Вы находитесь на странице: 1из 4

Skittles final project

Tanner Peterson
Period 7

After counting the amount of each color skittle in a regular pack, there was a sample
taken with 34 data entries. This data will be used to look deeper into a confidence interval and
hypothesis tests. Two classes of students entered their data on how many of each color skittle
their pack contained.

Data collection:

number of red number of number of yellow number of green number of purple


orange

10 8 15 16 11

369 339 410 358 385

Confidence Interval:
We are constructing a range of numbers where the desired number we are choosing is likely to
be. The amount of confidence we have in the number falling within that range is determined by
the question.

Based on the data that I collected, I will create a 99% confidence interval to estimate the true
proportion of yellow candies.

1. We are trying to estimate p= the true proportion of yellow candies. Our best guess is p
hat= .25 but because of sampling variation we are unlikely to be correct. So, we will
calculate a 99% z-interval for p.
2. Conditions:
● random sample: yes
● independent: yes
● sample is less than 5% of the population
● the population is reasonably assumed to be normal.
3. 99% CI= p^±z*√p^q^/n= ( .1404 , .3596)

.1404 .25 .3596


4. Thus, I am 99% confident that the interval from .1404, .3596 captures the true proportion
of yellow skittles in a regular package.
Construct a 95% confidence interval for the true mean number of candies per bag:
1. We are trying to estimate μ = the average number of skittles per package. Our best
guess is Ẍ = 60.548 but because of sampling variation we are unlikely to be correct. So
we will calculate a 95% t-interval for μ.
2. Conditions:
a. Random sample: Yes
b. Sample ﹤5% of population? Assume population: Yes
c. Normality-large sample size? n>30 or is population approx. normal? Yes
3. 95% CI= μ ± t*(s/√n) = ( 59.1971 , 61.8997 ) df = 32

59.1971 60.548 61.8997


4. Thus, I am 95% confident that the interval from ( 59.1971 , 61.8997 ) captures the true
average of skittles in a regular package.

The first interval relates that there is a 99% confidence that you would have between ( .1404 , .
3596) as the proportion of yellow candies to total candies. Because there are 5 colors you would
expect a 25% proportion which clearly falls in the range of the interval.

Secondly, the next interval represents the amount of total candies you would expect to find in a
package. Based on the information, there is a 95% confidence that the number would fall
between ( 59.1971 , 61.8997 ). The samples that have been collected reflect this as the average
of the samples is 60.548, which falls in the range.

Hypothesis tests:

After creating a previous hypothesis, it is possible to test it using the data that has been
collected. We will use a hypothesis test to interpret and test our skittles data.

1. Using a .05 significance level we will test the claim that 20% of skittles are red.

At first glance, it appears that the true proportion of red skittles is greater than .2 since p^ = .25.
However, it is also possible that the true proportion is p= .2 and we got a sample proportion this
low because of sampling variability. To decide, we will conduct a 1 sample z test for p (a = .05)

Ho = .2 Ha ≠ .2
Conditions:
a. Random sample: Yes
b. Independence: Yes
c. Normality : Yes
z= -.6455 p= .5186

Since the p-value of .5186 > .05, we do not reject the null hypothesis. There is not significant evidence
to suggest that the alternative hypothesis is correct.

2. Using a .01 significance level testing the claim that the mean number of skittles in a bag is 55

At first glance, it appears that the true proportion p of skittles is greater than 55 since p =60.548.
However, it is also possible that the true proportion is p= 55 and we got a sample proportion this
low because of sampling variability. To decide, we will conduct a 1 sample t test for p. (a =.01)

Ho = 55 Ha ≠ 55

Conditions:
a. Random sample: Yes
b. Independence: Yes
c. Normality: Yes

t= 8.363 p= .0000000015

There is a .01 significance level which the p value is smaller than. The lower the p value the
more evidence to suggest rejecting the null hypothesis. There is more evidence suggesting that
the alternative hypothesis is true.

The first hypothesis test suggests that there is more than 20% red skittles per bag, and the
second test suggests that there is more than 55 skittles total on average. There is a small
possibility that either of the alternative hypothesis tests are correct.

Reflection:

The conditions to be met for hypothesis tests and interval estimates are slightly different
but are used to determine whether the experiment or study is available to be tested on. The
interval estimates have to be random samples, less than 5% of the total population, and
approximately normally distributed. The study conducted on skittles was a random sample, it is
reasonable to assume that the 33 studies are less than 5% of the total, and it can also be
assumed to be normally distributed.
Hypothesis tests have the conditions of being a random sample, independent studies,
and normally distributed. In this case, they are again random samples, independent, and
reasonably assumed to be normally distributed. The skittles test is assumed to meet all of the
conditions that are needed for a hypothesis test.

In order to improve the normality condition of the color problems, it is possible to make
sure there is deviation in the amount of each color. The graph would likely be more unimodal
than normally distributed based on assumption. If the deviation is large enough it is possible that
it could become more normal.

Errors could come by the normal distribution not being enough, or the sample not being
random enough. Sampling errors could come from the process that they were collected. It was
likely a convenience sample, but there is a high possibility that the effect of that is small. By
looking at these interval estimates and hypothesis tests, the data can be interpreted and
discussed more fully. The mean amount per bag is likely over 55 and between 59.1971 and
61.8997.

Вам также может понравиться