Showing posts with label calculators. Show all posts
Showing posts with label calculators. Show all posts
Tuesday, November 3, 2015
The frequency table solution for TI-83 and TI-84
Put the list of numbers in one list (let's say L1) and the frequencies in a separate list (for simplicity's, make it L2 in this example.) Under the STAT menu, in the CALC sub-menu, choose
2-Var Stats L1, L2
Assuming the frequencies are in the second list, the sum of the y values is n, the size of the sample and the sum of the xy is the sum of all the x values. You then need to divide (sum of xy)/(sum of y) to get the average x-bar.
As for the median, you have to do it by hand, just like the folks with the TI-30xIIs.
Here are the lists from class, the first number x and the second number f(x). The answers are in the comments.
This year's daily differences, 2015 - (avg. 1999-2014)
-9, 1
-8, 3
-7, 4
-6, 7
-5, 13
-4, 15
-3, 12
-2, 16
-1, 11
0, 24
1, 27
2, 24
3, 24
4, 25
5, 18
6, 14
7, 11
8, 8
9, 14
10, 5
11, 7
12, 1
13, 1
14, 4
15, 2
16, 3
17, 5
18, 4
20, 2
21, 1
Here is the list of wins by NFL teams in 2014
12, 5
11, 4
10, 3
9, 4
8, 2
7, 4
6, 3
5, 1
4, 2
3, 2
2, 2
Answers in the comments.
Friday, June 26, 2009
Standard deviation using the TI-30x IIs
With this list, we can use the TI-30x IIs to get the average and both standard deviations, sx, the standard deviation for a sample and sigmax, the standard deviation for a population.
The buttons to press will be written in red.
Here is a data set we will input.
60, 60, 60, 61, 62, 62, 63, 63,64, 65, 65, 66, 66, 66, 66, 66, 66, 66, 66,
68, 68, 69, 69, 69, 70, 70, 71, 71, 71, 72, 72, 74, 76, 77, 78
[2nd][DATA][ENTER] This puts the calculator in one variable mode.
[2nd][DATA][LEFT][ENTER] This clears the data.
[DATA]
X1= 60 [DOWN]
FRQ= 3 [DOWN]X2= 61 [DOWN]
FRQ= 1 [DOWN]X3= 62 [DOWN]
FRQ= 2 [DOWN]X4= 63 [DOWN]
FRQ= 2 [DOWN]X5= 64 [DOWN]
FRQ= 1 [DOWN]X6= 65 [DOWN]
FRQ= 2 [DOWN]X7= 66 [DOWN]
FRQ= 8 [DOWN]X8= 68 [DOWN]
FRQ= 2 [DOWN]X9= 69 [DOWN]
FRQ= 3 [DOWN]X10= 70 [DOWN]
FRQ= 2 [DOWN]X11= 71 [DOWN]
FRQ= 3 [DOWN]X12= 72 [DOWN]
FRQ= 2 [DOWN]X13= 74 [DOWN]
FRQ= 1 [DOWN]X14= 76 [DOWN]
FRQ= 1 [DOWN]X15= 77 [DOWN]
FRQ= 1 [DOWN]X16= 78 [DOWN]
FRQ= 1 [DOWN]
[STATVAR] This shows the statistics variables.
n x-bar sx sigmax
35[RIGHT]n x-bar sx sigmax
67.37142857[RIGHT]n x-bar sx sigmax
4.747046848[RIGHT]n x-bar sx sigmax
4.67870455[RIGHT]
The buttons to press will be written in red.
Here is a data set we will input.
60, 60, 60, 61, 62, 62, 63, 63,64, 65, 65, 66, 66, 66, 66, 66, 66, 66, 66,
68, 68, 69, 69, 69, 70, 70, 71, 71, 71, 72, 72, 74, 76, 77, 78
[2nd][DATA][ENTER] This puts the calculator in one variable mode.
[2nd][DATA][LEFT][ENTER] This clears the data.
[DATA]
X1= 60 [DOWN]
FRQ= 3 [DOWN]X2= 61 [DOWN]
FRQ= 1 [DOWN]X3= 62 [DOWN]
FRQ= 2 [DOWN]X4= 63 [DOWN]
FRQ= 2 [DOWN]X5= 64 [DOWN]
FRQ= 1 [DOWN]X6= 65 [DOWN]
FRQ= 2 [DOWN]X7= 66 [DOWN]
FRQ= 8 [DOWN]X8= 68 [DOWN]
FRQ= 2 [DOWN]X9= 69 [DOWN]
FRQ= 3 [DOWN]X10= 70 [DOWN]
FRQ= 2 [DOWN]X11= 71 [DOWN]
FRQ= 3 [DOWN]X12= 72 [DOWN]
FRQ= 2 [DOWN]X13= 74 [DOWN]
FRQ= 1 [DOWN]X14= 76 [DOWN]
FRQ= 1 [DOWN]X15= 77 [DOWN]
FRQ= 1 [DOWN]X16= 78 [DOWN]
FRQ= 1 [DOWN]
[STATVAR] This shows the statistics variables.
n x-bar sx sigmax
35[RIGHT]n x-bar sx sigmax
67.37142857[RIGHT]n x-bar sx sigmax
4.747046848[RIGHT]n x-bar sx sigmax
4.67870455[RIGHT]
Monday, May 18, 2009
inputting the data for the worksheet using the TI-30XIIs
To get into the correct mode, follow these instructions.
[2nd][data]
move the underline so it is under 2-var, then press [enter].
[2nd][data]
move the underline so it so under clrdata, then press [enter].
now we enter the data.
[data]
x1 = 856.7[down]
y1 = 15.22[down]
x2 = 907.9[down]
y2 = 16.79[down]
x3 = 974.1[down]
y3 = 19.81[down]
x4 = 930.9[down]
y4 = 17.88[down]
x5 = 885.9[down]
y5 = 16.84[down]
x6 = 886.1[down]
y6 = 16.87[down]
x7 = 926.8[down]
y7 = 17.48[down]
x8 = 928.4[down]
y8 = 17.34[down]
x9 = 802.9[down]
y9 = 12.22[down]
x10 = 834.8[down]
y10 = 11.16[down]
x11 = 734.3[down]
y11 = 9.37[down]
x12 = 816.3[down]
y12 = 10.26[down]
x13=[statvar]
The numbers you need are
rx,y = .946053079... ~= .946 (Above the 99% threshold of .708)
a = 0.048098608... ~= .0481
b = -26.92322636... ~= -26.9232
So the formula to find the points on the line is yp = ax + b if you use the exact values and if using the approximation, yp = .0481x - 26.9232.
Answer for both versions to the nearest penny in the comments.
[2nd][data]
move the underline so it is under 2-var, then press [enter].
[2nd][data]
move the underline so it so under clrdata, then press [enter].
now we enter the data.
[data]
x1 = 856.7[down]
y1 = 15.22[down]
x2 = 907.9[down]
y2 = 16.79[down]
x3 = 974.1[down]
y3 = 19.81[down]
x4 = 930.9[down]
y4 = 17.88[down]
x5 = 885.9[down]
y5 = 16.84[down]
x6 = 886.1[down]
y6 = 16.87[down]
x7 = 926.8[down]
y7 = 17.48[down]
x8 = 928.4[down]
y8 = 17.34[down]
x9 = 802.9[down]
y9 = 12.22[down]
x10 = 834.8[down]
y10 = 11.16[down]
x11 = 734.3[down]
y11 = 9.37[down]
x12 = 816.3[down]
y12 = 10.26[down]
x13=[statvar]
The numbers you need are
rx,y = .946053079... ~= .946 (Above the 99% threshold of .708)
a = 0.048098608... ~= .0481
b = -26.92322636... ~= -26.9232
So the formula to find the points on the line is yp = ax + b if you use the exact values and if using the approximation, yp = .0481x - 26.9232.
Answer for both versions to the nearest penny in the comments.
Labels:
calculators,
correlation,
two variable statistics
Tuesday, April 28, 2009
Class Notes for 4/27 and 4/29
Dependence and Independence
The first line of Leo Tolstoy's Anna Karenina is "All happy families are alike; each unhappy family is unhappy in its own way." In statistics, all independent trials are alike, in that the probability of a particular outcome of one trial does not effect the outcome of later trials, nor was it effected by earlier trials. With dependent probability, the outcome of one trial is effected by the outcome of previous trials, but how that effects things is not always the same.
For example, if we talk about a 70% free throw shooter taking two shots, does missing the first shot effect the probability of missing the second shot? Let's look at this simple problem three different ways.
Predicting the future mathematically by carefully studying the past: Let's say I called this person a 70% free throw shooter because so far in the season, she has made 7 of 10 shots from the line. If she misses, she has now made 7 of 11 shots from the line, and she is now a 63.6% shooter. Should we factor that in to be more precise? Some mathematical models would say yes.
Using this method, missing a shot would make her percentage worse and making one would make it better, no matter how many shots she had taken. But if she had made 70 of 100 so far in the season, one miss would make her 70 of 101, which would lower her percentage to 69.3%, a much smaller effect than it has if she had made 7 of 10. If instead we were looking at her entire career instead a single season, perhaps she has made 700 of 1000, and missing one would mean she was 700 of 1001, which changes her percentage to 69.9%. When discussing free throw percentage, announcers on TV usually round to the nearest percent, so the first example of 7 of 10 to 7 of 11 would be a drop from 70% to 64%, the second example would be a drop from 70% to 69%, and in the third example, 69.9% would round to 70% and the change would be too small to notice.
Factors effecting success and failure: The free throw shooter goes to the line for two shots and misses the first. Is there a physical reason? We might treat free throw shooting as we would picking a random number from 1 through 100, where we count any number from 1 through 70 as a success and 71 through 100 as failures, but shooting free throws includes the human factor. Maybe she missed because she is nervous or distracted. Maybe it's late in the game and she is tired or injured, changing her technique. If any of these are the case, it might make more sense for us to downgrade the probability of making the next shot, though exactly how much it should be downgraded is no longer some simple formula like turning a fraction into a percentage.
Compensating for failure: Again, let's add in the human factor in this problem. She misses the first free throw, and her coach notices her technique looks inconsistent. "Elbow up!" the coach shouts from the sidelines, and the shooter hears the coach and readjusts her technique to match the way she shoots in practice. Should this change bring her back to being a 70% shooter, or even upgrade her chance of success? That is uncertain, but the failure on the first shot and the diagnosis of at least one reason for the failure could effect the odds, and that effect means the second shot should not be considered independent of the first.
The dependent probabilities in a 52 card deck
One of the simplest mathematical models of dependency is sampling without replacement, which is the way most card games or lotteries or the game of Bingo works. You have a set of outcomes which get effectively randomized and a trial is performed, meaning a card is taken from the deck or a ping pong ball is removed from the hopper or a bingo marker is removed from the spinner. Once removed, the number of possible outcomes has been reduced by one and probabilities for success and failure of certain outcomes change.
Looking for an ace: There are 52 cards is a standard deck and 4 of them are aces. If I draw a card from a randomized deck, the chances are 4/52 = 1/13 ~= 7.7% that the card will be an ace. What are the chances the second card is an ace?
That depends on the first card.
Probability that the second card is an ace, given the first card is an ace is 3/51 = 1/17 ~= 5.9%.
Probability that the second card is an ace, given the first card is not an ace is 4/51 ~= 7.8%.
Unlike the mathematical model of free throw shooting where we re-calculate the probabilities by adding the most recent make or miss into the precentage, which means a miss brings the odds down and a make brings the odds up, not getting an ace makes the odds a little better next time, and getting an ace makes the odds worse.

