~POPULAR TI-83 PLUS MENUS IN STATISTICS~

$counter$

~GENERAL~

1) Selecting a random sample of k integers out of n
USE MATH, PRB, MENU 5, ENTER, INSERT LOWEST#, HIGHEST#, K, ENTER.
Repeatly pressing enter will give you as many groups of k selections as you like.

2) Clearing a list
USE STAT, EDIT, ENTER, MOVE CURSER UP TO TOP OF LIST, PRESS CLEAR, ENTER OR
USE STAT, MENU 4, ENTER, 2ND L₁, ENTER. If clearing more, separate by commas.

3) Entering data into a list
USE STAT, EDIT, ENTER, MOVE CURSER TO THE LIST, ENTER DATA POINTS ONE AT A TIME. DO THE SAME FOR THE OTHER LISTS.

4) clearing the display when pressing clear doesn't work
USE 2ND, QUIT

5) Resetting your calculator when things go "haywire"
USE 2ND, MEM, MENU 7, ENTER, GO TO DEFAULTS, ENTER, GO TO RESET, ENTER

~HISTOGRAMS~ Press for details

~FINDING DESCRIPTIVE STATISTICS~
1) Enter data set in L₁ & use STAT, CALC, 1-VARSTATS, 2ND, L₁, ENTER.
The results will be displayed.

2) For frequency distributions, enter data points in one list and their corresponding frequencies in other list. Then proceed as in 1) STAT, CALC, 1-VARSTATS, 2ND L₁ COMMA 2ND L₂ ENTER.(Assuming the data set is in L₁ and their frequencies are in L₂)

~FINDING STATISTICS FROM PROBABILITY DISTIBUTIONS~

Use The same procedure as above in 2), but enter probabilities instead of frequencies in the list.

~BOXPLOTS~ Press for details

~FINDING COMBINATIONS OF n OBJECTS TAKEN r AT A TIME~

INSERT n FIRST THEN USE MATH, PRB, MENU 3, ENTER, r, ENTER.
To multiply two combinations, press multiply after entering r above, then repeat steps.
To divide two combinations, press divide after entering first r, then repeat steps.

~FINDING BINOMIAL PROBABILITIES~

1) For a fixed number of successes, USE 2ND VARS, MENU 0, ENTER (n,p,r), where n=# of trials, p=the probability of a success, & r= the # of successes.

2) For a cumulative number of successes, USE 2ND VARS, MENU A, ENTER (n,p,r), where r is the upper limit of successes.

Note: Menu A is often used in conjunction with the backdoor approach for problems dealing with "at least r" or "more than r". For these, the upper limit to enter would be r-1 for "at least" & r for "more than" & the result must be subtracted from 1.

This can be done in one step by 1 - (n,p,r-1) or 1 - (n,p,r).

~FINDING AREAS & SCORES UNDER THE STANDARD NORMAL CURVE~

1) For areas (probabilities), insert two z-scores that cover the area. If using x-scores, enter the two x-scores, then the mean, followed by the standard deviation.
USE 2ND VARS, MENU 2, (Z_left,Z_right) or (x_left, x_right, mean, standard deviation).
Note: When working with simple random samples dealing with means, (selecting a sample of n, n>1, divide the standard deviation by the square root of n in the last entry.

2) For finding z-scores or x-scores given the probability (area to the left of the score under the curve) use the following:
2ND VARS, MENU 3, ENTER AREA (PROBABILITY) to get z-score or 2ND VARS, MENU 3, ENTER (AREA, MEAN, STANDARD DEVIATION) to get x-score.

3) For percentile problems, enter percentile (decimal) for the area & follow the same procedure as in 2) above.

~FINDING AREAS UNDER t-CURVES~

USE 2ND VARS, MENU 5, ENTER (t_left, t_right, df), where df= degree of freedom (n - 1).

Note: t-scores are found by USING TABLE A-3, if areas are given.

~FINDING CONFIDENCE INTERVALS~

1) For proportions, USE STAT, TESTS, MENU A

2) For means (known population SD), USE STAT, TESTS, MENU 7

3) For means (unknown population SD), UES STAT, TESTS, MENU 8

Note: The margin of error can be found by subtacting the sample proportion or sample mean from the upper value in the confidence interval.

Note: For sample sizes (n) that are relatively large compared to population size (N),[n > or = .05N], multiply the SD entry by the finite population correction [square root of (N-n)/(N-1)]. This will result in slightly smaller confidence intervals & margins of error.

~HYPOTHESIS TESTING~ Press for details

~CORRELATION & REGRESSION~

1) For correlation, enter paired data into lists, then USE STAT, TESTS, MENU E.

Note: Freq must be 1

Note: Make sure you select r≠0 for the alternate hypothesis

Note: Skip RegEQ & go to Calculate

Note: For correlation, it doesn't matter which column is labeled x or y since, x correlated with y is the same as y correlated with x. Not so for regression.

2) For the regression line, use the same menu.

Note: It does matter which variable is independent (horizontal axis)(usually x), since this gives the regression line "y on x". The regression line "x on y" is the inverse and is different in most cases.

Note: Use table A-6 to check your correlation for significance at the .05 & .01 levels.

Note: Use the regression line for predictions, if the correlation is significant. Otherwise, use the mean for the y-values (for any x).