ProStats™

Online User Reference

Copyright (C) 2001 Software Engineering Inc.

All Rights Reserved

Introduction

ProStats™ is a graphically-oriented statistical analysis program which provides users with the opportunity to analyze data directly on their Palm device. Currently, over 40 statistical values are calculated on two data sets of up to 100 points each. Results are displayed both numerically and graphically to provide the user with the exact values as well as visual trends in the data. ProStats™ provides fast convenient and solid statistical answers to anyone doing statistical calculations. Our many users include Scientist, Researchers, Medical Professionals, Technicians, Engineers, and students world wide.

Features and Functions:

• Statistical calculations include: Mean, Standard Deviation, Standard error of the Mean, Variance, Minimum, Maximum, Range, Sum, and Skewness.
• Histogram and Scatter Plots may be done quickly to characterize data.
• Coordinate Tracking is available for scatter and regression plots to quickly locate individual data points.
• Hypothesis testing (1 and 2 tailed t-testing) may be done to at 1, 2.5, 5 and 10% significance levels, and test results displayed graphically.
• Confidence Intervals may be calculated for the means of the data sets.
• Box and Whisker Plots which calculate and display quartiles.
• Probability calculations may be done on all data or parameters may be entered individually.
• Sample Size Calculator
• Regression analysis includes linear and exponential curve fitting, plotting, and interpolation.
• Intelligent parsing algorithm with error detection.
• Data may be entered using the Data editor or imported from a comma-delimited memo file.
• Data is saved and loaded as comma-delimited (CSV) memo files.
• ProStats™ is compatible with all three leading spreadsheet programs.

Testing and Compatibility

All statistical calculations have been verified by comparing the calculated output to either textbook examples, hand calculations, or spreadsheet outputs.

ProStats™ was developed on a Handspring Visor Deluxe (OS 3.1H) and Platinum (OS3.5H) using the Quartus Forth Compiler. The program has been extensively tested on a Palm IIIxe (OS3.5) and used on various platforms.

During the development of ProStats™, it was used and tested with the following programs and/or hacks without conflicts:

HackMaster, X-Master, ClipHack, ClearHack, PhlemHack, MagicTextHack, QuickBits, Afterburner,

PopApp, MenuHack, SelectHack, RPNpopUpCalculator, PopUpNote

LapTopHack, TextPlus.

Files included with ProStats™

Filename Description

ProStats.Prc program module

Module1.Prc program extension module

Module2.Prc program extension module

Module3.Prc program extension module

Module4.Prc program extension module

Licence.txt user license agreement

ProStats.PDF this document

Installation

Copy the Zip file to a storage directory on the PC, and extract all of the files. Use the Palm Desktop Install function to install all PRC files onto the Palm device.

Upgrading from a previous version

ProStats™ was formerly PalmStats as such all registration codes will continue to work.

Users upgrading from any previous version must first delete the existing program files on their handheld device prior to installing this software.

Failure to do so may cause unpredictable behavior and or fatal errors which may result in data loss.

Registration

A registration code, along with instructions regarding how to enter it, will be e-mailed after registering this software through one of the sales sites listed below.

This will usually take less than 24 hours. The registration cost is $13.95. Registration codes are good for all future upgrades

Currently, ProStats™ may be registered at:

PalmGear HQ www.palmgear.com

Handango www.handango.com

Controls and Functions

Refer to the following figures for information regarding the user interface of ProStats™ .

1 – X and Y data fields are used to enter data into ProStats™. Use Graffiti characters followed by a return character.

Data Markers help in aligning data point pairs visually and may be turned off in the preferences screen.

2 – ENTER is used to enter the data from the X and Y data sets into ProStats™ . The n: field will display the number of data points which have been entered. ProStats™ uses IEEE floating point calculations internally. Data contained in the field is parsed and converted to this format by the Enter key. (See the next section regarding the parsing algorithm)

3 – Data Set Selector this control specifies what data set the calculation are performed on (X, Y or Both). This control has the same meaning and function throughout the program.

4 – Descriptive Statistical calculations are chosen by checking the appropriate box.

5 – The Data Editor Hotkey is used to quickly switch to the data editor

screen. A similar key is provided on the data editor screen for a quick return to the main screen.

6 – CALCULATE runs the calculations and displays the results on the results screen as shown above.

7 – Scroll Arrows The up and down arrows scroll the data up and down one line.

The left arrow scrolls to the beginning of the data field and the right arrow scrolls to the end of the data field.

The data field to be scrolled is selected using the Data Set Selector ( X, Y, or both).

For example, to scroll just the X data field, select X and use the scroll controls to move through the data.