This is the formula for the dependent probability model of sampling without replacement is given at the left. The two numbers in parentheses are a binomial coefficient, the numbers you get when you use nCr on your calculator, which I pronounce "n choose r" in class. The pairs of numbers that look like a base and an exponent, except that the exponent is underlined, is the convention developed by Donald Knuth at Stanford for writing the numbers that you get on your calculator using the nPr function, which I pronounce "n fall r", referring to the name "the falling factorial". If we think about a deck of cards, the lowercase letters refer to the size of the hand n, where r is the number of successful trials (r for right) and w is the number of unsuccessful trials (w for wrong), and r+w=n. The uppercase letters refer to the size of the deck, where T is the size of the deck, G is the number of cards we consider success if we draw them and B is the number of cards we consider a failed trial if we draw them. The letter T stands for Total, G for Good and B for Bad. Again, we have an equation, G+B=T.
Example: If we want consider drawing a heart a success and anything else a failure, what is the probability of drawing three hearts and two non hearts in a five card hand from a well-shuffled 52 card deck.
Here are the six numbers we need.
n = 5
r = 3
w = 2
T = 52
G = 13
B = 39
On a TI-30XIIs, here are the keys you would press.
5[prb][right]3×13[prb]3×39[prb]2÷52[prb]5[enter]
The calculator will read as follows.
5 nCr 3*13 nPr 3*39 nPr 2/52 nPr 5
0.081542617
This means the probability of exactly three hearts and two cards of some other suit is about 8.15%.
The Expected Value (EV) of a two outcome game
Let us assume we have a game that has only two outcomes, winning and losing. Let us further assume that two players have decided to wager on this game, both putting money into a pooled amount and the winner taking all at the end.
If we look at the game from the point of view of one of the players, we need to know the probability of winning p, how much that player put in, which we call Risk and how much the opponent put in, called Profit. The expected value EV equals the probability of victory p times the sum of Profit and Risk divided by Risk.
Different books use different formulas for this. Some do not divide by Risk. By dividing, the number we get is a percentage of return, and a game of flipping coins for $1 a game is equivalent to a game of flipping coins for $100 a game. Some subtract 1 from this formula. This just changes the most important number in identifying results from 1 to 0.
If EV = 1, we consider this a "fair game". For every $1 risked on this game, the expected value is that you will have that dollar returned to you, breaking even. Notice if we are flipping coins, that event never happens on any single play. Either the player makes a dollar profit or a dollar loss, but expected value is about the long run.
If EV > 1, the game is advantageous to the player. If EV < 1, the game is disadvantageous to the player.
In the game of roulette, there are 38 slots where the ball can land, and for simplicity's sake we will assume each has an equal chance of showing up, so p = 1/38. For every $1 you risk, you can make a profit of $35 if you correctly guess in the exact slot where the ball will land. To find the expected value using the TI-30xIIs, you should type in this.
1÷38×(35+1)÷1[enter]
The calculator will read as follows.
1/38*(35+1)/1
0.947368421
What this number means is that for every dollar risked on the spin of a roulette wheel, you should expect about 94.7 cents returned to you in change. In other words, about 5.3 cents is lost from every dollar you bet on every spin of the wheel.
Another way to play the game is to bet red or black. Of the 38 compartments, 18 are red and 18 are black and 2 are green. The probability of victory on betting one of the two major colors is 18/38 = 9/19 ~= 0.473684211. The profit and risk are now both $1. Here's what to type on the TI-30xIIs.
16÷38×(1+1)÷1[enter]
The calculator will read as follows.
16/38*(1+1)/1
0.947368421
The game has changed, both in probability and amount of profit compared to risk, but from the player's point of view, the expected value is precisely the same and still in favor of the casino.
No matter what the levels of profit and risk are, we can find a probability p that will make the expected value equal to 1, and that is p = Risk/(Profit+Risk). If the probability is increased with the profit and risk remaining unchanged, the game becomes advantageous. If is is decreased, the game becomes disadvantageous.
Modern and Classic Parimutuel odds
Profit and risk are listed either in classic form like 3-1 or 2-7 (or sometimes with colons 3:1 or 2:7), where profit is the first number and risk is the second.
In online betting sites, the numbers are given as numbers with absolute value greater than 100, with either a + or - in front of them. +250 means 250 is the profit and 100 is the risk, while -250 means 10o is the profit and 250 is the risk. The fourth page of the yellow sheet explains this in greater detail and shows how to switch back and forth between the two systems.
Practice problems
1. With a well-shuffled 52 card deck, find the probability of getting exactly r hearts in a five card hand when
a) r = 0
b) r = 1
c) r = 2
d) r = 3 (already solved above)
e) r = 4
f) r = 5
2. Find the break-even p when Profit and Risk are as given. Round to three places after the decimal point.
a) Modern parimutuel = +150
b) Modern parimutuel = -110
c) Classic parimutuel = 5:3
d) Classic parimutuel = 5:11
Answers in the comments.
The first line of Leo Tolstoy's Anna Karenina is "All happy families are alike; each unhappy family is unhappy in its own way." In statistics, all independent trials are alike, in that the probability of a particular outcome of one trial does not effect the outcome of later trials, nor was it effected by earlier trials. With dependent probability, the outcome of one trial is effected by the outcome of previous trials, but how that effects things is not always the same.
For example, if we talk about a 70% free throw shooter taking two shots, does missing the first shot effect the probability of missing the second shot? Let's look at this simple problem three different ways.
Predicting the future mathematically by carefully studying the past: Let's say I called this person a 70% free throw shooter because so far in the season, she has made 7 of 10 shots from the line. If she misses, she has now made 7 of 11 shots from the line, and she is now a 63.6% shooter. Should we factor that in to be more precise? Some mathematical models would say yes.
Using this method, missing a shot would make her percentage worse and making one would make it better, no matter how many shots she had taken. But if she had made 70 of 100 so far in the season, one miss would make her 70 of 101, which would lower her percentage to 69.3%, a much smaller effect than it has if she had made 7 of 10. If instead we were looking at her entire career instead a single season, perhaps she has made 700 of 1000, and missing one would mean she was 700 of 1001, which changes her percentage to 69.9%. When discussing free throw percentage, announcers on TV usually round to the nearest percent, so the first example of 7 of 10 to 7 of 11 would be a drop from 70% to 64%, the second example would be a drop from 70% to 69%, and in the third example, 69.9% would round to 70% and the change would be too small to notice.
Factors effecting success and failure: The free throw shooter goes to the line for two shots and misses the first. Is there a physical reason? We might treat free throw shooting as we would picking a random number from 1 through 100, where we count any number from 1 through 70 as a success and 71 through 100 as failures, but shooting free throws includes the human factor. Maybe she missed because she is nervous or distracted. Maybe it's late in the game and she is tired or injured, changing her technique. If any of these are the case, it might make more sense for us to downgrade the probability of making the next shot, though exactly how much it should be downgraded is no longer some simple formula like turning a fraction into a percentage.
Compensating for failure: Again, let's add in the human factor in this problem. She misses the first free throw, and her coach notices her technique looks inconsistent. "Elbow up!" the coach shouts from the sidelines, and the shooter hears the coach and readjusts her technique to match the way she shoots in practice. Should this change bring her back to being a 70% shooter, or even upgrade her chance of success? That is uncertain, but the failure on the first shot and the diagnosis of at least one reason for the failure could effect the odds, and that effect means the second shot should not be considered independent of the first.
The dependent probabilities in a 52 card deck
One of the simplest mathematical models of dependency is sampling without replacement, which is the way most card games or lotteries or the game of Bingo works. You have a set of outcomes which get effectively randomized and a trial is performed, meaning a card is taken from the deck or a ping pong ball is removed from the hopper or a bingo marker is removed from the spinner. Once removed, the number of possible outcomes has been reduced by one and probabilities for success and failure of certain outcomes change.
Looking for an ace: There are 52 cards is a standard deck and 4 of them are aces. If I draw a card from a randomized deck, the chances are 4/52 = 1/13 ~= 7.7% that the card will be an ace. What are the chances the second card is an ace?
That depends on the first card.
Probability that the second card is an ace, given the first card is an ace is 3/51 = 1/17 ~= 5.9%.
Probability that the second card is an ace, given the first card is not an ace is 4/51 ~= 7.8%.
Unlike the mathematical model of free throw shooting where we re-calculate the probabilities by adding the most recent make or miss into the precentage, which means a miss brings the odds down and a make brings the odds up, not getting an ace makes the odds a little better next time, and getting an ace makes the odds worse.

