You are on page 1of 9

COKE DATA PROJECT: PHASE TWO BRIAN JEAN STATS 1510 DAY BY NICOLE FROST

Abstract In this experiment, we compared three different types of soda in order to see if they stayed true to their advertised weights. The brands we used were Coke Zero, Diet Coke, and Diet Pepsi. Each can was labeled as containing 12 fluid ounces, or approximately 355 grams. By taking a random sample of 15 cans from each brand, we found that there was more diversity between individual cans than the brands themselves. We weighed a total of 45 cans and our results ranges from 363.01 grams to 375.48 grams. We discovered that Diet Pepsi weighed the least, Diet Coke weighed the most, and Coke Zero fell in between the other two. Based on this, the weight of the Diet Pepsi was the closest to the advertised amount of approximately 355 grams. If this study is truly representative of the population, then I would rather buy Diet Coke because there is more product per can.

Introduction For this project, we compared the actual weight of three different types of soda to the weight that was claimed on the soda cans. This is important to us as consumers because the experiment will help us see if we are being cheated out of product or if we are receiving extra product. Hypothesis: Our hypothesis was that the Diet Pepsi would weigh the less than Coke Zero and that Diet Coke would weigh the most out of the three. We decided that Coke Zero and Diet Coke would weigh more because Diet Pepsi is more highly carbonated than the other brands, meaning that there is more air and less product. We felt that Diet Coke would weigh the most because Coke Zero was advertised as a "lighter" and more dietconscious beverage.

The purpose of this project was to measure and compare the weight per can of three different brands of cola: Diet Coke, Diet Pepsi, and Coke Zero.

Methods The data was collected using the actual soda cans and a table balance. These were the only tools used except for our iPads and, of course, TC Stats. The Diet Coke had a population of 85, Diet Pepsi had a population of 96, Coke Zero had a population of 72. From each population, we took a sample of 15. The cans from each population were numbered. Using TC Stats and simple random sampling, we generated a set of 15 random numbers from each brand and we measured the weight of the cans that had those numbers. This was done to minimize the amount of error and bias in the study, as much as possible.

Results After we compiled our data, we found out that Diet Coke was in fact the heaviest cola, and Diet Pepsi was the lightest. The Coke Zero sample fell in between the other two colas

Figure 1

Figure 2 Possible error could have come from not balancing the scale, people moving the table,

and air affecting the scale. Also, bias could have affected the results because, if one believes that a sample is going to be heavier, they might hold it down firmer on the scale. Errors could have been made while inputing or recording the data with our iPads because of typos.

The way of measurement that would have been the safest in terms of possible error would be a census, but this would be an inefficient use of our time, even though it would be more accurate. As shown in Figure 1, the results from the Diet Pepsi and Coke Zero brands of soda produced distributions that were skewed left. The Diet Coke results produced a belldistribution. Also, as shown by Figure 2, the range of each of the sodas was at least 5 grams. Since some of the data is skewed, the five number summary was given in Figure 2 to show a better representation of the data than the box plot alone. The medians of the Diet Coke and Coke Zero were within 1 gram of each other but the Diet Pepsi data

had a median of 369.36 grams. The mean values were very close to the median values, meaning that the data values were relatively close to each other. For the data that was skewed, the median would be an appropriate measurement of location. For the Diet Coke data, the mean would be an appropriate measurement of location. In TC-STATS I constructed a confidence interval for Diet Coke, Diet Pepsi, and Coke Zero. Since these samples violated the assumptions for normality, I used a sign test with the medians and a confidence level of .95.

Figure 3-Diet Coke

Figure 4-Diet Pepsi

Figure 5-Coke Zero

The confidence intervals are Diet Coke- (370.8700, 373.2700), Diet Pepsi: (366.5000, 370.3800), and Coke Zero- (370.0500, 372.7000). This means that approximately 95% of the soda cans in this experiment weighed in between the values shown in the confidence intervals. After obtaining the confidence intervals, I used TC-STATS to perform a Kruscal-Wallis test to compare the different brands of soda. This was used instead of ANOVA because the distributions aren't normal. For this, my null hypothesis is satisfied by c=z=p and my alternative hypothesis is satisfied by one of the medians being different. The value of is 0.05.The p-value from this test is 0.0001 and, since this is less than , I reject the null hypothesis. This means that at least one of the medians is different and, as shown by Figure 2, Diet Pepsi is the most different from Coke Zero and Diet Coke. We used TC Stats to record our data and produce Figure 1 and Figure 2.

Conclusion My hypothesis was supported by the data in this experiment because Diet Pepsi was found to weight the least and Diet Coke weighed the most. Although the medians were all very close, by performing a Kruscal-Wallis test I saw that there was a difference in the data and the medians showed that Diet Pepsi was the most different.

Appendix Raw data:


Diet Coke Weight 371.14 370.76 371.18 373.6 371.8 373.27 370.87 372.93 373.16 368.25 373.69 371.22 372.16 373.33 375.48 Diet Coke can # 47 79 32 66 58 67 22 7 40 54 11 55 50 9 70 Diet Pepsi Weight 366.16 370.43 369.15 371.36 370.38 370.69 370.19 369.25 369.36 367.87 363.01 366.5 369.01 371.71 369.36 Diet Pepsi can # 14 56 21 43 50 33 17 46 86 15 19 4 90 79 58 Coke Zero can # 38 62 63 47 71 34 24 2 65 25 10 11 28 16 17 Coke Zero Weight 373.04 372.7 373.51 372.87 370.73 373.46 372.01 369.72 371.41 371.83 370.05 368.83 372.6 371.93 370.4

You might also like