group By column, with sub-sorts continuing to the right. Each sort is based on the natural order of the column. For example, for Code columns, rows are sorted according to the order in which you placed the codes; Demographics are sorted according to the order you defined for the demographic values; Respondents are sorted on last name. After an analysis is run, the sort order can temporarily be set on any column by clicking its column header. Clicking the column header multiple times toggles the sort order between ascending and descending.

Saving and Retrieving Queries

The query you define exists only as long as Standard View is open. Optionally you can save a query and retrieve it at a later time. Queries can be retrieved in Statistics View, Report View, or they can be embedded in a Word report. (You embed a query by creating a bookmark in the Word document, using the name of the query.)

Filter Group

The Filter Group enables you to specify a subset of the project for the analysis.

Group by Project Member

If the Project Member list is checked, a row is produced for each Staff member. The values in the result set represent averages for all of the Respondents assigned to each Project member.

Results Grid

The analysis results are output here. A result set is produced for each selected Filter List. If more than one Filter List is selected, additional rows are produced beneath the previous list.

Group by Respondent

Produces a result for each Respondent. This is most useful for comparing how Respondents rate against each other.

Group by Code Level

Produces results for each Code at the specified Code Level (in this case, Category, which is the default name for the top level Codes.) When a Code Level list is chosen, a row is produced for a Code only if the Code or one of its descendents -- for example, an Issue within a Category -- has a Rating value other than zero. Note that inclusion is based on the coding structure, and is independent of whether Comments are actually assigned to the Code. The values produced represent the averages for all Respondents within each Code.

Group by Demographic

When a Demographic is checked, a result set is produced that contains aggregates for each value within the Demographic. In the above example, two rows would be produced, one showing the averages for all Poets, and another showing averages for all Scientists.

Grouping Entity

This is the name of the entity or group of entities for which each row of data was produced. A column is included for each entity

Include Empty Results

When this box is checked a row is produced for each grouping entity, even if there is no data. As can be seen in the sample screen shot, there is no row for Scientists within Communications. this is because there are no scientists with Comment coded to the Communications category. If the Include Empty Results checkbox were checked, a row would have been included, with zero values in the appropriate columns. This is useful for situations such as feeding a data series to a chart and the chart expects the values either way.

Available Statistics

These are the various statistics that can be calculated. Below is each choice and a description of its meaning. In every case, the numbers are based on the project content after any filters have been applied. Note that any references to Demographics pertains to inclusion or exclusion based on Demographic values. If no Demographics are specified as Group By columns, results by demographic are always equal to overall results.

The best way to understand what the various statistics mean is to try simple queries where you can easily grasp the result, then add Group By columns, Statistics columns, and filters gradually and assess the results.

Count of Respondents

This is the number of Respondents who fit the Group By criteria for the current row. If no grouping is specified, this number will be equal to the number of Respondents included in the analysis. If Group By Respondent is specified, this number will be one.

Count of Comments

This is the number of Comments that fit the Group By criteria for the current row.

Count of Respondents Comments

This is the number of Comments belonging to the Respondents represented in the current row.

Demographic Size

This is the number of Respondents who match the demographic Group By criteria, if any. If other Group By criteria (for example, a Code Level or Tags) are specified in the analysis, this number does not change, so it provides a means of measuring how pervasive a demographic is within a given Group By criterion.

Sample Size

This is the total number of Respondents in the analysis. If no filter is set, this number will be equal t the number of Respondents in the project.

% of Respondents Comments

This number shows what percent of the Comments belonging to the Respondents represented in the current row match the Group By criteria. The calculation is ((Count of Comments)/(Count of Respondents Comments)) x 100.

% of Demographic Comments

This number shows for the current row, what percent of Comments belonging to Respondents with matching demographics match the Group By criteria. The calculation is ((Count of Comments)/(Count of Demographic Comments)) x 100.

% of Sample Comments

This number shows for the current row, what percent of Comments belonging to the Respondents represented in the current row match the Group By criteria. The calculation is ((Count of Comments)/(Count of Respondent Comments)) x 100.

Respondents % of Demographic

This is the number of Respondents with comments coded to the code(s) included in the current row, presented as a percentage of the Respondents whose demographic values match the current row demographic values -- that is, their Demographic values. The calculation is ((Count of Respondents)/(Demographic Size)) x 100. Note that if no Demographics are included in the grouping, the Demographic is equal to the Sample Size.

Demographic % of Sample

This is the number of Respondents whose demographics match the current row, presented as a percentage of the Respondents included in the analysis. The calculation is ((Demographic Size)/(Sample Size)) x 100.

Respondents % Of Sample

This is the number of Respondents with comments coded to the code(s) included in the current row, presented as a percentage of the Respondents included in the analysis. The calculation is ((Count of Respondents)/(Sample Size)) x 100.

Average Rating

This is the average of the ratings gives to the Respondents included in the current row (Count of Respondents).

Maximum Rating

This is the theoretical maximum rating. The theoretical maximum is reached only if every Respondent used to produce a row of data has content coded to every Code within the scope of the analysis, where the Rating is a positive number, and no content is coded to negative values. Depending on your rating system, this value may or may not be useful. When you code using emergent methods -- that is, your coding structure emerges by virtue of analysis -- it is normal for no Respondents or groups of Respondents to reach the maximum rating. If however your coding structure represents a classification of predefined value judgments -- a set of behaviors a job candidate is measured against for example -- it is possible that maximum rating could be reached.

% of Maximum Rating

This is the Average Rating as a percent of the Maximum possible rating. Depending on how your rating system is defined, this value may or may not be useful.

Selected Statistics

This is the list of statistics included in the analysis. A column is included for each statistic, in the order they appear in this list, and following the columns included for the Groups you selected.

Group by Tag

When this list is checked, results are grouped by Tag.

Miscellaneous Columns

These are non-statistical columns that can be added to queries.