This is the formula for the dependent probability model of sampling without replacement is given at the left. The two numbers in parentheses are a binomial coefficient, the numbers you get when you use nCr on your calculator, which I pronounce "n choose r" in class. The pairs of numbers that look like a base and an exponent, except that the exponent is underlined, is the convention developed by Donald Knuth at Stanford for writing the numbers that you get on your calculator using the nPr function, which I pronounce "n fall r", referring to the name "the falling factorial". If we think about a deck of cards, the lowercase letters refer to the size of the hand n, where r is the number of successful trials (r for right) and w is the number of unsuccessful trials (w for wrong), and r+w=n. The uppercase letters refer to the size of the deck, where T is the size of the deck, G is the number of cards we consider success if we draw them and B is the number of cards we consider a failed trial if we draw them. The letter T stands for Total, G for Good and B for Bad. Again, we have an equation, G+B=T.
Example: If we want consider drawing a heart a success and anything else a failure, what is the probability of drawing three hearts and two non hearts in a five card hand from a well-shuffled 52 card deck.
Here are the six numbers we need.
n = 5
r = 3
w = 2
T = 52
G = 13
B = 39
On a TI-30XIIs, here are the keys you would press.
5[prb][right]3×13[prb]3×39[prb]2÷52[prb]5[enter]
The calculator will read as follows.
5 nCr 3*13 nPr 3*39 nPr 2/52 nPr 5
0.081542617
This means the probability of exactly three hearts and two cards of some other suit is about 8.15%.
The Expected Value (EV) of a two outcome game
Let us assume we have a game that has only two outcomes, winning and losing. Let us further assume that two players have decided to wager on this game, both putting money into a pooled amount and the winner taking all at the end.
If we look at the game from the point of view of one of the players, we need to know the probability of winning p, how much that player put in, which we call Risk and how much the opponent put in, called Profit. The expected value EV equals the probability of victory p times the sum of Profit and Risk divided by Risk.Different books use different formulas for this. Some do not divide by Risk. By dividing, the number we get is a percentage of return, and a game of flipping coins for $1 a game is equivalent to a game of flipping coins for $100 a game. Some subtract 1 from this formula. This just changes the most important number in identifying results from 1 to 0.
If EV = 1, we consider this a "fair game". For every $1 risked on this game, the expected value is that you will have that dollar returned to you, breaking even. Notice if we are flipping coins, that event never happens on any single play. Either the player makes a dollar profit or a dollar loss, but expected value is about the long run.
If EV > 1, the game is advantageous to the player. If EV < 1, the game is disadvantageous to the player.
In the game of roulette, there are 38 slots where the ball can land, and for simplicity's sake we will assume each has an equal chance of showing up, so p = 1/38. For every $1 you risk, you can make a profit of $35 if you correctly guess in the exact slot where the ball will land. To find the expected value using the TI-30xIIs, you should type in this.
1÷38×(35+1)÷1[enter]
The calculator will read as follows.
1/38*(35+1)/1
0.947368421
What this number means is that for every dollar risked on the spin of a roulette wheel, you should expect about 94.7 cents returned to you in change. In other words, about 5.3 cents is lost from every dollar you bet on every spin of the wheel.
Another way to play the game is to bet red or black. Of the 38 compartments, 18 are red and 18 are black and 2 are green. The probability of victory on betting one of the two major colors is 18/38 = 9/19 ~= 0.473684211. The profit and risk are now both $1. Here's what to type on the TI-30xIIs.
16÷38×(1+1)÷1[enter]
The calculator will read as follows.
16/38*(1+1)/1
0.947368421
The game has changed, both in probability and amount of profit compared to risk, but from the player's point of view, the expected value is precisely the same and still in favor of the casino.
No matter what the levels of profit and risk are, we can find a probability p that will make the expected value equal to 1, and that is p = Risk/(Profit+Risk). If the probability is increased with the profit and risk remaining unchanged, the game becomes advantageous. If is is decreased, the game becomes disadvantageous.
Modern and Classic Parimutuel odds
Profit and risk are listed either in classic form like 3-1 or 2-7 (or sometimes with colons 3:1 or 2:7), where profit is the first number and risk is the second.
In online betting sites, the numbers are given as numbers with absolute value greater than 100, with either a + or - in front of them. +250 means 250 is the profit and 100 is the risk, while -250 means 10o is the profit and 250 is the risk. The fourth page of the yellow sheet explains this in greater detail and shows how to switch back and forth between the two systems.
Practice problems
1. With a well-shuffled 52 card deck, find the probability of getting exactly r hearts in a five card hand when
a) r = 0
b) r = 1
c) r = 2
d) r = 3 (already solved above)
e) r = 4
f) r = 5
2. Find the break-even p when Profit and Risk are as given. Round to three places after the decimal point.
a) Modern parimutuel = +150
b) Modern parimutuel = -110
c) Classic parimutuel = 5:3
d) Classic parimutuel = 5:11
Answers in the comments.
Thursday, April 23, 2009
Class notes for 4/22
Random and deterministic
It's common for people to use the word random in a casual manner, but in the field of the philosophy of science, discussion of whether something can be considered random or not is a subject of intense debate. The opposite of random is deterministic, which is to say that when we perform a task, we understand the possible outcomes thoroughly. For example, putting a key in a lock is a deterministic act. Will the door open? Not necessarily. It might be the wrong key. The key might be correct, but it might have worn out over time. A brand new key in a lock that has been used might have edges that are too sharp. The lock could malfunction. Maybe the person didn't turn the key in the right direction. So deterministic does not mean, "If you do a, then b will also happen." It can be more complicated than that. But in a completely deterministic act, we have an expected outcome, and even when it fails, we have explanations of why it fails.
The problem of determinism versus randomness is not a yes/no situation. We have some acts we consider random, like flipping a coin or rolling a die or choosing a card from a deck. If done under certain circumstances, even these can be deterministic. If the deck is removed from the pack and unshuffled, the top card will be the way the cards were sorted at the factory, and so it is completely deterministic. Some card tricks are done with decks that aren't actually randomly shuffled or the magician has ways of forcing the participant to pick a certain card, so picking a card is deterministic, or picking the card is random but returning it to the deck is deterministic, and so it can easily be found by being out of position. If we drop a coin or a die only a short distance, it might not bounce much, so the result is strongly determined by the original state of the die. Much of the randomness of things like coin flips or dice rolls or lottery balls being removed from a hopper have to do with physics problems that could be considered deterministic if we understood all the variables, but the equations are so difficult that solving them completely is beyond even the most sophisticated computer simulations.
This brings us to random numbers and computers. A computer is completely deterministic. It cannot truly produce a random number, though every computer and even many calculators, including the TI-30XIIs, have random number generators. Here, the computer takes some input unknown to the user, puts it through a function also unknown to the user and produces an output. This is called pseudorandom, and debate over these methods continue to this day. One of the fathers of computer science, the great Hungarian mathematician John Von Neumann, was quoted as saying, "Anyone who considers arithmetical methods of producing random digits is, of course, in a state of sin." A computer has nothing but arithmetical methods available. The tests done to see if a method produces output that is sufficiently random have to do with a distributions of a large set of pseudorandom output.

The probability of r successes in n independent trials, where p is the probability of success of any given trial
The simplest kind of problem in multiple trials is to assume independence between the trials, that the probability of success remains constant over all the trials, and we call that probability p. A basic question in these kinds of experiments is to ask what is the probability of getting a specific number of successes, calling that number r, in a specific number of trials, a number we will call n. If you have a TI-83 or TI-84, there is a function in the distribution menu (blue button then DISTR) called binompdf, that takes as its input n, p and r in that order. On the TI-30XIIs, we have to type in a formula.
Example #1: If we flip a fair coin, p = .5, ten times, what is the probability of exactly 6 heads and 4 tails?
TI-83 or TI-84: Go to the distribution menu, select binompdf and type in the following values
binompdf(10,.5,6) = 0.205078125
On the TI-30xIIs, here is what to type in
10[prb][right arrow]6x.5^6x.5^4[enter]
It will read
10 nCr 6 * .5^6 * .5^4
0.205078125

A different question would be what are the chances of at most r successes in n trials. Again, the TI-83 and TI-84 have a single function solution, this one called binomcdf(n, p, r). On the TI-30XIIs, we either have to calculate all the separate values, from the probability of 0 successes up through the probability of r successes then add these up, or if this is too much work, there is a method using normal distribution to approximate binomial distribution.
If you have Java on your computer, you can go to this website to see how the numbers across a row of Pascal's Triangle look like the bell-shaped curve. We can use z-scores and the lookup table to get values that will approximate these probabilities, and the approxomations get better as the numbers get bigger.
Example #2: If we flip a fair coin, p = .5, ten times, what is the probability of at most 6 heads?
TI-83 or TI-84: Go to the distribution menu, select binomcdf and type in the following values
binompcf(10,.5,6) = 0.828125
On the TI-30XIIs:
This entails a lot of work, but not an impossible amount. We need to find
p(exactly 0 heads) + p(exactly 1 head) + p(exactly 2 heads) + p(exactly 3 heads) + p(exactly 4 heads) + p(exactly 5 heads) + p(exactly 6 heads) = 0.828125
We have the normal approximation method, but we will see when n = 10, it's not very good.
n = 10, p = .5, r = 6
z = (6 + .5 - .5*10)/sqrt(.5*.5*10) = .3
Lookup table: .3 -> .6179
The approximate value at around 62% is very far from the true value at nearly 83%.
Different books have different cut-off points, but at a minimum, using the normal approximation to binomial should only be used when both np > 5 and nq > 5. In this case, both values equal 5, so this would not be a good candidate for using this method.
Example #3: If we flip a fair coin, p = .5, one hundred times, what is the probability of at most 60 heads?
TI-83 or TI-84: Go to the distribution menu, select binomcdf and type in the following values
binompcf(100,.5,60) = 0.982399899891... ~= .9824
TI-30XIIs: It's too many things to add up, so the approximation method is our only hope. Both np > 5 and nq > 5, so we can move forward.
(60 + .5 - .5*100)/sqrt(.5 * .5 * 100) = 2.1
Lookup table: 2.1 -> .9821
When n=100, the approximation is not perfect, but it is very close, unlike the results when n=10.

