- Do you think the experiment was valid? What would you change to improve the validity of the results?
- Was there anything about the activity you did not fully understand? Try to explain where you may have gotten lost.
- Do you believe the simulation resulted in good results? What would you have expected to happen if we had only simulated 25 taste test trials? 50,000 taste test trials?
- Why are histograms a good way to visualize data in this case?
- How would you change the dice simulation if you wanted to taste test the entire senior class as opposed to just our statistics class? What expected outcome would we anticipate if the seniors truly could not taste a difference?
Please post your response and comment on one classmate's response no later than the start of class on Fri, Sep 14.
I believe the Coke vs. Pepsi experiment was completely valid and showed that most people IN OUR STATISTICS CLASS do not know the difference between Coke and Pepsi. I believe to improve our validity of this experiment, all we can do is run a repeat experiment. The activity was easy to understand and there weren't any problems performing the experiment. The simulation showed good results and it was very interesting to see how the histogram distributed the numerical values in a standard bell curve manner based on the probability we had calculated in class. If we had simulated only 25 taste test trials I believe that we would have seen a slightly unequal distribution in the standard bell curve value, and we wouldn't get to clearly see the probability of rolling a specific number over another. If we had simulated 50,000 test trials I believe that the graph would still display a very well defined standard bell curve but it would also show some new variations such as rolling more "0"'s or more higher numbers; however, the most probable numbers would still be continued to be rolled the most. Histograms are a very good way to show data because they show a visual correlation in data and allow an individual to easily make conclusions about statistical experiments. As for the dice experiment, it would be very time consuming and difficult to use this method to calculate random probability for a whole senior class. If we were to roll 50 or 60 dice at a time (each dice simulating a test subject, just like in class)it would take forever to count them and sort out the probability. We would have to use some kind of computer generated dice model that can randomly roll and collect data for us.
ReplyDeleteUsing a computer to count the dice and collect data is a good idea.
DeleteI forgot to mention that the expected outcome for the senior class would just be the same results as our class except for at a larger scale. The bell curve would represent larger amount of numbers.
DeleteI believe the experiment was valid. The way we had it set up, there was really no way of knowing what the difference of the two drinks were besides tasting them. To improve on the validity, we can always repeat the experiment several times until we get consistent data.
ReplyDeleteNo, the setup was pretty clear.
I believe the simulation produced a clear result. The data we received at the end wasn’t really surprising. I believe that whether we do 25 taste trials or 50,000 taste trials, the law of probability will always even the results out to the point where the two trials are identical.
Histograms are used to show continuous data. It’s used to show when numbers can take on any value of a certain range and that’s exactly how our collected data was.
As far as doing the dice simulation with the senior class, several things need to be changed. The whole experiment itself will be long and time consuming. We would need more than the 19 dice per group we had in class so we need to figure out a way for everyone to have a dice to roll. Then there’s the problem of valid results. With that many people, it increases the chances of error. If the seniors could not taste the difference, then we also shouldn’t look too much into it. A sample of +50 people is not enough to make a valid outcome that people can’t taste the difference between Coke and Pepsi.
it is true that the senior class is small and 50 people couldn't really provide accurate and valid results
DeleteOur stats experiment was completely valid. I would have used a bigger group of people and did more trials. The activity was easy enough to understand and follow. The results we got were good if we had used 25 trials then the data would have been skew. For 50,000 people the results would leave more space for the extremes and a variety of results. Histograms are a useful tool to this experiment because it shows the normal distribution and frequency of 3s and 5s. The senior class isn't that large so each student could have more dice to do trials. It would take time though. If the seniors couldn't tell the difference between the two sodas then we can say that some would guess randomly and could get it right just based on chance.
ReplyDeleteI also think that doing a larger number of trials would increase the validity of the experiment.
Delete*The coke vs. pepsi test that we did in class was completely valid in almost all aspects, I can only see one potential flaw in it. The sample group that the experiment was performed on could have been more diverse. The entire sample group was from the same geographic area and around the same age. This could introduce unwanted variables into the experiment. The experiment definitely gives good results if the population is only the town of West Point, however if the population is bigger than that a more diverse sample group would produce better results.
