TI-84 Plus CE Linear Regression: Get r and r² to Show (Steps)
How to run a linear regression on a TI-84 Plus CE, key by key, and get r and r² on screen. Also covers saving the line to Y1, making predictions, drawing a residual plot and fixing the usual errors.
You run LinReg and the calculator gives you a and b, but no r or r². That's the most common TI-84 regression problem in statistics and algebra classes. It isn't broken. One setting is off.
This page takes you through the whole job key by key: entering data, turning on r and r², running the regression, saving the equation, predicting a value, and checking the residual plot. A worked example with real numbers lets you check your screen against ours.
Keys are written in brackets, like [STAT] or [2nd]. Menu items are shown as they appear on screen, like 4:LinReg(ax+b).
The worked example we'll use
A student records hours studied (x) and test score (y):
- Hours (x): 1, 2, 3, 4, 5, 6
- Score (y): 62, 66, 71, 74, 80, 83
Type these numbers in as you read. Your screen should match the answers below.
Step 1: Clear old data and enter your lists
- Press [STAT], then choose 1:Edit... and press [ENTER].
- If L1 already has numbers in it, arrow up until the L1 heading is highlighted. Press [CLEAR], then [ENTER]. That empties the whole list. (Don't press [DEL] on the heading. That removes the list from the editor.)
- Do the same for L2.
- Arrow into L1 and type each x value, pressing [ENTER] after each one: 1, 2, 3, 4, 5, 6.
- Arrow right to L2 and type the y values: 62, 66, 71, 74, 80, 83.
Check: L1 and L2 must have the same number of entries. If they don't, you'll get ERR:DIM MISMATCH later.
If L1 or L2 is missing from the editor: press [STAT], choose 5:SetUpEditor, press [ENTER] and [ENTER] again. That puts L1 through L6 back.
Step 2: Turn on r and r² (Stat Diagnostics)
You only have to do this once. The setting stays on until a RAM reset.
Method A (most TI-84 Plus CE calculators with recent OS versions):
- Press [MODE].
- Arrow down to STAT DIAGNOSTICS.
- Highlight ON and press [ENTER].
- Press [2nd] [MODE] (QUIT) to go back to the home screen.
Method B (works on any TI-84 model):
- Press [2nd] [0] to open the CATALOG.
- Press [x⁻¹]. Alpha-lock is already on in the catalog, and that key types D, so you jump to the D section.
- Scroll down to DiagnosticOn and press [ENTER].
- Press [ENTER] again. The screen should say Done.
Turning diagnostics on doesn't change any results you already have. Run the regression again to see r and r².
Step 3: Run the regression
Pick the right form. Under [STAT] then CALC, there are two linear regressions:
- 4:LinReg(ax+b) gives y = ax + b, where a is the slope.
- 8:LinReg(a+bx) gives y = a + bx, where b is the slope. AP Statistics usually writes the line this way.
Both give the same line. Only the letters change. Watch which letter is the slope when you write your answer.
With Stat Wizards on (the default on most CE calculators):
- Press [STAT], arrow right to CALC, and choose 4:LinReg(ax+b).
- A wizard screen opens. Set it up like this:
- Xlist: L1 (type [2nd] [1] if it isn't already there)
- Ylist: L2 ([2nd] [2])
- FreqList: leave blank
- Store RegEQ: put your cursor here, press [VARS], arrow right to Y-VARS, choose 1:Function, then 1:Y1. This saves the line in Y1 so you can graph it and use it to predict.
- Arrow down to Calculate and press [ENTER].
With Stat Wizards off (you only see "LinReg(ax+b)" on the home screen):
Type the arguments yourself, then press [ENTER]: LinReg(ax+b) L1, L2, Y1
To get them: L1 is [2nd] [1], the comma is the [,] key, L2 is [2nd] [2], and Y1 is [VARS], then Y-VARS, then 1:Function, then 1:Y1. To turn the wizards on or off, go to [MODE] and find STAT WIZARDS.
Step 4: Read the output
For our example, the screen shows:
- y=ax+b
- a=4.285714286
- b=57.66666667
- r²=.9941089838
- r=.9970501411
How to write the answer: predicted score = 57.67 + 4.29(hours), with r ≈ 0.997 and r² ≈ 0.994.