Another question that could be asked is the probability of getting at least r successes in n trials. "At least r" is the complement of "At most r-1", which means the two probabilities will add up to one.
Example #4: If a 70% free throw shooter takes ten shots, p = .7, what is the probability she makes 8 shots or more?
On the TI-83 or TI-84:
1 - binomcdf(10, .7, 7) = .3827827864..., which rounds to .3828 to four places.
On the TI-30XIIs
10[prb][right arrow]8x.7^8x.3^2[enter]
It will read
10 nCr 8 * .7^8 * .3^2
0.233474441
10[prb][right arrow]9x.7^9x.3^1[enter]
It will read
10 nCr 9 * .7^9 * .3^1
0.121060821
10[prb][right arrow]10x.7^10x.3^0[enter]
It will read
10 nCr 10 * .7^10 * .3^0
0.028247525
If we round these numbers to five places, add them up and round the answer to four places, it should agree with the answer above.
.23347 + .12106 + .02825 = .38278, which rounds to .3828, which agrees with the above answer.
Example #4: If a 70% free throw shooter takes 100 shots, p = .7, what is the probability she makes 75 shots or more?
On the TI-83 or TI-84:
1 - binomcdf(100, .7, 74) = .163130104..., which rounds to .1631 to four places.
On the TI-30XIIs: The problem is adding up too many parts, but np = 70 and nq = 30, both of which are more than 5, so let's go to the approximation method.
z = (75 - .5 - 70)/sqrt(.7*.3*100) = .98198..., which rounds to .98
Lookup table: .98 -> .8365, and 1-.8365 = .1635, which again is pretty close to the actual answer.
It's common for people to use the word random in a casual manner, but in the field of the philosophy of science, discussion of whether something can be considered random or not is a subject of intense debate. The opposite of random is deterministic, which is to say that when we perform a task, we understand the possible outcomes thoroughly. For example, putting a key in a lock is a deterministic act. Will the door open? Not necessarily. It might be the wrong key. The key might be correct, but it might have worn out over time. A brand new key in a lock that has been used might have edges that are too sharp. The lock could malfunction. Maybe the person didn't turn the key in the right direction. So deterministic does not mean, "If you do a, then b will also happen." It can be more complicated than that. But in a completely deterministic act, we have an expected outcome, and even when it fails, we have explanations of why it fails.
The problem of determinism versus randomness is not a yes/no situation. We have some acts we consider random, like flipping a coin or rolling a die or choosing a card from a deck. If done under certain circumstances, even these can be deterministic. If the deck is removed from the pack and unshuffled, the top card will be the way the cards were sorted at the factory, and so it is completely deterministic. Some card tricks are done with decks that aren't actually randomly shuffled or the magician has ways of forcing the participant to pick a certain card, so picking a card is deterministic, or picking the card is random but returning it to the deck is deterministic, and so it can easily be found by being out of position. If we drop a coin or a die only a short distance, it might not bounce much, so the result is strongly determined by the original state of the die. Much of the randomness of things like coin flips or dice rolls or lottery balls being removed from a hopper have to do with physics problems that could be considered deterministic if we understood all the variables, but the equations are so difficult that solving them completely is beyond even the most sophisticated computer simulations.
This brings us to random numbers and computers. A computer is completely deterministic. It cannot truly produce a random number, though every computer and even many calculators, including the TI-30XIIs, have random number generators. Here, the computer takes some input unknown to the user, puts it through a function also unknown to the user and produces an output. This is called pseudorandom, and debate over these methods continue to this day. One of the fathers of computer science, the great Hungarian mathematician John Von Neumann, was quoted as saying, "Anyone who considers arithmetical methods of producing random digits is, of course, in a state of sin." A computer has nothing but arithmetical methods available. The tests done to see if a method produces output that is sufficiently random have to do with a distributions of a large set of pseudorandom output.

The probability of r successes in n independent trials, where p is the probability of success of any given trial
The simplest kind of problem in multiple trials is to assume independence between the trials, that the probability of success remains constant over all the trials, and we call that probability p. A basic question in these kinds of experiments is to ask what is the probability of getting a specific number of successes, calling that number r, in a specific number of trials, a number we will call n. If you have a TI-83 or TI-84, there is a function in the distribution menu (blue button then DISTR) called binompdf, that takes as its input n, p and r in that order. On the TI-30XIIs, we have to type in a formula.
Example #1: If we flip a fair coin, p = .5, ten times, what is the probability of exactly 6 heads and 4 tails?
TI-83 or TI-84: Go to the distribution menu, select binompdf and type in the following values
binompdf(10,.5,6) = 0.205078125
On the TI-30xIIs, here is what to type in
10[prb][right arrow]6x.5^6x.5^4[enter]
It will read
10 nCr 6 * .5^6 * .5^4
0.205078125

A different question would be what are the chances of at most r successes in n trials. Again, the TI-83 and TI-84 have a single function solution, this one called binomcdf(n, p, r). On the TI-30XIIs, we either have to calculate all the separate values, from the probability of 0 successes up through the probability of r successes then add these up, or if this is too much work, there is a method using normal distribution to approximate binomial distribution.
If you have Java on your computer, you can go to this website to see how the numbers across a row of Pascal's Triangle look like the bell-shaped curve. We can use z-scores and the lookup table to get values that will approximate these probabilities, and the approxomations get better as the numbers get bigger.
Example #2: If we flip a fair coin, p = .5, ten times, what is the probability of at most 6 heads?
TI-83 or TI-84: Go to the distribution menu, select binomcdf and type in the following values
binompcf(10,.5,6) = 0.828125
On the TI-30XIIs:
This entails a lot of work, but not an impossible amount. We need to find
p(exactly 0 heads) + p(exactly 1 head) + p(exactly 2 heads) + p(exactly 3 heads) + p(exactly 4 heads) + p(exactly 5 heads) + p(exactly 6 heads) = 0.828125
We have the normal approximation method, but we will see when n = 10, it's not very good.
n = 10, p = .5, r = 6
z = (6 + .5 - .5*10)/sqrt(.5*.5*10) = .3
Lookup table: .3 -> .6179
The approximate value at around 62% is very far from the true value at nearly 83%.
Different books have different cut-off points, but at a minimum, using the normal approximation to binomial should only be used when both np > 5 and nq > 5. In this case, both values equal 5, so this would not be a good candidate for using this method.
Example #3: If we flip a fair coin, p = .5, one hundred times, what is the probability of at most 60 heads?
TI-83 or TI-84: Go to the distribution menu, select binomcdf and type in the following values
binompcf(100,.5,60) = 0.982399899891... ~= .9824
TI-30XIIs: It's too many things to add up, so the approximation method is our only hope. Both np > 5 and nq > 5, so we can move forward.
(60 + .5 - .5*100)/sqrt(.5 * .5 * 100) = 2.1
Lookup table: 2.1 -> .9821
When n=100, the approximation is not perfect, but it is very close, unlike the results when n=10.

Another question that could be asked is the probability of getting at least r successes in n trials. "At least r" is the complement of "At most r-1", which means the two probabilities will add up to one.
Example #4: If a 70% free throw shooter takes ten shots, p = .7, what is the probability she makes 8 shots or more?
On the TI-83 or TI-84:
1 - binomcdf(10, .7, 7) = .3827827864..., which rounds to .3828 to four places.
On the TI-30XIIs
10[prb][right arrow]8x.7^8x.3^2[enter]
It will read
10 nCr 8 * .7^8 * .3^2
0.233474441
10[prb][right arrow]9x.7^9x.3^1[enter]
It will read
10 nCr 9 * .7^9 * .3^1
0.121060821
10[prb][right arrow]10x.7^10x.3^0[enter]
It will read
10 nCr 10 * .7^10 * .3^0
0.028247525
If we round these numbers to five places, add them up and round the answer to four places, it should agree with the answer above.
.23347 + .12106 + .02825 = .38278, which rounds to .3828, which agrees with the above answer.
Example #4: If a 70% free throw shooter takes 100 shots, p = .7, what is the probability she makes 75 shots or more?
On the TI-83 or TI-84:
1 - binomcdf(100, .7, 74) = .163130104..., which rounds to .1631 to four places.
On the TI-30XIIs: The problem is adding up too many parts, but np = 70 and nq = 30, both of which are more than 5, so let's go to the approximation method.
z = (75 - .5 - 70)/sqrt(.7*.3*100) = .98198..., which rounds to .98
Lookup table: .98 -> .8365, and 1-.8365 = .1635, which again is pretty close to the actual answer.
Tuesday, April 21, 2009
Class Notes for 4/20
We dealt with probability in a single instance earlier in the class when we had the relative frequencies of the values of categorical variables. Relative frequencies, listed in a population as p and in a sample as p-hat, are numbers between 0 and 1. If we take all the relative probabilities of all the values of a variable, they will add up to 1, or something very close to 1 depending on rounding error.
We will now talk about probability in multiple event experiments, like flipping ten coins or rolling five dice or drawing a hand of four cards from a 52 card deck. The first important split in the types of multiple event experiments is between independent and dependent events.
Events are independent if the probability of a later event does not change based on the result of an earlier event. For example, if I flip a coin that I can assume is fair, there is a 50% chance of heads and a 50% chance of tails every time I flip it. If by chance, the coin comes up heads ten times in a row, even though earlier testing had shown it to be a 50%-50% chance each time, the eleventh flip is still 50%-50%. Unusually long runs of all heads or all tails are rare, but they are not impossible. Flipping coins and rolling dice are typical examples of independent random events.
Events are dependent if the probability of a later event changes based on the result of an earlier event. The typical example of this is drawing cards from a shuffled deck. If the deck has 52 cards and 4 aces, the probability of drawing an ace from the deck is 4/52 = 1/13 ~= .0769...
If the card has been drawn, what is the probability of the second card being an ace? That depends on what the first card is. If the first card is an ace, there are only 3 left in the deck, which now has 51 cards, so the probability is 3/51 = 1/17 ~= .0588..., which is a lower probability than getting an ace the first card.
If the first card wasn't an ace, the odds are 4/51, ~= .0784..., a slightly higher probability than drawing an ace the first time.
If I say someone is a 70% free throw shooter, is every free throw attempt independent of what happened before? Often, we set up such an experiment assuming independence just to make our work simpler, but the human factor is involved, so in reality it's very likely to be dependent. Some people get frustrated after a few misses and will do worse. Others will learn from the mistakes of a few misses and figure out what they are doing wrong and make improvements. A player might be having a bad day for some reason, or might instead have excellent concentration or just really good luck that day. But again, these kinds of experiments are often set up as though each free throw trial is independent of what came before.
Let's look at flipping coins. A list of all possible events is called the event space. Here are some examples of event spaces.
Event space for flipping one coin
Heads (H)
Tails (T)
ways to get one head = 1
ways to get no heads = 1
Event space for flipping two coins
HH
HT
TH
TT
ways to get two heads = 1
ways to get one head = 2
ways to get no heads = 1
Event space for flipping three coins
HHH
HHT
HTH
HTT
THH
THT
TTH
TTT
ways to get three heads = 1
ways to get two heads = 3
ways to get one head = 3
ways to get no heads = 1

