You can reach Dataxiom:
By Internet
sales@dataxiom.com
By Mail
Dataxiom Software Inc.
3700 Wilshire Blvd., Suite 1000
Los Angeles, CA 90010
U. S. A.
By Telephone
(213) 383-9973
By Fax
(213) 383-3542
The support area of the Dataxiom Website will be updated frequently to provide you with product updates and answers to Frequently Asked Questions (FAQ's).
Last FAQ update: 9/18/97
1. During the installation process, I got an error message 'Could not write to file C:\WINDOWS\SYSTEM\COMPOBJ.DLL.' What's wrong?
This means that the file COMPOBJ.DLL is in use at the time of installation. A possible cause is that you have Microsoft Office (or Lotus SmartSuite, PerfectOffice Suite, or a similar application) installed and placed in your Windows' StartUp group, so that the application starts running whenever you start Windows. There are two ways to solve this problem.
- If you still have the error message box, do not click on any of the Retry or Cancel buttons. Instead, press <Ctrl+Esc> to bring up the Task List dialog box. Select Microsoft Office or whatever application uses COMPOBJ.DLL, and click the End Task command button. Then click Retry in the error message box to continue the installation procedure.
- If you have already exited the StatMost installation program, double click the desktop empty area to bring up the Task List dialog box and close the program that uses COMPOBJ.DLL (as described in the above step), or find the program that uses COMPOBJ.DLL and terminate it manually. For Microsoft Office, you will see the topmost control buttons on the top-right corner of the desktop screen. Click the rightmost button on the Office Manager and choose Exit to close Microsoft Office.
Back to Top
2. I installed StatMost successfully, but when I run it, I get an error message 'Missing OLE file or incorrect version.' What's wrong?
Maybe your OLE2 related files are too old or incompatible with the ones shipped with the StatMost program disks. There are five OLE2 related files:
- OLE2.DLL
- OLE2DISP.DLL
- OLE2NLS.DLL
- STORAGE.DLL
- COMPOBJ.DLL
These files are all located under the WINDOWS\SYSTEM subdirectory. You can either copy these files to a floppy disk or rename them as *.BAK files. Then delete these files from your WINDOWS\SYSTEM directory and reinstall StatMost.
Back to Top
3. I have a 386 computer with 8Mb RAM. When I tried to install StatMost, it was very slow. How can I make it faster?
One way to make the installation process faster is to copy all the StatMost distribution files to your computer's hard disk and install StatMost from the hard disk.
Back to Top
4. I'm having trouble importing my database using the ODBC support feature in StatMost. What could be wrong?
One possibility is that you do not have the very latest ODBC drivers for your database program installed on your system. Although we try to provide the very latest versions of the ODBC drivers with each new release of StatMost, Microsoft and the various database developers are constantly publishing new file formats and ODBC drivers. You can download the latest version of the driver files you need from the Microsoft Website. If having the new driver file on your system does not take care of your ODBC importing problems, please contact Dataxiom Technical Support.
Back to Top
5. I have five columns in a datasheet (columns A, B, C, D, and E) and I want to see the descriptive statistics of column A and column D. How can I do it?
You can perform the descriptive statistics by selecting the Statistics | Descriptive - General option. This will bring up the Descriptive Statistics dialog box. In this dialog box, select column A by clicking on column A in the Data Column List box. Then hold the <Ctrl> key and click the mouse button on column D. This way you can select (or highlight) nonconsecutive columns in the Data Column List Box. See Chapter 2, Section 2.7 of the StatMost 3.0 User's Guide.
Back to Top
6. How can I select nonsequential items in a List Box?
See the answer to the preceding question.
Back to Top
7. I can edit and view my data in the datasheet window just fine, but when I print it to my HP LaserJet 4P printer, I get very tiny characters. Why?
The font size of datasheet output depends on your specific printer. The default font of StatMost is set to Courier New with size 18. This setting is fine for most printers. With the HP LaserJet 4P printer, however, you need to change the font size to 48 or so. You can change the print font by selecting the File | Page Setup option. This will bring up the Page Setup dialog box. Click the Print Font button in the dialog box to change the print font for datasheet output. In fact, you can also change other page settings in this dialog box.
Back to Top
8. How can I divide or multiply one column by another column?
Select the Data | Transform - Simple Math option. This will bring up the Simple Math dialog box which allows you to enter a formula for the calculation. For example, if you want to create a new column C which is the product of column A and column B, you simply enter 'C=A*B'. For more descriptions, see Chapter 5, Section 5.4, and Appendix B, Section B.1 in the StatMost 3.0 User's Guide.
