Sie sind auf Seite 1von 9

Sandy Vo Instructor Doug Richards Math 1040-003 September 19, 2013 Group Project - Part 2 As a group, we obtained our

samples with the data set provided by the canvas site. We chose the exhale study and headings. The height of each person was listed from 1 to 654. We chose simple random and systematic as our sample choices. For random simple, we used a website that generated random numbers. We input the first number and the last number that it could select from. We decided on thirty as a sample size because it was a nice and even number. We generated thirty random numbers and looked at the numbers listed on the chart with the height that correlated to that number. For example, the generator gave us the number 471. From the data (in spreadsheet form), we went down the population number 471 and found that the height for that individual was 65 inches. We did that for every number we generated. For systematic, we took the largest number of data we had (654) and divided it by the number we wanted in our sample (30). The product was 21.83 and we rounded that up to 22. Then we chose a random number between 1 and 22 as our starting point and selected our data accordingly. From the pie chart we created, the tall height (66 inches-74 inches) percentage stayed the same for both simple random and systematic. The medium height from the systematic had a sixteen percent difference from the simple random. Simple random has 43 percent while systematic had 60 percent. The short height varied from a sixteen percentage point as well. With systematic (because it was already in descending order), it looked like the pareto graph, but in ascending order. The bar graph for systematic went from shortest to tallest while the bar graph on simple random looked somewhat like a normal distribution. It would be difficult to tell because there are only three bars. The pareto chart on simple random looked somewhat uniform while the systematic, you could definitely tell it was skewed right.

If we were to compare our charts to the entire population of the data set, we can see that there is a significantly large difference between the short height and medium height. With the entire population, the data concludes that 18 percent is short while our data concluded 34 percent from the simple random sample and 17 percent from systematic. The data we got concluded that 43 percent of our simple random was of medium height and 60 percent (from systematic) were medium height while 59 percent were medium in the population data. Tall height wasnt very significant, but what surprised me the most was that the percentage from each sample and our population remained constant at 23 percent.

Population
Tall 23% (.2263) Short 18% (.1850)

Medium 59% (.5887)

Simple Random Bar Graph


3

Tall

Medium

13

Short

10

10

12

14

Simple Random Pie Chart

Frequency

23% Tall

34% Short

43% Medium

Height (in inches)

Simple Random Pareto


14 12 10

Medium Short Tall

Frequency

8 6 4 2 0 1

2 Height (in inches)

Systematic Bar Graph


Tall
7

Height

Medium

20

Short 3
0 5 10

Frequency 15

20

25

Systematic Pareto
25 20 Frequency 15 10 5 0 1 2 Height 3

Medium

Tall Short

Systematic Pie Chart


Short 17%

Tall 23%

Medium 60%

Group Project - Part 3 In the simple random, the shape of the box plot median a little more towards the right side. The right whisker is a bit longer than the left side. As compared the box plot from the cluster, the cluster box plot looks more normal. It is about symmetrical on both sides. The frequency histogram from simple random looks more skewed right with very little high numbers (in this example, its age). But there appears to be a small outlier. If we were to exclude the last age column, the diagram would be close to a normal distribution. The histogram for the cluster looks like a normal distribution with just little differences in the height of the bar. Comparing the two histograms, the simple random looks more equal except for that one data entry on the right side. The cluster mean had been much closer to the population mean than the simple random, but the simple randoms standard deviation is much closer than the cluster standard deviation. Population Mean: 9.9 years old Population Standard Deviation: 2.95 years old Simple Random: Sample Mean: 10.6 Sample Standard Deviation: 2.58 Cluster: Sample Mean: 9.63 Sample Standard Deviation: 2.07

Simple Random Box Plot:

Simple Random Frequency Histogram:

Cluster Box Plot:

Cluster Frequency Histogram:

Group Project - Part 4 From the simple random sample, I am 90 percent confident that the true proportion of the short people falls between 1.8 and 19 percent of the population. From the population pie chart, I can confirm that the population proportion is 18 percent. After analyzing this data set, I can say that this sample was within the 90 percent confidence level. From the systematic sample, I am 90 percent confident that the true proportion of the short people falls between 5.7 percent and 28.2 percent of the population. As stated above, the population proportion is 17 percent. From this confidence interval, I can confirm that the data from this sample falls in the limits of my 90 percent confidence level. From the mean confidence level, I chose 90 percent confidence. The mean of the population was 9.9 years and the population standard deviation was 2.95 years. The first sample I selected was simple random. As a group, we chose 30 people to be in our sample. The x-bar (mean) of this sample came out to be 10.6 years and the standard deviation came out to be 2.58 years. After performing a 90 percent confidence level on this data, I concluded that from the sample, I am 90 percent confident that the true population proportion for those in the exhale-inhale falls between 9.8 years and 11.3 years. This sample fits into my 90 percent confident level. From the cluster sample, I am 90 percent confidence that the true proportion of those in exhaleinhale data set falls between 8.9 and 10.3 percent. The population mean is 10.6. This data almost satisfies my 90 percent confidence level, but not quite.

Part 5 Null hypothesis: P=61.14358 Alternative hypothesis: P>61.14358 Alpha level: .10 Test statistic: 2.29 P-value .0220 Since the P-value .0220 is less than the significance level .10, we can reject the null hypothesis. The data meets the requirements because we have completed a simple random sample. There was a fixed number of trials (n=30), and the population proportion is greater than 5 percent. The numbers represent that there was not sufficient evidence to conclude that the proportion was not 61.14358. In my data, I could have made a type I error by rejecting a true null hypothesis (false positive). I could be concluding that a relationship exists when it in fact does not exist (in my case, the height of my sample and the height of my population). Part 6 During the duration of the class, we learned several concepts in statistics. This project followed the concepts we learned and applied it to a population data. Part two required the group to find a sample using one of the techniques we learned in class. Basically, this project taught us how to take a concept and apply it to some kind of real data. The project proceeded to have the group create histograms, box and whisker plot, find a sample standard deviation, and so on and so forth. In my opinion, it helped kind of glue the concept in our brains. The goal was to complete the steps required on the instruction sheet and eventually come up with a hypothesis for our data and whether it is reliable data. During this project, I have learned a tremendous amount of valuable information. I have learned that some statistics that are published in the world, may in fact, be unreliable due to the method that was used. Statistics has yet to be applied for any of my other classes, but as a required class for my medical lab major, I can be assured that statistics will come in handy during that time (otherwise, it wouldnt be required class). During the project, my group faced many problems. We initially had four members and now we are left with three. Setting up a group meeting was probably the most problematic part about the project. Our times conflicted numerous times, but we eventually came to a consensus and worked together as a group. As for our problem solving skills, we realized towards part 3 that we had messed up on our data. When we had transferred our data over from the excel spread sheet to statcrunch, we had left out about fifty numbers and that made a huge difference when we recalculated the population proportion.

This project changed my views on the real-world math applications because I could easily read about the studied being conducted or even conduct my own study. Ive never understood other math classes and how their math equations applied in the real world. It just didnt make sense in my head. Statistics made me think of math in a different way. I learned to enjoy this kind of math because its something that I could use in the real world.

Das könnte Ihnen auch gefallen