How to do linear, exponential, power, polynomial, sine and logistic regression on the Casio fx-CG50, TI-84 Plus CE and TI-Nspire CX II
On the fx-CG50: MENU → 2 (STATISTICS) … F3 (REG) → F1 (X) → F1 (ax+b).On the TI-84 Plus CE: stat → CALC → 4: LinReg(ax+b).On the TI-Nspire CX II: menu → Statistics → Stat Calculations → Linear Regression.
Least-squares lines and curves: linear, exponential, power, quadratic, cubic, sinusoidal and logistic models. Exact keystrokes for the fx-CG50 first, then the other calculators, a worked example you can check your answers against, the mistakes that cost marks, and the same thing in Desmos.
When you need it
AI students fit models in almost every Paper 2: the regression line of y on x and predictions from it (SL), and exponential, power, quadratic, cubic, sinusoidal and logistic models (HL, plus the coefficient of determination R²). AA students need the linear regression line and r. Quote the equation with its coefficients to 3 s.f.
Practise it: SL AI bivariate data · SL AI exponential modelling · SL AI sinusoidal modelling · HL AI non-linear regression · SL AA bivariate data.
Which IB course and paper
| Course | GDC papers | What you use it for |
|---|---|---|
| AA SL | Paper 2 (Paper 1 is calculator-free) | The regression line of y on x (and of x on y), Pearson's r and predictions. |
| AA HL | Papers 2 and 3 (Paper 1 is calculator-free) | As SL. |
| AI SL | Papers 1 and 2 | The regression line of y on x, r, and predictions from it. |
| AI HL | Papers 1, 2 and 3 | Also non-linear regression (quadratic, cubic, exponential, power and sine models) and the coefficient of determination R². |
A GDC is allowed on every AI paper and on AA Papers 2 and 3; AA Paper 1 is calculator-free, so learn the method by hand as well.
Topic notes: Correlation, regression, and Pearson's r (AI SL) · Exponential growth and decay models (AI SL) · Sinusoidal models for periodic data (AI SL) · Linear regression and residual analysis (AI HL) · Sinusoidal modelling with seasonal data (AI HL) · Exponential models for real-world decay (AI HL).
Keystrokes on your calculator
Pick your calculator: the page remembers it next time.
Regression on the Casio fx-CG50
Written for OS 3.80 (menus unchanged from OS 3.60 onwards as far as we can tell). Casio fx-CG50: graphic (GDC).
Watch the key presses on the fx-CG5015 presses, one at a time
The walkthrough needs JavaScript. The same key presses are in the numbered steps below.
- MENU 2 (STATISTICS). Type the x values into List 1 and the y values into List 2, pressing EXE after each. (Clear an old list first: highlight the list, F6 (▷) F4 (DELETE-ALL) F1 (Yes).)
- F2 (CALC) F6 (SET): 2Var XList = List1, 2Var YList = List2, 2Var Freq = 1. EXIT.
- F3 (REG) F1 (X) F1 (ax+b). The screen shows a, b, r, r² and MSe.
- Other models from the REG menu: F3 (X²) quadratic, F4 (X³) cubic; F6 (▷) then F1 (Log), F2 (Exp: F1 ae^bx or F2 ab^x), F3 (Pwr) axᵇ, F4 (Sin), F5 (Lgst) logistic.
- F5 (COPY) pastes the equation into the GRAPH Y= list; F6 (DRAW) plots it over the scatter diagram.
- Sin regression: set SHIFT MENU (SET UP) → Angle = Rad first. The model is y = a·sin(bx + c) + d.
- Logistic: y = c ÷ (1 + a·e^(−bx)). IB writes it L ÷ (1 + Ce^(−kx)), so L = c, C = a, k = b.
- Quadratic, cubic and the curve models show R² (the coefficient of determination), not r.
- Predict from the stored model: in RUN-MATRIX use OPTN F5 (STAT) → ŷ (7.5 F2 ŷ in linear mode).
Regression on the TI-84 Plus CE
Written for OS 5.8.x. TI-84 Plus CE: graphic (GDC); also the TI-84 Plus CE Python / CE-T Python Edition.
Watch the key presses on the TI-84 Plus CE15 presses, one at a time
The walkthrough needs JavaScript. The same key presses are in the numbered steps below.
- stat 1: Edit. Clear old data: move onto the L1 heading, clear, enter. Type the x values in L1 and the y values in L2.
- Turn diagnostics on once: mode → STAT DIAGNOSTICS: ON (on older OS versions: 2nd 0 (CATALOG) → DiagnosticOn → enter enter). Without it r and r² are not shown.
