Analysis reports providedOnly a few reports are currently provided. If you create your own useful reports and send them to me, I will integrate them into the next version of DeltaAccess. Also, please notify me about any problems with DeltaAccess analysis reports; so far they have not been tested exhaustively.
Reports are created the first time you use them, and unlike the queries, are not deleted after use. You do not have to save them under a new name as you would for queries.
To keep the reports readable, you are frequently asked to enter the first and last character or item number (CID or IID) to be plotted. You can enter, e. g., 1 and 999 to get all characters or items. Please note that in some graphs only every second or third character may be labeled if you try to plot too many characters.
NOTE: There is a bug in Microsoft Access, involving the use of parameters to restrict the scope of the report. Whenever you are asked to enter something like report!FirstCID, or report!Type (German version: e. g. Bericht!FirstCID) simply click OK. Normally this occurs only the first time a report is created from the DeltaAccess dialog box, not when opening existing reports from the database explorer.
Since charts, graphs, or diagrams are in fact reports with embedded charts, compare the following chapter: Analysis charts/graphs provided as well.
Reports to document your project definition or data
|
|
Name in database explorer = <Projectname>_ + ... |
SimpleCharDefinition |
|
Creates an overview of the character definition of your descriptor project, including Character headings, if available. When you open the report, you are asked about the inclusion of character states and modifiers. For character states you can include all states, only the normal states (without the special states (U/V/-), or no character states. The internal state 'TE' for text characters is never printed, the special state 'TE' for categorical characters is included with the normal states. Similarly, the modifiers defined for each character can optionally be included. Image definitions are currently not printed. You can use this report (like all other tabular reports) to export the information to Word or another RTF-compatible program. Click on the toolbar button |
|
SimpleItemDescriptionReport |
|
Can be used like natural language descriptions, but presents data in a tabular form for each character. Useful for error checking, creating printed output to work off-line. This is the only report which can potentially become very large (several 1000 pages for large projects with several 100 items and a large character definition). For very large data sets it may take 15 minutes or more until the formatting is finished (i. e. the last page reached in preview mode). When opening the report you are therefore prompted for the first and last item to print. You can enter, e. g., 1 and 9999 to print all items. If formatting takes too long, you can abort the report by pressing Ctrl-Break, and open it with fewer items, or open it based on an item subset. |
|
Item description report w. checkboxes for categorical char., includes notes, numerical/text char. |
ItemDescription_Checkbox |
A full item description which uses checkboxes for the states of categorical characters. This is one of the most powerful reports of DeltaAccess. It is documented in the separate chapter Checkbox report. |
|
Simple tabular item name report. 2 columns |
SimpleItemNameReport |
Basic item names, IID, abundances |
|
Simple tabular item name report, grouped by 'genus'. 2 columns |
SimpleItemNameReport_Genus |
As above, item names are grouped by 'genus' (= the 1st word in ItemName). Useful for overview in large item sets. |
|
Simple tabular item name report, sorted by % of unused chars (UV- = used) |
SimpleItemNameReport |
As Simple tabular item name report, plus the information about how many characters have been filled in like in the Unused Chars per Item queries is added. The special states (U/V/-/TE) count as normal character states. |
|
Simple tabular item name report, sorted by % of unused chars (UV- = unused) |
SimpleItemNameReport |
As above, except that the special states (U/V/-/TE) are treated as if non-existing. Compare Unused Chars per Item queries. Note that the last two reports have identical names, you can only use either of them. Rename the first before you create the other one to use them both concurrently. |
|
A report is available in the database explorer only after it has been created once using the DeltaAccess dialog box for the given descriptor project. Since the analysis charts/graphs described in the following chapter are always embedded in reports they are listed under reports in the database explorer as well.
*** Two further reports of a new project-independent generation of reports are available from the character dependency tab page of the new character editor or from the database explorer only. They are called DescrCharDependencyDefinition_ByControlled and DescrCharDependencyDefinition_ByControlling and are not listed in the Analysis pick list of the DeltaAccess dialog box. This type of reports will be increased in the future, after which the access to reports will be reorganized.
These character dependency reports list controlling characters, controlling states, and inapplicable characters (= controlled characters). The first report ('by controlling') lists all controlling characters and states together with the characters they control, the second report ('by controlled') lists all characters that are controlled together with the controlling characters and states. Each of the two reports contains the full character dependency definition, only the output is sorted the opposite way.
For more general information about Microsoft Access reports see the Contents of the Access 97 help file, Introduction to Microsoft Access 97, Reports, especially:
Reports: What they are and how they work
Ways to customize a report
Sections of a report
and the help topics in the Contents of the Access 97 help file, under Working with Reports.
List of all analysis options Contents of Analysis 
Main table of contents
Analysis group: Tabular reports, Name: Item description report w. checkboxes for categorical char., includes notes, numerical/text char.
This report is intended to print out form sheets, which can be corrected offline, or to enter data on paper. It exports all important item description data. The output is grouped into 4 sections:
categorical (UM/OM) characters with checkboxes (see figure above),
notes/comments associated with categorical data,
numerical characters, and
text characters.
Within each group the characters are sorted by character number (CID). This design was chosen in preference of a strict character sequence, to guarantee that the layout of the checkboxes is identical between items. This is very helpful for data entry and proofreading. The notes to categorical data are output separately, but modifiers are output immediately in front of the character states. Note that classical DELTA has no modifiers, so you had to use the Notes/Comments feature much more frequently. Each item will be printed starting on a new page, and the states are always kept together with the character. This might result in some free space at the bottom of the pages. The sequence of character states defined in the item description is ignored, since all available states are printed. Character headings are not yet included in the output; they might be included in a later version of DeltaAccess.