The list of numbers of ways to get r successes in n trials is often written in the pattern of the picture shown here, and this pattern is called Pascal's Triangle, at least in most of the world. The Italians call it Tartaglia's Triangle and the Chinese call it Yanghui's Triangle. None of these people actually invented it or claimed to have invented it. It's been around since before the time of Christ, and it has been studied all around the world.
While it is very common to see it presented in the form here as an equilateral triangle, it can also be presented where the first numbers in each row are lined up straight as follows
1
1 1
1 2 1
1 3 3 1
1 4 6 4 1
... etc.
It is standard to start counting the top row as row 0, and the left most column as column 0. For example, the 6 we see in the middle of the last row I typed in is row 4, column 2. Instead of having a copy of Pascal's Triangle around, our calculators have these numbers available. On Texas Instruments calculators, the function is under the probability menu. On the TI-30XIIs, the way to get that 6 is to type
4 [prb][right arrow]2[enter]
The calculator will read
4 nCr 2
6
All scientific calculators should have this function available, but all of them are slightly different. The TI-89 writes it as nCr(4,2) and Casio calculators write it as 4 C 2. I will pronounce it "4 choose 2", and when I type on the blog, I will type C(4,2). When I write it on the board or on tests, I will put a 4 on top of a 2 and surround both numbers with a large parentheses. These numbers are called the binomial coefficients.
The formula for finding the probability for exactly r successes in n independent trials where the probability of success on any single trial is p is shown here. In some books, they don't use the letter q, instead replacing it with (1-p). Likewise, sometimes w is replaced with (n-r). I use the extra letters and include the relationships between them. The letters r and w stand for right and wrong. The letter p and q are standard in probability texts for the probability of a success or a failure.
Let's do an example. You are given a four question multiple choice test, each question having five possible answers. The test is given in a language you do not read, so all you can do is guess. Each question is independent from the others, meaning that if C is the right answer to the first question, it's also possibly the answer to the second. The probability p of a correct guess is 1 chance in 5, or .2, The probability of failure q is 1-.2 = .8, and of course p + q = 1.
Probability of no correct answers = C(4,0)*.2^0*.8^4 = .4096
Probability of exactly one correct answer = C(4,1)*.2^1*.8^3 = .4096
Probability of exactly two correct answers = C(4,2)*.2^2*.8^2 = .1536
Probability of exactly three correct answers = C(4,3)*.2^3*.8^1 = .0256
Probability of four correct answers = C(4,4)*.2^4*.8^0 = .0016
The expected value of correct answers is n*p, so in this case it's 4*.2 = .8, which isn't possible. You can't get a fraction of correct answers on a multiple choice test. The expected value in this case says that over the long run, a test like this should average .8 right answers out of four. As we can see, the most likely thing to happen is actually a tie for first, where getting either no answers right or one answer right both have a probability of about 41%. If you need to get three answers right to pass the test, the odds are less than 3% to get either three or four right, and the odds of getting everything right by chance is a very slim 16 chances in 10,000.
If you have a TI-83 or TI-84, there is a function under the distribution menu called binompdf(n,p,r). All you have to is enter the function, then the three values in the order given, separated by commas.
The function for three right in four trials with probability .2 at each trial is binompdf(4, .2, 3), which as we see above is .0256.
Practice problem.
The test is changed. There are now five multiple choice questions and four choices for each, but it is still given in a language you do not read.
Round the probabilities to four places after the decimal.
1. What is the expected value?
2. What is the probability of no correct answers?
3. What is the probability of exactly one correct answer?
4. What is the probability of exactly two correct answers?
5. What is the probability of exactly three correct answers?
6. What is the probability of exactly four correct answers?
7. What is the probability of five correct answers?
Answers in the comments.
We will now talk about probability in multiple event experiments, like flipping ten coins or rolling five dice or drawing a hand of four cards from a 52 card deck. The first important split in the types of multiple event experiments is between independent and dependent events.
Events are independent if the probability of a later event does not change based on the result of an earlier event. For example, if I flip a coin that I can assume is fair, there is a 50% chance of heads and a 50% chance of tails every time I flip it. If by chance, the coin comes up heads ten times in a row, even though earlier testing had shown it to be a 50%-50% chance each time, the eleventh flip is still 50%-50%. Unusually long runs of all heads or all tails are rare, but they are not impossible. Flipping coins and rolling dice are typical examples of independent random events.
Events are dependent if the probability of a later event changes based on the result of an earlier event. The typical example of this is drawing cards from a shuffled deck. If the deck has 52 cards and 4 aces, the probability of drawing an ace from the deck is 4/52 = 1/13 ~= .0769...
If the card has been drawn, what is the probability of the second card being an ace? That depends on what the first card is. If the first card is an ace, there are only 3 left in the deck, which now has 51 cards, so the probability is 3/51 = 1/17 ~= .0588..., which is a lower probability than getting an ace the first card.
If the first card wasn't an ace, the odds are 4/51, ~= .0784..., a slightly higher probability than drawing an ace the first time.
If I say someone is a 70% free throw shooter, is every free throw attempt independent of what happened before? Often, we set up such an experiment assuming independence just to make our work simpler, but the human factor is involved, so in reality it's very likely to be dependent. Some people get frustrated after a few misses and will do worse. Others will learn from the mistakes of a few misses and figure out what they are doing wrong and make improvements. A player might be having a bad day for some reason, or might instead have excellent concentration or just really good luck that day. But again, these kinds of experiments are often set up as though each free throw trial is independent of what came before.
Let's look at flipping coins. A list of all possible events is called the event space. Here are some examples of event spaces.
Event space for flipping one coin
Heads (H)
Tails (T)
ways to get one head = 1
ways to get no heads = 1
Event space for flipping two coins
HH
HT
TH
TT
ways to get two heads = 1
ways to get one head = 2
ways to get no heads = 1
Event space for flipping three coins
HHH
HHT
HTH
HTT
THH
THT
TTH
TTT
ways to get three heads = 1
ways to get two heads = 3
ways to get one head = 3
ways to get no heads = 1

