Probability/Hypothesis testing help

  • Thread starter adeel
  • Start date
  • Tags
    Testing
In summary: Alpha is the symbol for the population mean; subscript a denotes the alternative hypothesis. So in this problem you are looking for the probability of an A or more given that the mean mark is 80% or greater.The probability is .8 or greater.The second question is about a binomial probability problem. You are looking for the probability that at least 20 out of a population of 2400 will receive an A. This is represented by the symbol P(x\geq .8). To solve for this, you use the following equation:P(x\geq .8)=.0946.The third question is about a two-tail test. To use a two-tail table for a one-tail test, you need to look
  • #1
adeel
45
0
I need help on some questions: The first two seem really easy, but i can't seem to figure out the right approach:

Marks in ECMB11 are normally distributed with a mean of 70% and standard deviation of 10%. Suppose an A is a mark of 80% or more.

1. If you take a sample of 120 students, what is the probability that at least 20 will receive an A.
2. What is the probability that average mark for 120 students will be at least 80%.

-------------

You are conducting a lower tailed test. Suppose alpha = 0.10, beta = 0.01, n = 16, s = 1460, and u subscript a (i guess population of a) = 2400. Determine u subscript o. (I guess population of o)

I have no idea what they are even asking for in this question


HELP is greatly appreciated.
 
Physics news on Phys.org
  • #2
I'm not an expert at stats but I will take a stab at this...

Notice that your mean is .7 and your SD is .1 so .8 is one SD from the mean. So what I am thinking is that you take

[tex] 1 - ( 1/2 * 0.68 + .5)[/tex]

to get the probability that any student will get .8 or more. Now that we know [tex]P(x \geq .8) [/tex] we can make it a binomial problem, either they are or they are not.

[tex] _{120}C_{20}* (.1587)^{20}*(.8413)^{100} = .0946[/tex]

I am :confused: about the rest of your questions...

I just hope that's all correct.

Regards
 
Last edited:
  • #3
Does anyone else know how to work these problems? I would like to see how they are worked out... :smile:
 
  • #4
your answer to the first question seems right, but i think the formula you gave is for the probability of having exactly 20, where as i Need at least 20. But I think we can use a chart to get it. I am glad you were able to help on this one. Just wondering if the binomial logic part makes sense. Does seem right.

After taht im...... :frown:
 
  • #5
adeel said:
You are conducting a lower tailed test. Suppose alpha = 0.10, beta = 0.01, n = 16, s = 1460, and u subscript a (i guess population of a) = 2400. Determine u subscript o. (I guess population of o)
Are you certain that you have ua and uo, not [itex]\mu_a \text{ and } \mu_o[/itex]? [itex]\mu[/itex] (mu) is the symbol for distribution mean; subscript a denotes the alternative hypothesis and subscript o denotes the null hypothesis.
 
  • #6
thats what i meant (mu) I am not sure how you guys get it to come out properly...

isnt mu population mean as well? Oh wait, they are equal to each other...Uhh so yeah...i guess i shuld have tried to type it like that. Hopefully can help me to solve the problems...
 
  • #7
i assumed that u would be interpreted as mu although i guess I should have made it clear.
 
  • #8
adeel said:
thats what i meant (mu) I am not sure how you guys get it to come out properly...

isnt mu population mean as well? Oh wait, they are equal to each other...Uhh so yeah...i guess i shuld have tried to type it like that. Hopefully can help me to solve the problems...
Right. So the picture can be described as follows. You have two partially overlapping "bell curves" (two normal density plots or graphs). The point of their intersection is the critical x (xc), marked on the horizontal axis. The area under the rightmost bell curve to the LEFT of xc is alpha. The area under the leftmost bell curve and to the RIGHT of xc is beta. The mean of the leftmost curve is mua and the mean of the rightmost curve is muo.

I hope this helps.
 
  • #9
If you look up the standard normal probability table for alpha = 0.1, you will see a value like z0.1=1.6. I am making this up as an example, you should look at the table yourself. But usually the table is for a two-tailed test. To use a two-tail table for a one-tail test, you want to look up alpha = 0.2. Let's say you did this and came up with a critical value z0.2. Because this is a lower-tail test, you want the negative value of z0.2, that is -z0.2. Now,
[tex]-z_{0.2}=-\frac{x_c-\mu_o}{\sigma}.[/tex]
Clearly if you knew xc you could determine muo. You have to use the alternative dist. to derive xc.
To do this, look up z0.02. Then solve xc from
[tex]z_{0.02}=\frac{x_c-\mu_a}{\sigma}.[/tex]
Now substitute xc into the previous equality and solve for muo.
 
Last edited:
  • #10
oh...ok i think i have an example like that from my class notes...thanks alot.


Anyone have ideas as to how to correctly do the first two i posted?
 
  • #11
For the first 2 questions you have to assume that the 120 random var.s are independent of each other. That's a standard assumption but was not stated explicitly in this problem.
adeel said:
Marks in ECMB11 are normally distributed with a mean of 70% and standard deviation of 10%. Suppose an A is a mark of 80% or more.

1. If you take a sample of 120 students, what is the probability that at least 20 will receive an A.
"At least 20 will receive an A" = (NOT "exactly 19 receive A") AND (NOT "exactly 18 receive A") AND ... AND (NOT "exactly 1 receives A") AND (NOT "exactly 0 receives A"). Now use the binomial formula to calculate each of the terms in quotes (e.g. Prob{"exactly 19 receive A"}), then take the complement ("NOT") to get each paranthesis term (e.g. Prob{NOT "exactly 19 receive A"} = 1 - Prob{"exactly 19 receive A"}), then multiply the complements. Formally this is equivalent to deriving the prob. distribution of the 20th highest order statistic out of a sample of 120 normal random variables. Typically this is a non-Normal distribution, closer (pehaps identical?) to a Log-normal.
2. What is the probability that average mark for 120 students will be at least 80%.
Need to derive the probability distribution of the sample average for a sample of 120. That distribution will also be Normal, you just need to figure out what its mean and the standard dsitribution are. Show that cumulative distribution as F. Then Prob(Average > 80) = 1 - F(80).
 
