MULTREG2 performs multiple linear regressions on data sets stored as lists in the TI-84. It provides essentially the same operations as the Stats-List Editor on TI-89's:

• Determines a linear regression from input lists.
• Calculates basic statistics like the correlation coefficient, the coefficient of determination, the F-score and its P-value, and the standard error of estimate.
• Calculates coefficient statistics like the coefficient standard errors, their corresponding T-scores and P-values.
• Provides predicted values with predicted confidence intervals for a given set of inputs

Anyone taking a Statistics class without a TI-89 or better (or equivalent computer software) will find this tool useful.

I WANT IT!!!

Choose the right program for your calculator, then download the file and copy it to your calc with TI-Connect. By using the program, you indicate your acceptance of the MULTREG License Agreement.

WARNING: THIS SOFTWARE IS IN BETA FORM. I have not yet confirmed that everything works yet. PLEASE be on the lookout for calculation errors! (i.e. by checking with your classmates using TI-89's).

How to Use

• Once you've downloaded it to your calculator, press the [PRGM] key, use arrow keys to select the program, and press [Enter] to run it.
• To make a new regression, select New Regression from the main menu.
• To view the results of the last regression, select View Last One. This screen automatically appears once a regression is created.
• There are a few minor differences between the TI-84 and TI-83 versions on the View Results screen.:
  • MULTREG2 for TI-84 creates predicted confidence intervals on the Predict option.
  • MULTREG3 for TI-83 shows all information regarding coefficients on the Coefficients option (i.e. the values, T-scores, P-values, and standard errors).

I FOUND A PROBLEM!!! Or...I want a new feature!

Come post on my forum here. I'll try to get back to you ASAP...

MULTREG License Agreement

MULTREG, the Program, is Copyright 2009 by Jeffrey Booth, the Author. You must never misrepresent the origin of the Program.

The Program is distributed on an "AS IS" basis, without warranties of any kind -- either express or implied. The Author shall not be held liable for any damages arising from the use of the Program.

You may use the Program for commercial purposes, but may not charge others a fee for access to and usage of the Program above any reasonable distribution costs (the costs associated with making the Program available to someone for use).

You may create derivative works from the Program, but may only receive compensation for your Additions to the Program. A "derivative work" is a software program that contains a copy (either of parts or of the entirety) of the Program in some form; it does not include output generated by the Program. "Additions" mean the results of any work that the Author did not perform.

If any part of this Agreement is deemed unenforceable, the remaining parts shall remain in effect.

Brief History

MULTREG2 (April 2009)

This version involved some significant modifications to MULTREG. With my college math skills, I was actually able to read and understand the Wikipedia material on multiple linear regressions. As it turns out, Wikipedia had all the answers: if I just typed everything in correctly, I would "magically" get the exact same results as the TI-89. This version took about 12 hours, and is now almost as featureful as the Stats-List editor. Oh...if only I knew how to read the math lingo back then...

MULTREG (April 2008)

I developed this version in about 30 hours. It was my first program in pre-89 TI-Basic. I programmed it on the school's TI-83, and it could calculate the coefficients and their standard errors -- as well as T-scores and P-values for the coefficients. However, it couldn't calculate confidence intervals for predicted values, disappointing the classmates who relied on it for their Statistics test.

Credits

I would like to acknowledge the following sources, which all contributed to the existence of this program...

• The article Wikipedia: Linear Regression at en.wikipedia.org/wiki/Linear_regression was the source for almost all the formulae used in this program.
• TIEmu - a TI-89 calculator emulator. I used this program to run the TI-89's Stats-List Editor, which gave me the expected output for my program. NOTE: I actually HAVE a TI-89, but it won't turn on for some reason.
• Virtual-TI - a TI calculator emulator. I used it to code and test my program on a TI-83.
• The com2com project allowed me to connect my virtual calculators to TI-Connect to download my program.
• The TI-89 Stats-List Editor was my basis for comparison. I compared my program's output to it, and modeled my program after it.
• The TI-83+ Graphing Calculator Guidebook described the commands I could use.
• NEW METHODS FOR SIMULATING THE STUDENT T-DISTRIBUTION -
  DIRECT USE OF THE INVERSE CUMULATIVE DISTRIBUTION by William T. Shaw. http://eprints.maths.ox.ac.uk/184/1/tdist.pdf This paper represents a Student's t-distribution as the inverse of a function. Therefore, I could make an inverse T-distribution in the future that might run much faster...
• Casey Daniels, my college roommate, let me borrow his TI-84+ for developing MULTREG2. It's a lot easier to program on the actual calculator than to mess around with emulators...
• Paul Rae's OHS AP Statistics Class - if everyone had TI-89's, I wouldn't have bothered with this program.