The list of numbers of ways to get r successes in n trials is often written in the pattern of the picture shown here, and this pattern is called Pascal's Triangle, at least in most of the world. The Italians call it Tartaglia's Triangle and the Chinese call it Yanghui's Triangle. None of these people actually invented it or claimed to have invented it. It's been around since before the time of Christ, and it has been studied all around the world.
While it is very common to see it presented in the form here as an equilateral triangle, it can also be presented where the first numbers in each row are lined up straight as follows
1
1 1
1 2 1
1 3 3 1
1 4 6 4 1
... etc.
It is standard to start counting the top row as row 0, and the left most column as column 0. For example, the 6 we see in the middle of the last row I typed in is row 4, column 2. Instead of having a copy of Pascal's Triangle around, our calculators have these numbers available. On Texas Instruments calculators, the function is under the probability menu. On the TI-30XIIs, the way to get that 6 is to type
4 [prb][right arrow]2[enter]
The calculator will read
4 nCr 2
6
All scientific calculators should have this function available, but all of them are slightly different. The TI-89 writes it as nCr(4,2) and Casio calculators write it as 4 C 2. I will pronounce it "4 choose 2", and when I type on the blog, I will type C(4,2). When I write it on the board or on tests, I will put a 4 on top of a 2 and surround both numbers with a large parentheses. These numbers are called the binomial coefficients.
The formula for finding the probability for exactly r successes in n independent trials where the probability of success on any single trial is p is shown here. In some books, they don't use the letter q, instead replacing it with (1-p). Likewise, sometimes w is replaced with (n-r). I use the extra letters and include the relationships between them. The letters r and w stand for right and wrong. The letter p and q are standard in probability texts for the probability of a success or a failure.Let's do an example. You are given a four question multiple choice test, each question having five possible answers. The test is given in a language you do not read, so all you can do is guess. Each question is independent from the others, meaning that if C is the right answer to the first question, it's also possibly the answer to the second. The probability p of a correct guess is 1 chance in 5, or .2, The probability of failure q is 1-.2 = .8, and of course p + q = 1.
Probability of no correct answers = C(4,0)*.2^0*.8^4 = .4096
Probability of exactly one correct answer = C(4,1)*.2^1*.8^3 = .4096
Probability of exactly two correct answers = C(4,2)*.2^2*.8^2 = .1536
Probability of exactly three correct answers = C(4,3)*.2^3*.8^1 = .0256
Probability of four correct answers = C(4,4)*.2^4*.8^0 = .0016
The expected value of correct answers is n*p, so in this case it's 4*.2 = .8, which isn't possible. You can't get a fraction of correct answers on a multiple choice test. The expected value in this case says that over the long run, a test like this should average .8 right answers out of four. As we can see, the most likely thing to happen is actually a tie for first, where getting either no answers right or one answer right both have a probability of about 41%. If you need to get three answers right to pass the test, the odds are less than 3% to get either three or four right, and the odds of getting everything right by chance is a very slim 16 chances in 10,000.
If you have a TI-83 or TI-84, there is a function under the distribution menu called binompdf(n,p,r). All you have to is enter the function, then the three values in the order given, separated by commas.
The function for three right in four trials with probability .2 at each trial is binompdf(4, .2, 3), which as we see above is .0256.
Practice problem.
The test is changed. There are now five multiple choice questions and four choices for each, but it is still given in a language you do not read.
Round the probabilities to four places after the decimal.
1. What is the expected value?
2. What is the probability of no correct answers?
3. What is the probability of exactly one correct answer?
4. What is the probability of exactly two correct answers?
5. What is the probability of exactly three correct answers?
6. What is the probability of exactly four correct answers?
7. What is the probability of five correct answers?
Answers in the comments.
Sunday, March 15, 2009
using the TI30X IIs for confidence intervals

Let's say we have the following statistics from a data set and we want to find the 95% confidence interval for mux, the underlying population's true average value.
n = 36
x-bar = 25.35
sx = 6.78
The degrees of freedom will be n-1 = 35, but 35 is not one of the choices on Table A-3, so we have to go with row 34 instead. The Confidence Level Multiplier for 95%, or CLM95%, is 2.032. If you have a TI-30xIIs, key in the following numbers into your calculator to get the high threshold.
25.35+2.032*6.78/[2nd][x^2]36[enter]
The answer should be 27.64616, which we can round to 27.65.
All we have to do to get the low threshold is to change the + to a -. Key in the following.
[up][right][right][right][right][right]-[enter]
After you key this in, the equation line should look like
25.35-2.032*6.78/sqrt(36
and the answer is 23.05384, which rounds to 23.05.
This says we are 95% confident using this data set that the true average of the underlying population is between 23.05 and 27.65.
Tuesday, March 3, 2009
Class notes for 3/2
For most of the rest of the semester, we will be working with the idea of standard deviation, a way to measure how spread out a set of data is. There are going to be a lot of different ways to compute standard deviation depending on what kind of set we are dealing with, and the first two we will learn are sx and sigmax, which are the standard deviations for a sample of numerical data and a population of numerical data, respectively. I am going to go through the steps of calculating these numbers with a small set of data first, then show how to key in the data to the TI-30XIIs, which is a huge time saver.Data set #1: 1, 2, 3, 4, 5, 6
Step #1: Find the average. 1+2+3+4+5+6 = 21, and 21/6 = 3.5. So x-bar or mux is 3.5, depending on whether we have a sample or a population.
Step #2: Take the squares of all the values of data minus the average, then add them together.
(1-3.5)^2 = (-2.5)^2 = 6.25
(2-3.5)^2 = (-1.5)^2 = 2.25
(3-3.5)^2 = (-0.5)^2 = 0.25
(4-3.5)^2 = 0.5^2 = 0.25
(5-3.5)^2 = 1.5^2 = 2.25
(6-3.5)^2 = 2.5^2 = 6.25
sum = 17.5
Step #3: Divide the sum by N or n-1, depending on population or sample.
17.5/6 = 2.91666...
17.5/5 = 3.5
Step #4: Take the square root of the value from Step #3.
sigmax = sqrt(2.91666...) ~ 1.707825...
sx = sqrt(3.5) ~ 1.870828...
For a small set of data and an average that is exact, this isn't so hard. As data sets get larger, this becomes a lot of work to do by hand, which is why a calculator is so valuable.
Steps for TI-30x
Step #1: Get into one variable mode and clear the data set.
If the word STAT is on your screen, this key sequence will do the trick.
[2ND][STATVAR][ENTER][2ND][DATA][ENTER]
If the word STAT is not on your screen, type these key strokes.
[2ND][DATA][ENTER]
Now you are ready to enter in the data. Type in the stuff written in red, where the stuff in black is what is already on the screen.
[DATA]
X1= 1 [DOWN]
FRQ = 1 [DOWN]
X2= 2 [DOWN]
FRQ = 1 [DOWN]
X3= 3 [DOWN]
FRQ = 1 [DOWN]
X4= 4 [DOWN]
FRQ = 1 [DOWN]
X5= 5 [DOWN]
FRQ = 1 [DOWN]
X6= 6 [DOWN]
FRQ = 1 [DOWN]
[STATVAR]
The read out will now give you the following information as your scroll left and right.
n = 6
x-bar (or mux) = 3.5
sx = 1.8708...
sigmax = 1.7078...
sum(x) = 21
sum (x^2) = 91
If you move the underline to the x-bar and press [ENTER], the equation line will now have the symbol x-bar on it, which means the calculator can do equations with the exact values of the average and the standard deviations in them, which will be useful in calculating z-scores.
The standard deviations are roughly equal to the average distance away from the average of all the data in the set. The reason for the n-1 instead of n in the sx equation is the idea of degrees of freedom in a data set. If you know the average of a set of data and the size of the set, you know the total, and if I give you the total of all but one of the values of a set, you can subtract to find the last value.

Why we work with these rough approximations of the average distance instead of the exact value comes from calculus. The normal curve is a bell shaped curve that has area = 1 under the curve from negative infinity to infinity, so any vertical line we draw can cut the area into two parts, where the area under the curve to the left of the line is x and the area to the right of the line is 1-x. A lot of data sets, though not all, have this kind of distribution, and by using z-scores of the raw scores, which is (raw-average)/(standard deviation), we can compare two data sets that are normally distributed, even if they have different averages and different standard deviations. We will look at this in greater detail next class.
Labels:
calculators,
class notes,
instructions,
standard deviation
Monday, February 2, 2009
Class notes for 2/2
Graphical representation of data
Frequency tables and dotplots: A frequency table is a list of values, which we will call x, followed by the number of times each value shows up on the list, which is called the frequency, denoted by f or f(x). A dotplot is a graph that represents a frequency table. For example, let's take the frequency table from the last class notes and turn it into a dotplot with a vertical scale.
x || f(x)
20 || 2
19 || 1
18 || 2
17 || 5
16 || 7
15 || 2
14 || 5
13 || 5
12 || 3
11 || 5
_9 || 2
_8 || 1
_7 || 2
_4 || 3
For the dot plot version, I type in the information in the Courier font, which is mono-spaced. This was the same number of characters on a line will take up the same width
20| **
19| *
18| **
17| *****
16| *******
15| **
14| *****
13| *****
12| ***
11| *****
10|
9 | **
8 | *
7 | **
6 |
5 |
4 | ***
3 |
2 |
1 |
Notice that every value from 1 to 20 is included in the scale, even the ones that did not appear on the list. This is so the dotplot will be correctly spaced.
The five number summary, box and whiskers, IQR and outlying data. Examples of five number summaries and box and whisker plots can be seen in the previous class notes, but John Tukey added one more idea to the box and whisker definition, which would take into account the idea of outlying data, values that are far larger or far smaller than most of the other data. The question is how to define "far larger" and "far smaller", and we will see that in all the different systems we use, the cut-off point for these decisions is arbitrary. Tukey's arbitrary decision involves a number called the Inter-quartile Range or IQR for short. The formula is as follows:
IQR = Q3 - Q1
This is the distance between the values that define the cut-off points for the 75th percentile and the 25th percentile, so the middle half of the data lies inside that range. That is represented by the box, and Tukey decided that the neither the left or right whisker should ever be more than one and a half times as long as the width of the box. If there is any data that is higher that Q3 + 1.5*IQR, it is represented by a dot. Likewise, any data that is lower than Q1 - 1.5*IQR is represented by a dot. The whiskers only extend to the highest data less than the high outlier threshold or the lowest data more than the low outlier threshold.
Example: Here are the number of times the most popular twelve stories on the Huffington Post website were read today, where the scale is in thousands or readers.
759 165 162 152 132 129 128 118 113 65 55 44 35
Obviously, the most popular story* is being read by more people than the next five stories combined, so we would expect it to be outlying data. Let's create the five number summary, the IQR and find the outlier thresholds.
High: 759
Q3: 157 (The average of 162 and 152)
Q2: 128.5 (The average of 129 and 128)
Q1: 60 (The average of 65 and 55)
Low: 35
IQR = 157 - 60 = 97
High outlier threshold = Q3 + 1.5*IQR = 157 + 1.5*97 = 157 + 145.5 = 302.5
Low outlier threshold = Q1 - 1.5*IQR = 60 - 1.5*97 = 60 - 145.5 = -85.5
As we might expect, there is no low outlying data, but the highest value of 759 is much higher than any of the rest of the data. The way this will be presented is by the upper whisker only extending to the value of 165, which is the highest non-outlying value, while the outlier will be represented by a dot. Here is a horizontal box and whisker plot of this data. You can click on this picture to get a larger version.

