Chapter 7 Project by Cody, Ashley, and Lauren

We decided to cover a topic regarding the student GPA of a random survey of 50 FHS
students. In this study, we conducted the survey, received our feedback, and calculated the mean
to be a GPA of 3.519 and a standard deviation of .57726. These are two very important
calculations because they allowed us to make inferences and calculate probabilities pertaining to
the GPAs of Fowler High School students. Here is a table of our results:

3.9 3.98 2.6 3.6 3.06 3 3.7

4 3.5 3.8 3.871 3.94 3.825 3.8

3.0857 3.6 3.9 3.27 3.99 3.4 3.5

3.7 3.77 3.75 3.83 3.8 3.85 3.7

3.6 1.78 3.92 2.3 2.9 3.8 3.6

1.35 3.8 3.85 3.93 3.8 3.65 2.5

3.7 3.85 3.8 3.95 3.9 3.27 1.75

After collecting 50 GPAs, we decided to put the 50 data values into L1 on the
calculator. We then calculated 1 VARI stats to get the five number summary which included the
mean (3.159), sample standard deviation (.57726), minimum (1.35), maximum (4.0), Q1 (3.5),
and Q3 (3.85). From the five number summary we were able to calculate the Interquartile Range
(IQR). The formula to calculate IQR is Q3-Q1: 3.85 - 3.5 = .35; this value shows us the spread of
our data within the middle of the data, which is between Q1 and Q3. Our median was found by
finding the middle of all 50 data values. The median is 3.76, which was higher than our mean of
3.519.
The outliers for the data are 1.35, 1.78, 2.3, 2.5, 2.6, 2.7 and 2.9. We found these outliers
by first using the formula 1.5 x IQR or 1.5 x 0.35 = 0.525. Next, we subtracted this number from
Q1 and added the score to Q: 3.5 - 0.525 = 2.975 and 3.85 + 0.525 = 4.375. If there is any
numbers below 2.975 or above 4.375, these data values are outliers. Therefore, we found there
were 7 numbers below 2.975, so these numbers are outliers.
The histogram that we graphed does not look like a normal distribution. The histogram
more so looks like a skewed left graph. Pearsons index for skewness measures how much as
skewness a graph has from a normal distribution. 3(mean - median)/ standard deviation =
3(3.519 - 3.76)/ 0.57726 = -1.252 Since this value is less than -1, it indicates there is skewness in
the graph. When we look at the graph, this skewness number is confirmed because the graph is
skewed left.

Using the Empirical Rule, for 68% of the data, the values fall between 2.94 and 4.09. We
found these values by subtracting and adding one standard deviation to the mean (3.519 -
0.57726 = 2.94 and 3.519 + 0.57726 = 4.09). For 95% of the data, the values fall between 2.369
and 4.67. We found these values by subtracting and adding two standard deviations from the
mean (3.519 -1.15 = 2.369 and 3.519 + 1.15 = 4.67) For 99.7% of the data, the values fall
between 1.787 and 5.25. We found these values by subtracting and adding three standard
deviations from the mean (3.519 - 1.73 = 1.787 and 3.519 + 1.787 = 5.25).
When looking at our groups GPAs, we are able to make comparisons to our data by
standardizing the raw data scores into z scores. The first of the three GPAs that we compared
was 3.63, which we converted to a z-score using the formula: (3.63 - 3.519)/(.57726) = .19.
Looking at the left-tail z score distribution, we were able to conclude that the probability of a
FHS student getting a score lower than this GPA is 58%, whereas the probability of getting a
score above this value is 42%. The second GPA that we compared to our data using the z score
distribution is the GPA of 4.0. The formula for converting this value to a z score is: (4.0 -
3.519)/(.57726) = .83. This is a relatively above average z score, therefore about 80% of the data
is probable to occur; leaving only 20% of the data to be ahead of it. It is important to note that
although the z score shows that 20% of the GPAs would be above a 4.0 based on our data, this is
inaccurate when we consider that the highest GPA is, in fact, a 4.0. This issue might also have an
effect on the probabilities of our other calculations, and it is an unfortunate drawback to the
standard comparison in this circumstance. The third and final GPA that we compared with our
data was a 3.1. This data value is different in the respect that it will have a negative value, since
it is below the mean. Our formula for this GPA is as follows: (3.1 - 3.519)/(.57726) = -.73. This z
score, when shown on the z distribution tells us that the probability that someone is below this
GPA value is 23%, leaving the probability that someone has a higher GPA to be 77%. One final
calculation that we made with these z scores was made in order to calculate a probability of
GPAs that lie between two of these values. In this case, we calculated the probability that the
GPAs in our data would be between a 3.1 and a 4.0. In order to figure this problem out, we must
refer back to the z scores: 3.1 = -.73 and 4.0 = .83. All that needs to be done after this is the
probabilities from our z distribution must be subtracted: P(.83) - P(-.73) = .7967 - .2327 = .564
or 56%.
Our GPA scores (that we used to compare to our data) are also useful in the calculations
of samples. An important aspect of samples to remember is that the standard deviation is
different, which is /n. We do not need to worry about the average, which is the same, but the
standard deviation equals: (.57726)/(30) = .105393 An example of this is when looking at a
sample of 30 FHS students, which is the n value in these calculation; since there are 30
students in this example, we implemented this into calculating the standard deviation above.
With the mean and new standard deviation, we are able to create new z scores. With these z
scores, we are trying to find the probability that the 30 students from this sample are above the
GPA of 3.63, as well as between 3.1 and 3.63. The formula for calculating the probability that a
student has a GPA greater than 3.63 is: (3.63 - 3.519)/.105393 = 1.05; P(1.05) = .8531 or 85%.
Again, these probabilities are found by using the same z distribution as before. Finally, the
probability of GPAs from the sample that will fall between the GPAs of 3.1 and 3.63 is
calculated by: P(1.05) - P(-3.97) = .5753 - .0000 = .5753 or 58%. The reason for the large
difference in z scores and probabilities is attributed to the decrease in the size of the standard
deviation. In our previous calculations, the standard deviation was .57725, whereas it is now
.105393, and this change in values causes a big difference in our z scores, and finally our
probabilities.
Z-scores are obtained from raw scores, which are the initial scores that are present in the
data. While taking the the z-distribution into account, we are able to calculate the raw data values
below which only 3% of our data should fall, as well as the raw data value that includes the top
3% of our data above it. The value at which 3% of our data should fall below is a 2.43 GPA, and
the other value is 4.60. We found these values by finding the percentage on a z-distribution
probability chart, and then found the raw data value based off knowing the z-value. It is as
simple as solving for x in the formula z=(x-)/. Again, it is important to recognize the
although the raw data value above which only 3% of our data should fall (4.60) is what we
calculated, it is an impractical value due to the cap at a 4.0 GPA.
Our group created a survey of FHS students GPAs, and we were able to calculate many
interesting values based off of these GPAs and their z-scores. Although this subject was not
perfectly normal, we learned a large amount about practical and life-applications for these
methods. Although not all of our values were accurate, if we used a different source of data, the
outcome may end differently.