Last edited:
  • #12
How do I derive the probability distribution?
 
  • #13
EnumaElish said:
Need to derive the probability distribution of the sample average for a sample of 120. That distribution will also be Normal, you just need to figure out what its mean and the standard dsitribution are. Show that cumulative distribution as F. Then Prob(Average > 80) = 1 - F(80).

Could one use the central limit theorem to say that the mean of the sample is approximately equal to the mean of the population?
 
  • #14
adeel said:
How do I derive the probability distribution?
Well, the sample average (Xbar) is a linear combination of 120 normal random var's (RVs). Each RV ~ N(mu,sigma) and they are independent. A linear combo of several Normal RV's is itself a Normal RV, and all you need to derive is its mean and std. dev. For the mean, you can use an expected value theorem that says expectation is a linear operator, hence:

E(Xbar) = E(Sum[Xi/n]) = E(Sum[Xi]/n) = E(Sum[Xi])/n = Sum[E(Xi)]/n = Sum[mu]/n = n mu/n = mu.

So the mean of the sample average is identical to anyone of the RV's mean (all of which are equal to mu = 70). No big surprise here.

Next, you need to derive the std. dev. of the sample mean. It should be equal to sigma/sqrt(n), where sigma is the std. dev. given in the problem. But you should check this formula. I'll think about how it is derived and come back if I can remember it.
 
Last edited:
  • #15
This theorem (like the linear expectation theorem) is from Mood, Graybill, Boes' Introduction to Probability and Statistics textbook (I am sure about the author names but slightly less so about the title):

For RV's X, Y and constants a, b, Var(aX+bY) = a2Var(X) + b2Var(Y) + abCov(X,Y). In your case the X is identical to Y, there are 120 RV's, and all Cov terms = 0 (because of independence). "Sum" denotes summation over i = 1 through 120. Therefore: Var(Xbar) = Var(Sum[Xi/n]) = Var(Sum[(1/n)Xi]) = (1/n)2Sum[Var(Xi)] = (1/n)2Sum[sigma] = (1/n)2n sigma2 = sigma2/n. Since std. dev. = sqrt(Var), it follows that Std. Dev. of the sample average = sigma/sqrt(n).

We have proven that Xbar is ~ N(mu, sigma/sqrt(n)) where mu = 70, sigma = 10, n = 120. This is your F distribution.
 
Last edited:
  • #16
ok thanks for your help everyone. I think i understand how to do it.
 
  • #17
59 years since Cox, and people still insist on the tail area tests?
 

1. What is the difference between probability and hypothesis testing?

Probability is a mathematical concept that measures the likelihood of an event occurring. It is based on data and can be used to make predictions about the future. Hypothesis testing, on the other hand, is a statistical method that is used to determine if there is a significant difference between two groups or if a certain relationship exists between variables.

2. How is probability used in hypothesis testing?

In hypothesis testing, probability is used to calculate the likelihood of obtaining a certain result if the null hypothesis (the statement being tested) is true. This probability is known as the p-value. If the p-value is low enough, typically 0.05 or less, it is considered significant and the null hypothesis is rejected in favor of the alternative hypothesis.

3. What is the purpose of hypothesis testing?

The purpose of hypothesis testing is to determine if there is enough evidence to support a claim or hypothesis about a population based on a sample of data. It allows scientists to make conclusions about a larger population based on limited data.

4. What are the key steps in hypothesis testing?

The key steps in hypothesis testing are:

  • Step 1: State the null and alternative hypotheses.
  • Step 2: Choose the appropriate test statistic and level of significance.
  • Step 3: Collect data and calculate the test statistic.
  • Step 4: Determine the p-value and compare it to the chosen level of significance.
  • Step 5: Make a decision to either reject or fail to reject the null hypothesis.
  • Step 6: Interpret the results and draw conclusions.

5. What are some common misconceptions about hypothesis testing?

One common misconception is that a significant result always means that the alternative hypothesis is true. However, it is important to remember that a significant result means that the null hypothesis can be rejected, but it does not necessarily prove the alternative hypothesis. Another misconception is that a non-significant result means that the null hypothesis is true. In reality, a non-significant result simply means that there is not enough evidence to reject the null hypothesis, but it does not prove that the null hypothesis is true.

Similar threads

  • Set Theory, Logic, Probability, Statistics
Replies
11
Views
1K
  • Set Theory, Logic, Probability, Statistics
Replies
8
Views
2K
  • Set Theory, Logic, Probability, Statistics
Replies
20
Views
3K
  • Set Theory, Logic, Probability, Statistics
Replies
5
Views
2K
  • Set Theory, Logic, Probability, Statistics
Replies
3
Views
1K
  • Set Theory, Logic, Probability, Statistics
Replies
4
Views
4K
  • Set Theory, Logic, Probability, Statistics
Replies
4
Views
908
  • Set Theory, Logic, Probability, Statistics
Replies
6
Views
1K
  • Set Theory, Logic, Probability, Statistics
Replies
1
Views
2K
  • Precalculus Mathematics Homework Help
Replies
29
Views
2K
Back
Top