- stat → CALC → 4: LinReg(ax+b). Xlist: L1, Ylist: L2, Store RegEQ: Y1 (alpha trace → Y1). Calculate.
- Other models in the same CALC list: 5: QuadReg, 6: CubicReg, 9: LnReg, 0: ExpReg (y = a·bˣ), A: PwrReg, B: Logistic, C: SinReg.
- SinReg gives y = a·sin(bx + c) + d and works in radians.
- Logistic gives y = c ÷ (1 + a·e^(−bx)).
- Storing to Y1 lets you predict with Y1(7.5) on the home screen.
Regression on the TI-Nspire CX II
Written for OS 6.x. TI-Nspire CX II: graphic (GDC), non-CAS (the CX II CAS is a different model: check the exam rules before using one).
- Add a Lists & Spreadsheet page (ctrl doc → Add Lists & Spreadsheet). Name column A xs and column B ys in the top row, then type the data.
- menu → Statistics → Stat Calculations → Linear Regression (mx+b). X List: xs, Y List: ys, Save RegEqn to: f1. OK.
- The results (m, b, r², r) appear in the next empty columns.
- Other models in the same menu: Quadratic, Cubic, Power, Exponential (y = a·bˣ), Logarithmic, Sinusoidal, Logistic (d = 0).
- Predict with f1(7.5) on a Calculator page.
- Logistic (d = 0) is y = c ÷ (1 + a·e^(−bx)).
For teachers: Show on the board Print a keystroke card Teaching GDC skills
Worked example
(a) x = 1, 2, 3, 4, 5, 6; y = 3.1, 4.9, 7.2, 8.8, 11.1, 12.9. Find the regression line of y on x, r, and predict y when x = 7.5. (b) Same x, y = 2.0, 3.1, 4.4, 6.6, 9.8, 14.5. Fit y = a·bˣ and predict y at x = 8.
Answers (checked with Python)
(a) y = 1.97714x + 1.08, r = 0.999191; at x = 7.5, y ≈ 15.9. (b) y = 1.37065 × 1.48177ˣ (equivalently y = 1.37065e0.393240x); at x = 8, y ≈ 31.9 (an extrapolation, so less reliable).
Now try it on an exam-style question
Use the keystrokes above on IB-style correlation and regression questions. Free to try, no sign-up.
Common calculator mistakes
- Mixing up a and b: fx-CG50 ax+b and TI LinReg(ax+b) use a for the gradient; the ClassWiz and TI LinReg(a+bx) use b for the gradient.
- Leaving old data in the lists (or a frequency list set to List 3 from a previous question).
- Regression of x on y when the question wants y on x: the x values go in the first list.
- Quoting r for a non-linear model: for quadratics and curves use R².
- Sine regression in degree mode gives nonsense coefficients.
- Predicting far outside the data range (extrapolation) without saying it may be unreliable.
- Predicting x from the line of y on x: AA needs the regression line of x on y for that (swap the lists).
- Writing the model with rounded coefficients and then predicting from them: predict from the stored equation, and round the prediction.
- Accuracy: unless the question says otherwise, IB markschemes want answers exact or to 3 significant figures. Keep the full calculator value for later parts and round only the answer you write down (x = 2.094551… is written 2.09, not 2.1 or 2.0946).
If you get an error
fx-CG50: Ma ERROR
Ma ERROR Casio's Math ERROR: the result is outside the allowable calculation range, or you are trying to do an illegal mathematical operation, such as dividing by zero. Check: Look for a zero denominator, sin⁻¹ or cos⁻¹ of a number bigger than 1 (often a wrong side or angle in the sine or cosine rule), the log of 0 or of a negative number, and a value stored in a variable earlier that you forgot about. Fix: Casio's advice: check your input values and change them so the result is within the allowable range.
TI-84 Plus CE: ERR:DOMAIN, ERR:DIM MISMATCH, ERR:INVALID DIM, ERR:SYNTAX, ERR:STAT
ERR:DOMAIN A value given to a function or instruction is outside the values it accepts. In a regression, TI says it means the data do not suit the model you chose. Check: Is every input allowed? σ must be positive, p between 0 and 1, and sin⁻¹ or cos⁻¹ needs a number from −1 to 1. Exponential and power models need positive data. Fix: Change the value that is out of range, or choose a model that suits the data.
ERR:DIM MISMATCH Lists or matrices in the same operation have different sizes (TI's example: multiplying two lists with different numbers of elements). Check: L1 and L2 with different numbers of values (one value missed or typed twice), a frequency list of the wrong length, or matrices whose sizes cannot be multiplied. Fix: Make the lists the same length, or check the size of each matrix.