(* The most popular story the day this data was taken were pictures of Olympic champion Michael Phelps smoking a bong in the dorm room of a friend.)
Rounding: There are several different rules for rounding values, depending on what the data looks like. In general, when taking data from a list and finding the average, we round to one decimal place more than the data is represented. If numbers are given as whole numbers, we round to the nearest tenth. If numbers are given to the nearest penny, we round the average to the nearest tenth of a penny.
The numbers I gathered from the website I decided to round to the nearest thousand then write the numbers in thousands. For this decision, I used the significant digit method. If you put too many places of numbers, people can get caught up in the precision of the number instead of the general size of the number. For big numbers reported in newspapers and magazines, most writers use either one or two or three significant digits, which is to say how many digits are reported before you round the rest of the number to zeroes.
Here are some examples. Let's say that the bailout of the banks cost $763,922,575,873.22. That's too much information, and who knows if being precise makes any sense, since the final numbers might change.
One significant digit: The bank bailout costs $800 billion.
Two significant digits: The bank bailout costs $760 billion.
Three significant digits: The bank bailout costs $764 billion.
All of these answers are correct, depending on how many significant digits we decide to use. The more significant digits, the more precise the answer is, but using more than three or so significant digits for data presented to the public is probably too much information. If we are using numbers in calculations, we usually want to be as precise as possible, and only do the rounding at the end, when the number is presented as the answer.
Order of operations: You may have learned the order of operations by remembering the word PEMDAS or the mnemonic device
Please (parentheses)
Excuse (exponents)
My Dear (multiplication and division)
Aunt Sally (addition and subtraction)
With calculators like the TI-30X II s or the TI-83 or TI-84, the display screen has a line which shows you the equation as you press the keys, so you can enter an entire equation just once and then press the [ENTER] key. The calculator understands the order of operations rules, but sometimes you still have to be careful how you type your formulas in.
A fraction is really two sets of parentheses. You might remember the formula for the slope of a line in algebra class as being the difference of the y values divided by the difference of the x values. Let's say the two points were (8, 7) and (4, 2). Because of the fraction sign we want to do 7-2 = 5 and 8-4 = 4 BEFORE we do the division, so to make that clear to the calculator, we have to put parentheses around the numerator equation, press the division key, then parentheses around the denominator equations, then press [ENTER].
On your calculator, the square root is followed by an open parentheses. In some texts, a formula that includes a square root might show two number being multiplied together by putting the numbers in parentheses and next to each other instead of using the multiplication sign. That can be a problem on the calculator, because the square root starts it own set of parentheses. Instead of typing (.3)(.7), on the calculator you should type .3*.7. The instructions for different equations are shown below.
On most calculators, you don't have to close every set of parentheses. The only calculator I know that is a stickler for parentheses agreeing with one another left and right is the TI-89. On the other hand, if you use a spreadsheet program, it will always complain if your left and right parentheses don't agree.
Frequency tables and dotplots: A frequency table is a list of values, which we will call x, followed by the number of times each value shows up on the list, which is called the frequency, denoted by f or f(x). A dotplot is a graph that represents a frequency table. For example, let's take the frequency table from the last class notes and turn it into a dotplot with a vertical scale.
x || f(x)
20 || 2
19 || 1
18 || 2
17 || 5
16 || 7
15 || 2
14 || 5
13 || 5
12 || 3
11 || 5
_9 || 2
_8 || 1
_7 || 2
_4 || 3
For the dot plot version, I type in the information in the Courier font, which is mono-spaced. This was the same number of characters on a line will take up the same width
20| **
19| *
18| **
17| *****
16| *******
15| **
14| *****
13| *****
12| ***
11| *****
10|
9 | **
8 | *
7 | **
6 |
5 |
4 | ***
3 |
2 |
1 |
Notice that every value from 1 to 20 is included in the scale, even the ones that did not appear on the list. This is so the dotplot will be correctly spaced.
The five number summary, box and whiskers, IQR and outlying data. Examples of five number summaries and box and whisker plots can be seen in the previous class notes, but John Tukey added one more idea to the box and whisker definition, which would take into account the idea of outlying data, values that are far larger or far smaller than most of the other data. The question is how to define "far larger" and "far smaller", and we will see that in all the different systems we use, the cut-off point for these decisions is arbitrary. Tukey's arbitrary decision involves a number called the Inter-quartile Range or IQR for short. The formula is as follows:
IQR = Q3 - Q1
This is the distance between the values that define the cut-off points for the 75th percentile and the 25th percentile, so the middle half of the data lies inside that range. That is represented by the box, and Tukey decided that the neither the left or right whisker should ever be more than one and a half times as long as the width of the box. If there is any data that is higher that Q3 + 1.5*IQR, it is represented by a dot. Likewise, any data that is lower than Q1 - 1.5*IQR is represented by a dot. The whiskers only extend to the highest data less than the high outlier threshold or the lowest data more than the low outlier threshold.
Example: Here are the number of times the most popular twelve stories on the Huffington Post website were read today, where the scale is in thousands or readers.
759 165 162 152 132 129 128 118 113 65 55 44 35
Obviously, the most popular story* is being read by more people than the next five stories combined, so we would expect it to be outlying data. Let's create the five number summary, the IQR and find the outlier thresholds.
High: 759
Q3: 157 (The average of 162 and 152)
Q2: 128.5 (The average of 129 and 128)
Q1: 60 (The average of 65 and 55)
Low: 35
IQR = 157 - 60 = 97
High outlier threshold = Q3 + 1.5*IQR = 157 + 1.5*97 = 157 + 145.5 = 302.5
Low outlier threshold = Q1 - 1.5*IQR = 60 - 1.5*97 = 60 - 145.5 = -85.5
As we might expect, there is no low outlying data, but the highest value of 759 is much higher than any of the rest of the data. The way this will be presented is by the upper whisker only extending to the value of 165, which is the highest non-outlying value, while the outlier will be represented by a dot. Here is a horizontal box and whisker plot of this data. You can click on this picture to get a larger version.