What the numbers mean in context:
- Slope (a = 4.29): for each extra hour studied, the predicted score goes up by about 4.29 points.
- Intercept (b = 57.67): the predicted score for 0 hours of study. That value is outside our data range, so be careful reading too much into it.
- r = 0.997: a very strong, positive linear relationship.
- r² = 0.994: about 99.4% of the variation in score is explained by the linear relationship with hours studied.
Sign check: r always has the same sign as the slope. If your slope is negative and r is positive, you've mixed up your lists.
Step 5: Predict a value using Y1
You stored the line in Y1, so you don't need to retype the equation.
- On the home screen, press [VARS], arrow right to Y-VARS, choose 1:Function, then 1:Y1.
- Type ( 7 ) so the screen reads Y1(7).
- Press [ENTER].
Result: 87.66666667. A student who studies 7 hours is predicted to score about 87.7.
Seven hours is just past the data (1 to 6 hours). Predicting far outside the data range is extrapolation, and teachers often ask you to say so.
Step 6: Graph the scatter plot and the line
- Press [2nd] [Y=] (STAT PLOT) and choose 1:Plot1.
- Highlight On and press [ENTER].
- Type: choose the first icon (scatter plot).
- Xlist: L1. Ylist: L2.
- Press [ZOOM] and choose 9:ZoomStat. The window fits your data, and Y1 is drawn through the points.
Use [TRACE] and the up/down arrows to move between the points and the line.
Step 7: Make a residual plot (to check whether linear fits)
The calculator saves residuals in a list called RESID every time you run a regression. Run the regression first, then:
- Press [2nd] [Y=] and choose 1:Plot1.
- Keep Xlist as L1. Move to Ylist, clear it, then press [2nd] [STAT] (LIST). Under NAMES, scroll to RESID and press [ENTER].
- Press [Y=] and turn Y1 off so the line doesn't cover the plot. Arrow onto the = sign next to Y1 and press [ENTER] to unhighlight it.
- Press [ZOOM] then 9:ZoomStat.
For our example, the residuals are roughly 0.05, −0.24, 0.48, −0.81, 0.90 and −0.38. They bounce above and below zero with no curve, so a linear model fits.
What to look for: random scatter around zero means linear is reasonable. A clear U shape or curve means a linear model isn't a good fit.
To see the residual numbers, go to [STAT], 1:Edit, arrow to an empty list heading, and type the name RESID.
Fixing the most common problems
- No r or r² on screen: Stat Diagnostics is off. Do Step 2, then run the regression again.
- ERR:DIM MISMATCH: L1 and L2 have different numbers of entries. Check the bottom of each list in the editor.
- ERR:INVALID DIM when graphing: a stat plot is pointing to an empty or mismatched list. Turn off any plots you aren't using: [2nd] [Y=], 4:PlotsOff, [ENTER].
- Weird answers from an old problem: old numbers are still in the lists. Clear them from the heading as in Step 1.
- Line doesn't appear on the graph: you didn't fill in Store RegEQ, or Y1 is turned off in [Y=].
- Slope and intercept seem swapped: you used 8:LinReg(a+bx) and read it as ax+b, or the other way around.
- QuadReg shows R² but no r: that's normal. For non-linear models the calculator gives R² (the coefficient of determination) instead of r.
Quick-reference card (copy this into your notes)
- [STAT] 1:Edit: x values in L1, y values in L2
- Diagnostics on (once): [MODE] then STAT DIAGNOSTICS ON, or CATALOG then DiagnosticOn
- [STAT] then CALC then 4:LinReg(ax+b): Xlist L1, Ylist L2, Store RegEQ Y1 ([VARS] Y-VARS 1 1), Calculate
- Read a, b, r², r and write the equation in context
- Predict: Y1(value) [ENTER]
- Scatter plot: [2nd] [Y=] Plot1 On, then [ZOOM] 9
- Residuals: Plot1 Ylist = RESID, then [ZOOM] 9
This is an independent guide. It isn't affiliated with or endorsed by Texas Instruments. Menu wording can vary a little between OS versions. If a menu item looks different on your calculator, look for the closest match on the same menu.
CalcFlip DIY Kit: Find, Test and Flip One Used TI-84 Plus CE Graphing Calculator on eBay · $15.99
Get it on Gumroad