The Data Editor

ProStats™ provides a data editor screen for quick and easy data entry. Data may be entered using the keypad (2) or Graffiti.

To use the data editor, tap the stylus inside one of the fields and use the keypad to enter data points followed by Enter. The Keypad (2) is equipped with a backspace key which deletes the character immediately to the left of the cursor.

The Return hotkey (3) is used to quickly return to the main screen.

Data may be scrolled using the Scroll Controls (4) in conjunction with the Data Field Selector (1). For example, to scroll just the X data field, select X and use the scroll controls to move through the data.

Entering Data and Intelligent Input Data Field Parsing

ProStats™ is equipped with an fast, accurate and intelligent input parsing algorithm. Although the data entry process may seem a bit different to the novice,

it is quite powerful and provides an extremely high degree of data input integrity. Data points are entered for processing via the Graffiti system or the Data Editor previously described. These points may consist of any numeric value/character (either as shown or floating point format 1e3, 2.23e4 etc.) followed by a new line or carriage return. When ENTER is pressed, the parsing algorithm will attempt to convert the characters in the X and Y data fields to the internal IEEE floating-point format and store them internally as such. During this process, any errors encountered stop the parsing algorithm and are flagged as shown in the figure below.

1 – The point containing the error is highlighted and the field is scrolled so that the point is in view.

The character or characters must be edited out before parsing will continue successfully.

2 – The field is flagged by an ERROR label.

Notes:

• The parsing routine will ignore spaces (white space characters) in the data fields. This may be helpful when editing the data fields using Graffiti.
• Extra new line characters at the end of a data field will cause errors.
• These errors are flagged by the cursor positioned after the last data point of the data field showing the error. This is illustrated below.

The cursor will be positioned after the 2, if extra newlines

are present at the end of the X dataset.

• It is a common occurrence for an inexperienced user to have extra newline characters at the end of their data set

ProStats will only parse 100 data points into its internal format. Any additional character data in the data fields is ignored.

Preference Selection

The Markers check box is used to toggle the marker lines seen in the X and Y data fields. The numerical format may be changed using the Digit(s) and Decimal/Scientific controls. ProStats™ supports up to 6 digits after the decimal point.

The histogram and plot displays may be configured using the remaining controls. Gauss Plots refer to the normal distribution curve approximated when a histogram plot is drawn. Grids may be selected for scatter and regression plots. Dot Size may be Small (approx. 1 mm squares), Medium (1.5 mm circles), or Large (2 mm circles). This setting also changes the dot size for the scatter plot and regression plots. The # of bins is the number of bins the data will be sorted into prior to drawing the histogram. The width of each bin is equal to the data range divided by the number of bins. For example, if the range of a data set is 1500 and # of bins is set to 20 (as shown above) the width will be 75, or 1/20 the range. There can be from 10 to 100 bins in increments of 5. The X axis setting controls the minimum and maximum values over which the histogram will be plotted. When set to +/- 3.125 standard deviation, the histogram will be centered on the screen with both tails decreasing to zero. This is the view seen in the histogram figure. When set to Range of Data, the ends of the plot will be the minimum and maximum data points.

Special Menu Functions

In addition to the normal Palm editing functions shown below, two new functions have been added to the edit menu. Clear Data erases all the data contained in the field selected with the Data Selector (X, Y, or both). The Swap X-Y function exchanges the X and Y datasets.

All other editing functions behave the same as in other Palm applications. ProStats™ has been tested with the following edit-enhancing operating system extensions: ClipHack, ClearHack, MagicTextHack, MenuHack, and SelectHack.

Analysis Functions

ProStats™ provides the user very flexible data analysis capabilities. The figure below shows the analysis function available in this version of ProStats™.

Confidence intervals for the means of the X and Y data sets may be calculated at 80, 90, 95, 98, and 99% confidence levels. Calculations done on small sample sizes (n<30) are done with a t distribution. Sample sizes of greater than 30 points are calculated using the normal gaussian distribution.

A Sample Size Calculator has been provided which calculates the minimum sample size required for a given confidence level, mean, standard deviation, and total error of the estimate. The interval containing the mean is calculated along with the minimum number of samples based on a normal distribution.

Scatter Plots may be done to visualize the relationship between the X and Y data sets. This function treats the data sets as coordinate pairs with the X data set containing the X coordinate and the Y data set containing the Y coordinate. The figure below illustrates a typical scatter plot.

1 – Coordinate tracking is available by moving the stylus across the screen. The (x,y) position of the stylus is displayed.

2 – The Y axis is scaled with respect to the range of the Y data set. This value represents the maximum Y value in the data set.