Back to Top
9. How can I generate random integers between 1 and 100?
Use the Data | Fill option. This option brings up the Block Fill dialog box (see Figure 5.1.1 in the StatMost 3.0 User's Guide). In the dialog box, enter the Minimum value as '1,' and the Maximum value as '100;' set the Series Option to Random; and check the Integer checkbox. Once the Integer checkbox is checked, the resulting values will be integers.
Back to Top
10. When I perform some statistical calculations, I see strange characters like 'NaN' or '1.#IND' in the report window. Why?
The symbol 'NaN' means Not-A-Number, and '1.#IND' is an undefined numeric value. This means that you may have missing values or text values in your data columns. Some statistical computations (such as ANOVA, Student t-test, etc.) ignore missing values, but other calculations (such as the Chi-square contingency table) take paired values so that the missing values or text values are used in computations. That's why you get all the 'NaN' or '1.#IND' characters. To avoid this situation, you can trim out the missing values (Data | Trim) or replace missing values by some other valid values (Data | Recode). In fact, there are several other strange characters in StatMost reports that signal missing values, such as '1.#QND', '1.#QNAN0', etc.
Back to Top
11. How do I set up my data and factors to run a 2 factor ANOVA (3 treatments with 2 blocks) in StatMost? I want to use the Analysis | Experimental Design | Two Factor "Treatment-by-level" method.
When specifying your factor size in the analysis dialog box, keep in mind that your slowest moving factor should be specified first (i.e., as factor A). For example, suppose the data are entered into the datasheet as follows: Low1 Med1 High1 Low2 Med2 High2. That is, all the variables for block 1 followed by all the variables for block 2. For this data setup, you should assign your factors as follows: Factor A: 2 (block), Factor B: 3 (treatment). Alternatively, if the data are organized somewhat differently: Low1 Low2 Med1 Med2 High1 High2, that is, all variables for treatment 1 (low), then all variables for treatment 2 (med), followed by all variables for treatment 3 (high), then the factors should be assigned as follows: Factor A: 3 (treatment), Factor B: 2 (block).
Back to Top
12. How do I analyze a 2 factor repeated measures design in StatMost? Both factors are within factors, one with 2 levels (area), the other with 6 levels (trial). My data are arranged in the datasheet like: area1trial1 area1trial2 ... area1trial6 area2trial1 area2trial2 ... area2trial6. I cannot figure out which ANOVA option to use or how to specify my model.
For this repeated measures design you should use the Experimental Design option in the Analysis menu and choose the Two Factors option, followed by the Treatment by Treatment by Subjects option. Given the way your data are arranged, you should specify your slowest moving factor (as entered into your datasheet) as Factor A (in this case, area, with 2 levels) and your fastest moving factor as Factor B (in this case, trial, with 6 levels).
Back to Top
13. I have selected data columns in a plot setup dialog box, but after I click OK I get a 'No Data Added' error message. What's wrong?
You did not add a curve to the new graph. In some of the graph setup dialog boxes (for instance, the 2D Scatter Dialog Box), there is an Add Curve button. By default, StatMost puts the initial columns in the X and Y positions and the value for the curve number next to the Add Curve button is '0.' This means there is no curve added to the new graph. You need to click the Add Curve button to add the new curve to a graph before you click OK.
Back to Top
14. When I set up a graph, if I click the Add Curve button, my Y value goes away. Why?
Each time you click the Add Curve button, StatMost automatically clears the Y position to allow you to choose another Y for the next curve. This prevents you from adding the same curve twice. Because StatMost can plot multiple curves in one graph, this Add Curve feature allows you to add many curves to the same graph.
Back to Top
15. I know how to add two or more curves in the same graph when I first create a graph. But how can I add more new curves to an existing graph?
Adding new curves to an existing graph is similar to creating a new graph with multiple curves. The only difference is that you need to select the existing graph you want to add curves to. Once that graph is selected, choose one of the options from the Plot menu, and set up the curve options the same way you create a new graph. Notice that the curve count (number displayed in the setup dialog box next to the Add Curve button) is for the newly added curves only. It does not include the curves in the existing graph.