This report will be rather slow: the number of records equals the sum over all characters of the product of items times character states. I recommend testing it on a subset before running it on a large project with several hundred items. When you start the report you are asked for the first and last item number, so you can easily print, e. g., 50 items at a time. You can create item subsets, to restrict it other than by a range of item numbers.
You are also asked whether you want to output the special states (U/V/-/TE) or not. If you click OK without entering anything, special states will be ignored in the output. If you enter, e. g., 'U' instead of the requested '*', a checkbox for only this special state will be added to the output. See also the development note.
The report is explicitly designed so that it can be exported in RTF format to Microsoft Word. This is convenient, if you want to send proofreading forms to your colleagues using email. Most Windows word-processors should be able to read the RTF format.
Finally, if you want to print a report with only part of the data, e. g., without the text characters, you can modify the saved report. Open the report in the database explorer in Design mode. In the center of the report, under the gray horizontal bar labeled "Detail", you see 4 subreport controls for the 4 report sections mentioned above. Select the control you want to remove, press the delete key on the keyboard. Optionally you could rearrange the remaining subreport controls to eliminate superfluous space. Save the report under a new name.
Compare the chapter Creating item descriptions as HTML forms for similar report using checkboxes. The HTML forms can be used directly in a www browser to edit exiting or create new item descriptions.
Tabular reports Contents of Analysis Main table of contents
Analysis charts/graphs provided
In MS Access, a chart, graph, or diagram is always embedded in a report. The available analyses are described in the following sections:
Charts: Item analysis
Charts: Character analysis
Charts: Two-character comparison
Overview of analysis options List of all analysis options Contents of Analysis Main table of contents
Currently there is only a single report with a graph which provides item specific data. Please tell me what other graphs you would like to see:
Graph of mean/ranges of numerical characters for each item: A graph with mean (as a bar chart), normal range (as a box), and min/max extremes (as a line) or all items, for each numerical character on a separate page. The graph is sorted by item mean, of no mean is recorded for an item, the middle of the ranges is used instead. If you want more complex item sets than simple ranges or item numbers, you can create an item subset and plot that. -- This is currently the only graph in which items are compared. Which other are graphs do you consider desirable (other than phylogenetic/similarity comparisons which require special algorithms)?
Analysis charts/graphs provided List of all analysis options Contents of Analysis
Character / character state analysis:
A graph similar to the one comparing numerical characters for each item is provided for multistate characters: Mean, range, min/max of categorical/multistate characters, over all items (Graph). Note that the mean of ordered categorical characters can not be interpreted as the mean of a numeric character. By definition, the interval between ordered character states is not defined. Yet, the calculation of a mean assumes that the interval is identical. This may sometimes be a good approximation, in which case the graph for these characters may be useful. The graph for the unordered categorical character on the second page of the report is even more difficult to interpret, in most cases one of the following character state usage graphs will be more useful. The name of the saved report in the database explorer is <Projectname>_CategoricalStatGraph.
Char-state usage BY Char (graph): This is the graphical equivalent of the Char-state usage BY Char (crosstab) query! Note that because of the logic of the graph (x/y graph!) all characters are present on the x-axis, but only the character states of ordered/unordered categorical characters (UM/OM) are plotted. The special character states 'U' and 'V' are plotted as 0, the non-applicable marker '-' as -1. Only the first 15 character states are plotted, if you have characters with more states all higher states are ignored. The first page of the report contains an overview of all items, the following pages contain the graphs for each item separately. The name of the saved reports in the database explorer is <Projectname>_CIDxCS_Graph.
While the previous graph only plots the presence or absence of a character state, Character state usage over all items, separate bar graphs for types: UM/OM/RN/IN plots the distribution of character states (average of all items) as percent of all items for which this character has been used. Thus the base is not total items, and is different for each character. The name of the saved report in the database explorer is <Projectname>_CIDxCS_BarGraph_by_Type.
Analysis charts/graphs provided List of all analysis options Contents of Analysis
Charts: Two-character comparison
Three different reports with charts are available to compare the character states of 2 characters. The results are graphical representations of the cross-tabulations in the analysis queries Char-state usage Char 1 x 2 (crosstab). In each case you are asked which characters you want to analyze each time you open the report. The number and name of the selected characters and character states are printed on top of the report.
|
|
Name in database explorer = <Projectname>_ + ... |
Char. state usage of Char 1 x 2 (bar graph using CS; like a crosstab) |
CIDxCID_BarGraph_CS |
If you have a large number of character states, consider the bubble graph chart below, which is much less confusing in such cases. |
|
Char. state usage of Char 1 x 2 (bar graph using CS-Name; like a crosstab) |
CIDxCID_BarGraph_CSName |
Identical with previous graph, except that the character state names are plotted directly on the axis. This is not always better, since it leaves less space for the chart, and the states will be sorted alphabetically. |
|
Char. state usage of Char 1 x 2 (bubble graph using CS; like a crosstab) |
CIDxCID_BubbleGraph |
The area of the bubbles is proportional to the number of items which posses the given cross of two characters. Further, a linear regression of the item counts is calculated. The correlation coefficient r2 can give you an idea of the amount of total variation that could be explained by a linear relationship between the characters. Note that a correlation analysis is only meaningful if both characters are ordinal. It is not possible to create a version with character state names plotted on the axis like in the bar graph chart. |
|
Previews of Two-character comparison graphs:
Char. state usage of Char 1 x 2 (bar graph using CS; like a crosstab):
Char. state usage of Char 1 x 2 (bar graph using CS-Name; like a crosstab):
Char. state usage of Char 1 x 2 (bubble graph using CS; like a crosstab):
Analysis charts/graphs provided List of all analysis options Contents of Analysis 
Next
or select Tools, Office Links, Publish it with Word from the menu.