3 – Grids are available and may be toggled on and off in the preferences screen. The grids are located at 25, 50 and 75% points on the each axis.

4 – This is the Origin point for the X-Y axis. This point represents the minimum x and y point and is displayed as a coordinate.

5 – The X axis is scaled with respect to the range of the X data set. This value represents the maximum X value in the data set.

The number of points in the X and Y data set must be equal in order to do a scatter plot. This screen is exited by tapping the HOME key.

Histogram Plots may be done to quickly determine the distribution of a data set. The figure below illustrates a typical plot of both X and Y data sets. The data set to plot is selected with the Data Set Selector on the main screen.

This screen is exited by tapping the HOME key.

1,2 – The Mean and standard deviation are calculated and displayed along with corresponding error bars.

3 – Gauss plots are available and may be toggled on or off in the preferences menu.

4 – the Median is calculated and shown with the small triangle marker.

Hypothesis Testing may be done on all data. These tests include both X and Y as individual data sets against a given mean, X and Y against each other, or tests on user supplied means. All tests may be done with 1, 2.5, 5, and 10% significance levels. This screen is exited by tapping the HOME key. The figure below shows the test selection list.

xbar = m and ybar = m tests the means of the respective data sets against a user supplied value (m).

x=y, S² equal test the means of the x and y data sets against each other assuming the sample size and variance are the same. This test is available in most spread sheet or data analysis packages and is commonly referred to as: t-Test: Two-Sample Assuming Equal Variances.

x=y, S² equal test the means of the x and y data sets against each other assuming the sample size and variance may not be the same. This test is commonly referred to as: t-Test: Two-Sample Assuming Unequal Variances.

Enter Params allows user supplied entries for the test parameters.

ProStats™ is equipped with an internal t-table for use in testing data sets up to n=30. Data sets containing more than 30 points are tested with respect to a normal distribution. The following figures show the test output of two t-tests.

1, 2 – The two-tail test limits (1) are shown shaded and the one-tail test limits (2) are displayed as solid lines. The test significance may be selected with the pick list. In this case it is set to 10%. The tail areas correspond to this value.

3 – The location of the test value (m) is displayed as a circle and line.

Testing two data set against each other may be done as shown below. Each distribution is drawn along with the respective tails and test results.

The data sets from the T-test may be displayed in several different ways to help visualize the relationship between the data.

Box plot with error bars indicating the Standard Deviations and

Boxes for the Standard Error of the Mean.

Default Histogram Plot.

Dot Plot with error bars

indicating the Standard Deviations.

Regression Analysis may be done using eight regression models:

y = bx + a, y = bx, exponential curve y = ae^bx, or polynomials.

The figure below shows the regression screen. This screen is exited by tapping the HOME key.

The Regression output may be copied to the clipboard

for use in the memopad.

After ProStats™ calculates the regression information, individual points may be calculated using the point calculator. All error values associated with the regression constants are also calculated. The following figures show typical regression plots of data. The plot screen is exited by tapping the HOME key.

This will return to the regression screen seen above.

This figure shows Residuals plotted.

The figures above show typical regression plots with the data fitted to polynomial curves. As with the scatter plot, the coordinate tracking function may be used by dragging the stylus over the screen.

Polynomial Regression:

ProStats™ will fit up to a 6^th order polynomial to a data set provided the figures below show several typical screen shots. The regression output may be selected and copied in to the clipboard for use in other documents.

The fitting algorithm used is a least-squared routine based routines from Bevington and Robertson (see reference). This routine is fast and accurate to about 5 decimal places.

The 95% confidence interval of the regression line may be displayed for the Y=ax+b function. The confidence band is chosen by checking the 95% Limit checkbox as seen below in the figure.

Notes on Curve Fitting

ProStats™ does not use iterative methods for its curve fitting routines. Therefore, when comparing its output to a system using for example the Levenberg-Marquardt Method the results will be slightly different. This will mainly be seen in the polynomial fitting routines.

Probability Calculations may be done on all data. These calculations include both X and Y with respect to one or two values, or user specified data input. All calculations are done assuming a normal distribution with n>30.

The locations of m1 and m2 are marked by small and large circles respectively. The probability of a value being less than m1 is shaded. This screen is exited by tapping the HOME key.

The Filer

ProStats™ is capable of reading and writing comma-delimited memo files. These files are used to store datasets or transfer data to and from a PC during synchronization.

1 – Filenames are displayed, along with their number (left margin) as shown in above. The filename list will display all of the memos in the memo pad database twenty at a time. The total number of memos is shown below the list. The Forward and Back controls are used to step though the memos in increments of 20.