Back to Top
16. How can I change the default color palette?
There are two ways to change colors in the color palette. One way is to use the File | Preference option. This will bring up the Preference dialog box. There is a Color Palette list box. Double-click (or right mouse click) on the color you want to change to bring up the Color dialog box. Another way is to change color in any one of the graphic object properties dialog boxes. At any color drop-down list box, double-click (or right mouse click) on the color you want to change to bring up the Color dialog box. Choose a color in the Basic Color list or click Define Custom Color to define your own color. The new color will replace the color in the Color Palette and will be used as a default color in your future Color Palettes in StatMost. The Color Palette will be saved to your STATMOST.INI file for future use. The Color Palette is used when you modify drawing colors of graphic objects. If you want to go back to the original default Color Palette, you can delete the ColorPaletteR, ColorPaletteG, and ColorPaletteB values in the STATMOST.INI file.
Back to Top
17. How can I select an individual bar in a stacked bar or group bar graph, or an individual area slice in a stacked area graph?
Group (stacked) bar and stacked area graphs are quite different from multiple scatter/line curves. There is a significant difference in the graphs' internal structure. For example, a group of five bars contains five pairs of data series but is treated inside StatMost as a single data series, while a graph with five scatter curves has five individual pairs of data series. In order to select an individual bar series in a group bar (stacked bar or stacked area) graph, you can click on the bar once. If you want to select another bar series, click on the other bar twice (not double-clicking). Clicking on it once will select the graph (not the bar), and clicking on it the second time will select it.
Back to Top
18. How can I print a graph without borders and crop marks?
By default, the graph page output options show paper header, paper footer, paper border, and crop marks. You can turn these options off by selecting the File | Page Setup option. This option will bring up the Page Setup dialog box, which allows you to check or uncheck the paper header, paper footer, paper border, and crop marks options.
Back to Top
19. The graph size is too small on my screen. How can I make it bigger?
You can change the graph size by selecting the View | Set Page Option. This option will bring up the Set Page dialog box, which allows you to set the page dimension and orientation. Note that the page size is different from the paper size. Normally the page size is smaller than the paper size. StatMost always places the graph page at the center of your paper. In order to print the paper information such as paper header, footer, border, and crop marks for 8.5''x11'' paper, you need to set your page size to 7.5''x10'' or smaller.
Back to Top
20. How can I make the tick marks on an axis point toward the inside of the graph?
By default, the tick marks on an axis are pointing toward the outside of the graph. To change them to point to the inside, modify the Tick Size to a negative value in the 2D Axis Properties dialog box (Figure 10.3.3). See Section 10.3 in the StatMost 3.0 User's Guide for details.
Back to Top
21. I'm drawing a scatterplot with error bars in StatMost and seeing the line connecting my datapoints is very confusing. The connecting line comes up automatically even though I don't ask for it. Is there any way to turn it off or hide it?
Many of the individual objects in a StatMost graph can be either shown or hidden, or even turned off completely. First, you must select your plot by clicking your mouse pointer anywhere inside the plot frame. Once you do this, "handles" should appear on the plot frame itself. Now to turn off the connecting line, go to the Format menu and choose Select Curve. This brings up the Select Plot Curves dialog box. From here, select Properties to go the the Error Bar Properties dialog box and click None under Line and OK to remove the connecting line from your plot.
In addition, if you also wanted to hide an object like the datasheet name from your graph, you can choose Show/Hide Objects from the Format menu. You then scroll down the list of plot objects until you come to the datasheet name object, select it, and click Hide followed by OK. This object will then disappear from your graph.
Back to Top
22. When I print a graph in StatMost, it appears on the page squished into the top left-hand quarter of the page. What's wrong?
This is a known driver problem on some printers. If you notice this problem, select the Preference option on the File menu from any StatMost window, select "Special Printer" in the Preference dialog box and then reprint your graph. Note that you may need to toggle the Special Printer option off and on up to three times before your graphs will print correctly. If you do not see this option in the Preference dialog box of your version of StatMost, please contact Dataxiom Technical Support.
Back to Top
23. When I print a graph in StatMost, the y-axis label is reversed and slightly shifted from how it appears on my screen and in Print Preview. What's wrong?
This is a known driver problem on some printers. If you notice this problem, select the Preference option on the File menu from any StatMost window, select "Reverse Font Rotation on Printing" in the Preference dialog box, and then reprint your graph. If you do not see this option in the Preference dialog box of your version of StatMost, please contact Dataxiom Technical Support.