Why are lingreg calculations not accurate?

SOURCE: Wrong answer with squared negatives.

Your calculator will square a number before it applies the negative sign, so what you're actually doing is squaring 2, then applying the - sign to your answer. You need to put the -2 in parenthesis, so (-2)^2. That will give you 4. Hope this helps!

SOURCE: wrong trig ratio

Your calculator is set in radians. Press mode key. Arrow down to 3rd line. Highlight degree and press enter. Your calculator is now in degrees, not radians. To get out of the mode screen and back to home screen, press 2nd and then mode.

SOURCE: TI-84 Plus Linear Regression (ax+b) ERR:DIM MISMATCH

Hello,
Some old configuration settings, or old variables are spoiling your fun. Clear the RAM and start again.
Hope it helps.

SOURCE: On my TI-30xs Texas Instruments, I cant figure out

Use the toggle button above the enter key. It looks like 2 arrows

SOURCE: TI-83 Texas calculator, exponential regression.

Read this post. It answers all your questions and shows you HOW to do find your equation of best fit, correlation coefficients, etc.

Before doing statistics, you have to prepare the calculator by clearing old data that may generate errors.

A. TO CLEAR OLD DATA

1. Press [2nd][+] (MEM) to open the memory management screen.
2. Press [4:ClrAllLists]
3. The command echoes on main screen, press [ENTER] to run it.
4. Calculator displays Done.

B. RUN the SETUP EDITOR Command

1. Press [STAT][5:SetUpEditor] [ENTER] ; calculator displays Done.

C. TURN DIAGNOSTICSON option (optional)

To obtain the correlation coefficient in a statistical analysis on the TI8xPlus family of calculators you must activate the DiagnosticsOn option. To access the command

1. You press [2nd] [0] (CATALOG)
2. While the list of commands is displayed, press [X^-1] (D)
3. Cursor jumps to first command that starts with D, dayOfWk(
4. Press the DownArrow 9 times to reach DiagnosticsOn
5. Select it and press [ENTER]
6. Command echoes on main calculator screen with blinking cursor
7. Press [ENTER] to set the option
8. Calculator displays Done.

From now on, if a correlation coefficient has meaning for a regression model, it will be calculated along with the other statistical results. If it has no meaning to the model it will not be calculated.

D. ENTER THE DATA
Press [STAT][1:Edit]
Type in your D (X ) values in L1 and the N (Y) values in L2 (see picture 1). Sorry the picture does not have YOUR values.


TO OBTAIN THE STATISTICAL RESULTS (optional)

1. After you finish entering the data, press [2nd][Quit] to exit StatEditor.
2. Press [STAT] and highlight Menu (CALC)
3. In the list select [2:2-Var Stats]
4. Command echoes on main screen.
5. Press [ENTER] to calculate the stats

Here is what you get (THESE~ARE NOT YOUR RESULTS). Your results will be different.

Notice the down arrow next to n=7. That means the results screen has more values. Press the down arrow to display.

You notice that the n=7 line now has an arrow pointing Up. Notice the Down Arrow on the bottom line (more data is coming). Here is the rest of the statistical results.

There is no arrow at the bottom of the picture. End of statistical results.
The screen captures with 2-Var Stats summarize all the satistics.

However there is no correlation coefficient, no line of best fit, etc. To obtain that you have to perform a regression analysis.

E. PERFORM A REGRESSION ANALYSIS
To perform a regression analysis, you must try several regression models and then find the one that mimicks the raw data best. In this particular case, you have already decided to do a linear regression, and that is good for me. I do not have the time to take you through all of them. In this step you must select the exponential regression model.

1. Press [STAT] , highlight [CALC]
2. In the list of regression models select [0:ExpReg]
3. The command ExpReg echoes on main calculator screen.
4. If you press [ENTER] you will get you equation but you will not be able to draw the curve on top of your data to verify the goodness (no it is not a swear word) of the fit. So you must save the equation as a function in Y1 variable.
5. While ExpReg command is still in command line, press the [VARS] key.
6. Highlight the [Y-VARS] menu and select [1:Function], then press [ENTER].
7. In the Y function variable list select [1:Y1]
8. At the command line, you have ExpReg Y1.
9. Press [ENTER] to save the equation of best fit in Y1.

Here are your results. (These are your actual results)

You also have r and its square. You need to read your theory to know which of the two correlation coefficients you have been asked to obtain.

Since you saved the equation in the Y1 function you can draw it. You just have to press GRAPH and it will be drawn. But that will not tell how good the fit is, until you see the function drawn on top of the raw data.

DRAWING THE SCATTER PLOT AND BEST FIT CURVE ON SAME SCREEN

Regardless of the curve of best fit, you can draw a scatter plot as follows.

1. Press [2nd][Y=] (STATPLOT)
2. Select [1:Plot1...Off] and press ENTER to open plot1 configuration screen
3. Use arrow to highlight On and press ENTER
   
4. Use the down arrow to move cursor to the line Type.
5. Select the first icon for a scatter plot. NOTE: To navigate the 6 statistical stat plots on the line Type, use the right or left arrow (to backtrack), but not the Up or down arrows.

Here is the configuration screen

Here plot1 is active (ON), it is scatter plot (first icon) the Xlist is L1 by default but you can change it to some other list. Notice the highlighted A cursor on XLIST line. It means that the [ALPHA] keyboard is active.
On the line Mark you can choose the appearnce of the raw data point.

And here are your two graphs: the raw data scatter plot and the line of best fit on the same screen. On the picture the ZoomStat option was used.


Now, it is up to you to accept this model or to try another model.

Testimonial: "You are brillant. Wish you would write a textbook on how to use this calculator."