2 – Source/Destination controls the place the data is coming from during a SAVE operation or the place the data will be going to during a LOAD. This control functions the same as the Data Selector on the main screen. For example to LOAD the data file into the X data set - - choose a file, select X and tap LOAD.

3 – The filename (Name:) may be specified during a SAVE to add a descriptive name to the data set. ProStats™ adds a ~ character to the front of all the files it saves to distinguish data files from other memos. The filer will only load files beginning with the ~ character. If a Name is not specified during a SAVE, the filer will create a unique name based on the date, time, and data sets specified with Source/Destination.

Data File Format

The figure below shows the comma-delimited format of a ProStats™ data file. Line one is the title and can be up to 40 characters long. Each remaining line may contain any numeric value separated by a comma. Spaces are allowed between numbers. Each line in the data file, including the title line, must be followed by a new line character.

Figure 1 Figure 2

Figure 1 has an equal number of points in the X and Y datasets. Figure 2 shows the filer output when there are more points in X than Y.

Note: When data is loaded from a data file into BOTH X and Y data sets, data will be alternately put into X and Y. The first, third, fifth, etc… go into X and the second, fourth, sixth, etc… go into Y.

Export of Data to other Programs

ProStats™ is compatible with all three leading spreadsheet programs.

The CSV and TAB to clipboard functions, available though the options menu,

allow easy transfer of ProStats™ data to Quicksheet 5.0 and MiniCalc Spreadsheets. CSV files written by the ProStats™ filer may be imported directly into Iambic TinySheets.

Step 1 Enter data into ProStats™

Step 2 From the options menu choose CVS-> Clipboard for

Quicksheet 5.0 or TAB->Clipboard for MiniCalc.

Both X and Y data sets will be copied to the clipboard independent of the

setting of the Data Field Selector

Step 3 Run the appropriate program and Paste the data directly into

the open spread sheet.

Quicksheet 5.0 MiniCalc

References

The calculations used in ProStats™ have been derived from the following texts:

Glantz, S. A., Primer of Biostatistics, 4^th ed. New York: McGraw-Hill, 1997.

Bevington, R., and Robertson, D., Data Reduction and Error Analysis for the Physical Sciences, 2^nd ed. New York: McGraw-Hill, 1992.

Crow, E., Davis, F., and Maxfield, M., Statistics Manual, New York: Dover, 1960.

Taylor, John R., An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements, 2^nd ed.

Sausalito, CA: University Science Books, 1982.

Contacting Us:

Email may be sent to seisupport@home.com

Our Web Page is http://members.home.net/seisupport/index.html

Any comments or observations are greatly welcome. ProStats™ will be updated, improved and features added based on user’s comments. Currently, ProStats™ is being used by students, engineers, medical professionals, and scientists world wide. It is our goal to produce the finest statistical analysis program available for the Palm Platform and value for the registration cost.

License Agreement

You should carefully read the following terms and conditions

before using this software. Unless you have a different license

agreement signed by Software Engineering Inc., your use of this

software indicates your acceptance of this license agreement and

warranty.

Registered Version

One registered copy of ProStats™ may only be used by a single person

on one Palm Computing^(R) Platform device. Since the registration

code is derived from the PalmOS user name, a new registration code

is required for each Palm Computing Platform device.

Governing Law

This agreement shall be governed by the laws of the State of

Indiana.

Disclaimer of Warranty

THIS SOFTWARE AND THE ACCOMPANYING FILES ARE SOLD "AS IS" AND

WITHOUT WARRANTIES AS TO PERFORMANCE OR MERCHANTABILITY OR ANY

OTHER WARRANTIES WHETHER EXPRESSED OR IMPLIED. Because of the

various hardware and software environments into which ProStats™ may be utilized, NO WARRANTY OF FITNESS FOR

A PARTICULAR PURPOSE IS OFFERED.

Good data processing procedure dictates that any program be

thoroughly tested with non-critical data before relying on it.

The user must assume the entire risk of using the program. ANY

LIABILITY OF THE SELLER WILL BE LIMITED EXCLUSIVELY TO PRODUCT

REPLACEMENT OR REFUND OF PURCHASE PRICE.

Evaluation and Registration

This is not free software. Subject to the terms below, you are

hereby licensed to use this software for evaluation purposes for

30 days only. If you use this software after the 30 day evaluation

period a registration fee of $13.95 is required. Registration is

available through one of the purchasing sites listed. When

payment is received you will be sent a registration code which

should be entered into ProStats™.

Unregistered use ProStats™ after the 30-day

evaluation period is in violation of U.S. and international copyright

laws.