Why does my TI-84 not show r?

Diagnostics are off. On a TI-84 Plus CE, press MODE, scroll to STAT DIAGNOSTICS, and set it to ON. On older models, press 2nd then 0 for the CATALOG, scroll to DiagnosticOn, press ENTER twice. Run LinReg again and r and r squared appear under the slope and intercept.

How do you find the correlation coefficient on a TI-84?

Enter the two variables in L1 and L2 under STAT, 1:Edit. Turn diagnostics on. Then press STAT, arrow to CALC, choose 4:LinReg(ax+b), set Xlist to L1 and Ylist to L2, and select Calculate. The r value prints with the regression output.

Can a TI-84 calculate Spearman's rank correlation?

Not directly. There is no rank function. You rank each variable by hand into two new lists, giving tied values the average of the ranks they span, then run LinReg on the two rank lists. The r it returns is Spearman's rho.

Does the TI-84 give a p value for the correlation?

Not from LinReg. You can get one with LinRegTTest under STAT, TESTS, which reports t and p alongside r. LinReg by itself stops at the coefficient.

What does ERR:DIM MISMATCH mean?

L1 and L2 hold different numbers of values. Correlation needs paired data, so every x must have a y. Go back into STAT, 1:Edit and check that both lists are the same length.

How to Find the Correlation Coefficient on a TI-84

The short answer

STAT, then 1:Edit. Put one variable in L1 and the other in L2.
Turn diagnostics on, or r will not print. See below.
STAT, arrow right to CALC, choose 4:LinReg(ax+b).
Set Xlist to L1, Ylist to L2, then Calculate.

The screen returns the regression equation, then a, b, r², and r. The last line is your correlation coefficient.

If r is missing, diagnostics are off

This is the whole problem for most people. The TI-84 computes r either way, then declines to display it unless a setting is switched on. You get the slope and intercept, no correlation, and no error message explaining why.

The fix depends on how new the calculator is.

TI-84 Plus CE, and any model on OS 2.55 or later: press MODE, scroll down to STAT DIAGNOSTICS, highlight ON, press ENTER.
Older TI-84 Plus and TI-83 Plus: press 2nd then 0 to open the CATALOG, scroll to DiagnosticOn, press ENTER, then ENTER again to run it. The screen says Done.

Run LinReg again and r² and r now appear at the bottom. The setting persists until the memory is reset, which is why it works on a classmate's calculator and not yours.

Step by step with real numbers

Twelve students, hours studied against exam score.

L1 (hours)	L2 (score)
2	57
3	69
4	56
5	63
6	50
7	75
8	71
9	81
10	69
11	76
12	80
13	77

With diagnostics on, LinReg(ax+b) on L1 and L2 returns:

y=ax+b
a=2.020979021
b=53.50932401
r²=.5183986799
r=.7199990833

So r = .72 and r² = .52. Hours and scores move together, and about 52% of the variance in scores is shared with hours studied.

A note on which LinReg to pick. 4:LinReg(ax+b) and 8:LinReg(a+bx) both work and both give the same r. They differ only in which coefficient is called a and which is called b. If your textbook writes the line as y = a + bx, use option 8 so the labels match.

Spearman's rank correlation on a TI-84

The TI-84 has no rank function and no Spearman command. You rank by hand, then correlate the ranks.

Rank each variable from smallest to largest. Hours are already in order, so L3 is simply 1 through 12.
Rank the scores into L4. The lowest score, 50, gets rank 1. The highest, 81, gets rank 12.
Tied values share the average of the ranks they span. Two students scored 69, occupying ranks 5 and 6, so both get 5.5. This is the step people skip, and it changes the answer.
Run LinReg(ax+b) on L3 and L4.

For this data that gives r_s = .75, slightly higher than Pearson's .72, which is what you expect when one point sits away from the line.

Sorting help: 2nd, STAT, arrow to OPS, 1:SortA( will sort a list ascending so you can read ranks off in order. Sort a copy, not your original paired lists, or you will break the pairing between L1 and L2.

Getting a p value

LinReg stops at the coefficient. An r of .72 from twelve students and an r of .72 from twelve hundred are different findings, and this screen shows the same number for both.

The TI-84 can give you the test, just not from that menu. Press STAT, arrow to TESTS, and choose LinRegTTest. Set Xlist to L1, Ylist to L2, Freq to 1, and leave the alternative at ≠ 0. It returns t, p, df, and r together. For the data above: t = 3.28, df = 10, p = .008.

What it will not give you is a confidence interval for r, which needs a Fisher z transformation the calculator does not carry.

Skip the menus. Paste the same two columns into the correlation calculator and get r, r², the exact p value, a 95% confidence interval, and an APA 7 write-up in one step.

Open the correlation calculator

Errors you will hit

ERR:DIM MISMATCH. L1 and L2 have different lengths. Correlation needs pairs, so every x needs a y. Check both lists in STAT, 1:Edit.
No r, only a and b. Diagnostics are off. See the section above.
ERR:DOMAIN. Usually a list with fewer than two values, or a variable where every entry is identical so there is nothing to correlate.
Old data in the lists. Clear a list with STAT, 4:ClrList, then L1, ENTER. Leftover values from the last problem are a quiet source of wrong answers.

On Desmos instead

If you have a browser rather than the handheld, Desmos will do it. Open a graphing calculator, add a table, put your pairs in x₁ and y₁, then type y_1 ~ mx_1 + b on a new line. Desmos fits the line and reports r² underneath. Take the square root for r, and give it the sign of the slope, since r² alone cannot tell you direction.

Correlation coefficient calculator with p value, confidence interval, and APA 7 output
How to calculate a correlation coefficient in Excel with CORREL and the Analysis ToolPak
Descriptive statistics calculator for means, SD, and skewness
Which statistical test should I use?