(* The most popular story the day this data was taken were pictures of Olympic champion Michael Phelps smoking a bong in the dorm room of a friend.)
Rounding: There are several different rules for rounding values, depending on what the data looks like. In general, when taking data from a list and finding the average, we round to one decimal place more than the data is represented. If numbers are given as whole numbers, we round to the nearest tenth. If numbers are given to the nearest penny, we round the average to the nearest tenth of a penny.
The numbers I gathered from the website I decided to round to the nearest thousand then write the numbers in thousands. For this decision, I used the significant digit method. If you put too many places of numbers, people can get caught up in the precision of the number instead of the general size of the number. For big numbers reported in newspapers and magazines, most writers use either one or two or three significant digits, which is to say how many digits are reported before you round the rest of the number to zeroes.
Here are some examples. Let's say that the bailout of the banks cost $763,922,575,873.22. That's too much information, and who knows if being precise makes any sense, since the final numbers might change.
One significant digit: The bank bailout costs $800 billion.
Two significant digits: The bank bailout costs $760 billion.
Three significant digits: The bank bailout costs $764 billion.
All of these answers are correct, depending on how many significant digits we decide to use. The more significant digits, the more precise the answer is, but using more than three or so significant digits for data presented to the public is probably too much information. If we are using numbers in calculations, we usually want to be as precise as possible, and only do the rounding at the end, when the number is presented as the answer.
Order of operations: You may have learned the order of operations by remembering the word PEMDAS or the mnemonic device
Please (parentheses)
Excuse (exponents)
My Dear (multiplication and division)
Aunt Sally (addition and subtraction)
With calculators like the TI-30X II s or the TI-83 or TI-84, the display screen has a line which shows you the equation as you press the keys, so you can enter an entire equation just once and then press the [ENTER] key. The calculator understands the order of operations rules, but sometimes you still have to be careful how you type your formulas in.
A fraction is really two sets of parentheses. You might remember the formula for the slope of a line in algebra class as being the difference of the y values divided by the difference of the x values. Let's say the two points were (8, 7) and (4, 2). Because of the fraction sign we want to do 7-2 = 5 and 8-4 = 4 BEFORE we do the division, so to make that clear to the calculator, we have to put parentheses around the numerator equation, press the division key, then parentheses around the denominator equations, then press [ENTER].
On your calculator, the square root is followed by an open parentheses. In some texts, a formula that includes a square root might show two number being multiplied together by putting the numbers in parentheses and next to each other instead of using the multiplication sign. That can be a problem on the calculator, because the square root starts it own set of parentheses. Instead of typing (.3)(.7), on the calculator you should type .3*.7. The instructions for different equations are shown below.
On most calculators, you don't have to close every set of parentheses. The only calculator I know that is a stickler for parentheses agreeing with one another left and right is the TI-89. On the other hand, if you use a spreadsheet program, it will always complain if your left and right parentheses don't agree.
Thursday, January 29, 2009
Inputting a data set into the TI-30X II s
In the following instructions, I will write comments in this font and the stuff from the calculator in a blocky font called Courier. The things the calculator will put on the screen are in black, while the keys you type in will be in red.
First, you want to be in one-variable stat mode. If you are in stat mode, the word STAT shows up on the output screen in small letters.
Step 0: If you don't see the word STAT on your screen, type the following.
[2nd][DATA][ENTER]
This puts you in stat mode. Before you hit enter, the screen should read
1-VAR 2-VAR
and by pressing enter when 1-VAR is underlined, you go into 1-variable statistics mode. You can skip now to Step 2.
Step 1: If you see the word STAT on the screen, type the following.
[2nd][DATA]
What you will see are the words
1-VAR 2-VAR [left][left]
come up on the screen, and then when you scroll left twice, you will see the word
CLRDATA [ENTER]
underlined. That's when you press enter, which clears out any old data you have.
Step 2: Inputting a data set. The two important buttons here are [DATA] and [STATVAR]. [DATA] lets you start inputting data, and when you are finished pressing [STATVAR] will get you to the statistics associated with the data set you just entered.
One variable statistics in the TI-30X II s are input as a frequency table, a value followed by how many times it shows up on the list. Consider the following list.
20, 20, 19, 18, 18, 17, 17, 17, 17, 17, 16, 16, 16, 16, 16, 16, 16, 15, 15,
14, 14, 14, 14, 14, 13, 13, 13, 13, 13, 12, 12, 12, 11, 11, 11, 11, 11,
9, 9, 8, 7, 7, 4, 4, 4,
[DATA]
x1= 20 [down]
frq= 2 [down]
x2= 19 [down]
frq= 1 [down]
x3= 18 [down]
frq= 2 [down]
x4= 17 [down]
frq= 5 [down]
x5= 16 [down]
frq= 7 [down]
x6= 15 [down]
frq= 2 [down]
x7= 14 [down]
frq= 5 [down]
x8= 13 [down]
frq= 5 [down]
x9= 12 [down]
frq= 3 [down]
x10= 11 [down]
frq= 5 [down]
x11= 9 [down]
frq= 2 [down]
x12= 8 [down]
frq= 1 [down]
x13= 7 [down]
frq= 2 [down]
x14= 4 [down]
frq= 3 [down]
x15= [STATVAR]
The screen will go blank, then show the word CALC for a few seconds, then the statistics for the data set appear. You can use the [left] and [right] buttons to look at different statistics. The number that appears is the underlined statistic. On this list, the numbers are as follows.
Step 3: Reading the statistics of the sample. (Or it could be the parameters of the population, depending on how the data set was defined.)
What you will see on the screen is in black, what keys you will press are in red. The only key in these instructions are the left button, but the right button works for scrolling as well.
n x-bar sx sigmax
45 [left]
n x-bar sx sigmax
13.333333... [left]
n x-bar sx sigmax4.073193967 [left]n x-bar sx sigmax
4.027681991 [left]
sum(x) sum(x^2)
600 [left]
sum(x) sum(x^2)
8730
To get out of the screen showing the statistics, press [CLEAR]. If you press [ENTER] instead, the screen will show the equation line with the symbol of the statistic you were looking at added to whatever equation you were just looking at. This will be useful later in the class when we need the average x-bar and one of the standard deviations, either sx or sigmax, for calculating z-scores.
If you have any questions about how to use one-variable mode on the TI-30 X II s, leave a comment or send me an e-mail.
First, you want to be in one-variable stat mode. If you are in stat mode, the word STAT shows up on the output screen in small letters.
Step 0: If you don't see the word STAT on your screen, type the following.
[2nd][DATA][ENTER]
This puts you in stat mode. Before you hit enter, the screen should read
1-VAR 2-VAR
and by pressing enter when 1-VAR is underlined, you go into 1-variable statistics mode. You can skip now to Step 2.
Step 1: If you see the word STAT on the screen, type the following.
[2nd][DATA]
What you will see are the words
1-VAR 2-VAR [left][left]
come up on the screen, and then when you scroll left twice, you will see the word
CLRDATA [ENTER]
underlined. That's when you press enter, which clears out any old data you have.
Step 2: Inputting a data set. The two important buttons here are [DATA] and [STATVAR]. [DATA] lets you start inputting data, and when you are finished pressing [STATVAR] will get you to the statistics associated with the data set you just entered.
One variable statistics in the TI-30X II s are input as a frequency table, a value followed by how many times it shows up on the list. Consider the following list.
20, 20, 19, 18, 18, 17, 17, 17, 17, 17, 16, 16, 16, 16, 16, 16, 16, 15, 15,
14, 14, 14, 14, 14, 13, 13, 13, 13, 13, 12, 12, 12, 11, 11, 11, 11, 11,
9, 9, 8, 7, 7, 4, 4, 4,
[DATA]
x1= 20 [down]
frq= 2 [down]
x2= 19 [down]
frq= 1 [down]
x3= 18 [down]
frq= 2 [down]
x4= 17 [down]
frq= 5 [down]
x5= 16 [down]
frq= 7 [down]
x6= 15 [down]
frq= 2 [down]
x7= 14 [down]
frq= 5 [down]
x8= 13 [down]
frq= 5 [down]
x9= 12 [down]
frq= 3 [down]
x10= 11 [down]
frq= 5 [down]
x11= 9 [down]
frq= 2 [down]
x12= 8 [down]
frq= 1 [down]
x13= 7 [down]
frq= 2 [down]
x14= 4 [down]
frq= 3 [down]
x15= [STATVAR]
The screen will go blank, then show the word CALC for a few seconds, then the statistics for the data set appear. You can use the [left] and [right] buttons to look at different statistics. The number that appears is the underlined statistic. On this list, the numbers are as follows.
Step 3: Reading the statistics of the sample. (Or it could be the parameters of the population, depending on how the data set was defined.)
What you will see on the screen is in black, what keys you will press are in red. The only key in these instructions are the left button, but the right button works for scrolling as well.
n x-bar sx sigmax
45 [left]
n x-bar sx sigmax
13.333333... [left]
n x-bar sx sigmax4.073193967 [left]n x-bar sx sigmax
4.027681991 [left]
sum(x) sum(x^2)
600 [left]
sum(x) sum(x^2)
8730
To get out of the screen showing the statistics, press [CLEAR]. If you press [ENTER] instead, the screen will show the equation line with the symbol of the statistic you were looking at added to whatever equation you were just looking at. This will be useful later in the class when we need the average x-bar and one of the standard deviations, either sx or sigmax, for calculating z-scores.
If you have any questions about how to use one-variable mode on the TI-30 X II s, leave a comment or send me an e-mail.
Labels:
calculators,
inputting a data set,
instructions,
TI-30X II S
Sunday, January 25, 2009
The basics of the TI-30X II S
The TI-30X II S is the recommended calculator for the class. It has many features that will do some very difficult calculations automatically. In the instructions, any time a key press or series of key presses is discussed, the instructions will be in red in the font called Courier.
One of the most important keys on the calculator is the blue key [2nd] at the upper left. Using this key, almost all the black keys and the bottom row of the white keys have double uses. For example, the
[ON]
button is at the lower left. If you press
[2nd][ON]
that turns the calculator off. Some, but not all the functions on the regular black buttons have an inverse function associated with it by pressing 2nd. For example, the 4th button from the bottom in the left column, with [x^2] on it, will square a number, while if you key in [2nd][x^2], it takes the square root of a number. Be careful. To get a square root, you need to type
[2nd][x^2] 3 [ENTER]
which will give you 1.732050808, while to get 3 squared, you type
3 [x^2][ENTER]
to get 9.
To enter in data sets, we will be using the DATA and STATVAR keys, as well as the four arrow buttons in the upper right. Since the four symbols are not available in the text editor I use for this blog, I shall type [up], [down], [left] and [right] to identify the four buttons.
Getting in and out of STAT mode: There are some words written in a small font below the output line for answers. If you see STAT written on that line, this means the calculator has a current set of data it is storing. If you don't see that word, there is nothing currently stored.
If you don't see STAT: Press
[2nd][DATA]
and the equation line of the output screen will give the choices of 1-VAR and 2-VAR. You can use the [left] and [right] buttons to toggle back and forth. Pick the one you want, then press
[ENTER]
(The first data sets we are going to deal with will be 1-variable applications.)
If you want to start a new data set, press
[2nd][DATA][left]
This gives you the CLRDATA option. If you press
[ENTER]
all data that was stored will be erased.
Now we know how to get in and out of STAT mode and erase old data. Next we will learn how to enter a data set and how to get the parameters or statistics, which is to say some of the important numbers associated with the data set, which could be a population or a sample.
Monday, January 12, 2009
The minimum necessary calculator

There is no assigned text for the class, but it is mandatory that you have a calculator that can do statistics. If you already have a calculator, check with the teacher to see if it can do all the tasks needed for the class. For example, a TI-83 or a TI-84 are very good choices, but if you are going to buy a calculator specifically for this course, you can get one much cheaper. (Note: If you have a TI-89 and you aren't planning to take any math after statistics, you have too much calculator for your purposes, and it might be a good idea to buy a new, cheaper calculator. The TI-89 is designed for people who are going on in math to calculus and beyond, and it makes doing statistics a little harder than it should be.)
I recommend the TI-30XIIs calculator. Texas Instruments does not pay me to say this. The reasons I recommend it are these.
1. It's cheap. The highest price I have seen for it is $19.99 at Walgreen's. I have seen it as low as $13.50 at Radio Shack.
2. It's available at a lot of places. The Laney bookstore should have this calculator, but it is also sold at drug stores like Walgreen's and Long's, electronics stores like Radio Shack and Best Buy, office supply stores like Office Max and Staples and department stores like Target and Wal-Mart. There should be one of these stores somewhere convenient for you.
3. It does a lot of work for you. The TI-83 and TI-84 are a little more useful than the TI-30XIIs, but the price difference is about $100 more. If this is the last math class you will be taking, I can't recommend spending the extra cash for just a few features. (Note: I do NOT recommend the TI-30Xa. It's not quite powerful enough.)
4. It's solar powered. There is also the TI-30XIIb, and the only difference is b is for battery and s is for solar. Spend two bucks more now, save on batteries and help the environment.
5. Even after class is over, it's nice to have a good calculator around the house. Or at least that's my experience.
As I wrote in the syllabus, there is no textbook for the class, but a calculator is mandatory. You will not be allowed to share a calculator with a classmate on quizzes and exams, and you will be expected to know how to use your own calculator. There will be class time alloted to teaching people how to use their calculators, and I would like the class to have as few different types of calculators as possible. I recommend Texas Instrument calculators because I have used most of the models that are useful for a statistics course and should be able to instruct students on their uses. Other brands like H-P or Casio are not be as well known to me, and a student with one of these may be forced to read the instruction manual to find out how to do certain things needed in the course.
Subscribe to:
Posts (Atom)
