Sunteți pe pagina 1din 3

The Taller You Are, The Wider You Will Be

My partner and I chose the topic of height and wingspan. We wanted to see if
there was a strong correlation between someone’s height and someone’s wingspan.
Before starting our project, we predicted that the two topics would have a strong and
positive correlation. We decided to take a different approach than most groups to collect
our data. To collect our data, we found an online data source that had heights and
wingspans of 50 different people. Here is a link to the data set we found:
https://statproject159.weebly.com/data.html
Our explanatory variable was height while our response variable was wingspan.
We thought that it would make most sense if as you got taller, your wingspan would also
grow.

After collecting our data, we were able to create a scatter plot depicting the data
we collected and estimate the potential correlation between our selected topics.
Originally we realized that our data had some unrealistically small numbers (potentially
the height and span of children compared to adults) which was throwing off our data. To
compensate, we chose to remove the first sixteen points that were significantly
impacting our data set and added sixteen more at the bottom. After making this change,
our scatter plot gave a more accurate representation of our data that suggested a
positive, potentially strong correlation.
After this, we were able to begin making our calculations. Our correlation
coefficient r turned out to be 0.854, confirming that our data set had a positive, strong
correlation.
After we calculated our correlation coefficient r, we calculated X-Bar and Y-Bar.
For our data, X-Bar was for height and Y-Bar was for wingspan. We found out that X-
Bar (height) was 166.755102, while Y-Bar (wingspan) was 168.0653061. X-Bar was the
average for the 50 heights and Y-Bar was the average for the 50 wingspans.
After making these calculations, we were able to determine our least squares

regression line for our data set was y=0.854x+25.7


Our marginal change was 0.854, meaning that for every one centimeter increase in
height, there will be a 0.854 centimeter increase in wingspan.
When we first started our project we took out sixteen influential points. These
points were very small compared to the rest of our data set After removing these points.
we found that there were no influential points in our data set. Therefore, there was no
influential points on our graph. Influential points are significant because they can
change the y-hat equation and many other variables like X-Bar and Y-Bar.
Our coefficient of determination (𝑟 2 ) is 0.73. This is the number that represents
our explained variation and can then be used to determine our unexplained variation.
The amount of explained variation in our regression is 73%. The amount of
unexplained variation in our regression is 27%. This tells us that there is more explained
variation in our regression than there is unexplained.
Originally, our data set had multiple lurking variables because we realized it
including the height and span of children rather than just adults, which left a significant
gap between some of our points. This prevented us from even having a readable graph.
To fix the issue we removed all of the outliers, which were irrelevant to our intended
data set, and added an equal amount of points from our source that were better fit for
our data set. If we hadn’t done this, our variables would not have appeared to have any
correlation and we would not have a proper graph.
For interpolation we used the equation y=0.854x+25.7 and plugged in the
number 175 for the letter x. After calculating the equation, we got the number 175.15.
For extrapolation we used the equation y=0.854x+25.7 and plugged in the number 250
for the letter x. After calculating the equation, we got the number 239.2.
After choosing the variables of height and wingspan, we discovered that our
variables had strong, positive correlation after completing our calculations. This
confirmed our original estimates that height and wingspan are very closely related and
have direct impacts on one another.

S-ar putea să vă placă și