ERR:INVALID DIM TI's knowledge base: one or more stat plots are turned on, but the lists they use are empty. Check: A stat plot left on from a statistics question, so the error appears when you graph a function. Fix: Turn off the stat plots you are not using, or put the data back in the lists.
ERR:SYNTAX The command contains a syntax error: a misplaced function, argument, bracket or comma. Check: TI points out that subtract − and negative (−) are different keys, and using the wrong one causes this error. Then check the brackets and commas. Fix: Find the misplaced function, argument, bracket or comma and correct it.
ERR:STAT A statistics calculation was tried on lists that are not suitable. A statistical analysis needs at least two data points. Check: An empty or one-value list, or the wrong list named in the command. Fix: Check the lists named in the command and the data in them.
TI-Nspire CX II: Domain error, Dimension mismatch, Syntax error
Domain error An argument must be in a specified domain, and this one is not (TI's example: rand(0)). Check: Is every input allowed? σ must be positive, p between 0 and 1, and sin⁻¹ or cos⁻¹ needs a number from −1 to 1. Fix: Change the value that is out of range.
Dimension mismatch Two or more arguments must have the same dimension and do not (TI's example: [1,2]+[1,2,3]). Check: Lists of different lengths in a statistics command, or vectors or matrices of different sizes. Fix: Make the lists, vectors or matrices the sizes the operation needs.
Syntax error The syntax checker found a problem. Check: A missing bracket or comma. Fix: Correct the expression and try again.
Do it in Desmos
Not in the exam. Desmos is not an approved calculator for IB exams, so never rely on it for Paper 2. It is excellent for learning, checking a GDC answer and building graphs for your Internal Assessment (IA).
- Add a table (+ → table) with the data in x₁ and y₁.
- Type y₁ ~ m x₁ + b. Desmos shows m, b, r and R² and draws the line.
- For an exponential model type y₁ ~ a b^{x₁}. Desmos fits the curve directly; your GDC fits a straight line to ln y. Turn on "log mode" in the regression's options to get the GDC's answer.
- Sinusoidal and logistic work the same way: y₁ ~ a sin(b x₁ + c) + d, y₁ ~ L/(1 + C e^{−k x₁}).
Open desmos.com/calculator (free, in any browser) and type:
table: x_1 = 1, 2, 3, 4, 5, 6; y_1 = 3.1, 4.9, 7.2, 8.8, 11.1, 12.9y_{1}~ mx_{1}+b
Questions
Can I use a calculator for regression in the exam?
In IB Maths you can use one approved GDC on every AI paper and on AA Papers 2 and 3, in exam mode. AA Paper 1 is calculator-free, so learn the method by hand as well.
What are the fx-CG50 keystrokes for regression?
MENU 2 (STATISTICS). Type the x values into List 1 and the y values into List 2, pressing EXE after each. (Clear an old list first: highlight the list, F6 (▷) F4 (DELETE-ALL) F1 (Yes).). Then F2 (CALC) F6 (SET): 2Var XList = List1, 2Var YList = List2, 2Var Freq = 1. EXIT. Then F3 (REG) F1 (X) F1 (ax+b). The screen shows a, b, r, r² and MSe. The full steps and the other calculators are in the keystrokes section above.
Does this work on other calculators?
Yes. This page has keystrokes for the Casio fx-CG50, TI-84 Plus CE and TI-Nspire CX II. The worked example has the same answers on every model, so use it to check.
What should I write as working?
Write the maths, not the key presses: say what you are finding in normal notation (the equation, the probability statement or the integral), then the calculator's answer exact or to 3 significant figures unless the question says otherwise. Keep full accuracy in later parts.
More calculator skills
Sources
- Casio fx-CG50 Software User's Guide, OS 3.70 (PDF) : Casio fx-CG50, OS 3.80 (menus unchanged from OS 3.60 onwards as far as we can tell).
- TI-84 Plus CE guidebook (Texas Instruments) : TI-84 Plus CE, OS 5.8.x.
- TI-Nspire CX II guidebooks (Texas Instruments) : TI-Nspire CX II, OS 6.x.
- Desmos Help Center for the Desmos functions.
Menu names are taken from each manufacturer's manual for the models shown. Menus can move slightly between OS updates: if a screen looks different on yours, the manual for your version is the reference, and please tell us so we can update this page. A calculator with no tab here has no checked steps for this skill yet. Worked-example answers were recomputed independently in Python.