During my runs I stumble upon many sizes of pine cones. Just for curiosity, I was wondering whether or not the length of a pine cone had an effect on the number of scales a pine cone had.
What I plan on doing is collecting a bunch of pine cones, measure the height, and then dismantle it's scales and count them. I predict that there will be a positive association between the height, and the number of scales on a pine cone, and that there will be a strong linear relationship.


DATA COLLECTION OF PINE CONES:
Screenshot 2017-11-14 at 7.31.52 AM.png
PHOTO OF COLLECTING DATA
IMG_20171110_080044052_HDR.jpg
SCATTER PLOT OF PINE CONE DATA
Screenshot 2017-11-15 at 1.47.36 PM.png

Screenshot 2017-11-14 at 9.50.09 AM.png

THIS CHART INCLUDES THE CORRELATION COEFFICIENT

THE CORRELATION OF DETERMINATION IS 40.57 %

THE LEAST SQAURED REGRESSIONAL EQUATION IS -1.2919536 + 4.7854295 x

THIS IS A GRAPH SHOWING WHERE THE FITTED LINE GOES THROUGH THE DATA
Screenshot 2017-11-14 at 10.03.14 AM.png

THIS IS THE RESIDUAL PLOT COMPARING X VALUES TO THE RESIDUAL:
Screenshot 2017-11-14 at 10.00.56 AM.png
RESIDUAL GRAPH ( POINTS WERE SQUARED ):
external image kujQu3aDgAh6pFhWi0lKMKdD9XvxV3jR37M-D277MpLYZyf4YGifTwJetT-djciQshLstzlO_71mta-gUCF-P1_uG9uN1uRfw16lq0YU30O2PFZzL10nguAuv0aJxVb_Sx_i2M_Q
DATA ANALYSIS:

My hypothesis was correct that the association between the length of the pine cone and the number of scales it had was positive. It was a moderately linear, not as strong as I thought it would of been. I thought that the longer the pine cone was, the more room available for scales to grow. The correlation was right about .64 which helps concluded a positive linear association. The residual graph that was generated from this data came out pretty spread out indicating a rather linear / straight line in my data. The residuals scale, which was between -20 and 20, was a bit on the larger side of what I wanted for the scale of the residual to be. In the second residual graph the data is squared. Having the data squared the residual graph is more scattered ( less of a recognizable pattern ) indicating that the data has been straitened out more. The new residual graph is more spread out then the previous one and it has a scaling between (-1.5 and 1). This graph is much better than the first one because it has a smaller scale and it's data points are more spread out. Some influences in my data that could of made it possible for errors to have occur were that the pine cone could of had missing scales ( I tried my hardest to find pine cones that were full ) and also possible miss counts of scales during the counting process.