ReplyDelete*The experiment was fairly easy to understand, I didn't run into any problems.
*I think that the simulation that we did ended in very good results. The pattern in the numbers was pretty obvious. Even though it wasn't completely even on both sides of the curve it was easy to get the idea. If we had only done 25 trials the pattern of the numbers would not have been as consistent and we might get very skewed results from the simulation. If we had done 50,000 simulations I believe that the bell curve would look similar to the one we had in class. The only difference would be that the bell curve would be more accurate and symmetrical. The peak of the bell curve would probably sit at 6 1/3 because this is the exact average number of people to guess correctly out of 19
*Histograms are a good way to visualize data because it is an easy way of graphically showing the portions of numbers in data. This can help many people to learn because the majority of our class consists of visual learners.
*In order to test the entire senior class I would use the same amount of dice as there are seniors. Then I would roll the dice and collect enough simulations to make a smooth and fairly symmetrical bell curve (maybe 500 or so). Assuming that the seniors have 60 people in their class I would say the results of the test (if the seniors couldn't taste a difference) could range anywhere from 12 to 28.
Ya, I agree on the range of the results.
DeleteI think the Coke vs. Pepsi experiment was valid. The more times we do the experiment will show if it was actually valid or are getting mixed results. Everything in the experiment was clear-cut and easy to understand. I was very surprised when the simulations gave us the results we were looking for. If we had only done 25 simulations and 50,000 taste trials I think our overall results would be the same. In the simulations we probably would not have gotten as good results because we only would have done 25 trials, which is not a lot. When they showed up on the histogram it was easy to understand and visualize because it shows where the curve would be and how all the dice rolls were distributed. The dice simulation would change due to the fact that the senior class is larger then our statistics class. We only used 17 dice because there are 17 people in the class. So for the senior class you would need to increase the number of dice used. If the seniors really could not taste the difference between the two we might see the same results because they still have three cups and two sodas to choice from so the chance of them getting it right might be the same.
ReplyDeleteI agree that the histogram was easy to understand. It made it much easier to visualize the data.
DeleteI agree that the chance of the seniors getting it right would be the same.
DeleteI do think this experiment was valid. It fulfilled the purpose and showed that most people cannot differentiate between Coke and Pepsi. We could run more trials to validate our results. It would give us more data to conclude from.
ReplyDeleteThere was nothing in this activity that I did not understand. It was well explained. It was also thoroughly explained how the simulation with the dice tied into the experiment with the Coke and Pepsi. It showed the probability of students guessing correctly if they could not tell the difference.
I believe the simulation resulted in very good results. It was what we expected. If we had only simulated 25 taste test trials, we would most likely not get the results we expected. I don’t believe we would get the same curve. More trials are needed. If we had done 50,000 taste test trials, I think the curve would be about the same as what we got. The most probable numbers would still be rolled the most which is six or seven.
Histograms are a good way to visualize the data because it shows a clear curve and the number three and five were rolled the most. It makes it easy to view and interpret the results.
If we wanted to test the senior class, we would need to increase the amount of dice we were rolling. You would probably need one die per senior to represent each student. We would expect around one-third of the senior class to guess the correct cup based on probability if they could not tell the difference.
I agree that we would probably not get the same results with 25 simulations as we did with our simulations in class, but with the 50000 simulations I think we would get very similar results as our simulations.
DeleteI believe the experiment we did was a valid one. I do not think there is much way to improve the validity of our experiment, the way we had our experiment was set up there really wasn't any way to tell the difference besides actually tasting the difference. The only other thing we could do to verify our experiments results more would be to do more trials than we did.
ReplyDeleteI think I fully understood the experiment and how it was set up.
The simulation trials we did gave us what I believe to be very valid results, and I was not surprised about the results. If we were only to do 25 trials we could get similar results but it is more likely that our results would jump around a bit more, in other words they wouldn't be as stable. If we did 50000 simulation trials though it would give us very similar results as the 500 trials did. The only difference we might have seen is that the edges may have had more results.
Histograms are very helpful to see the layout of your graph. The curved line shows you the high points and the low points of the data. It also shows you a constant line so your could easily estimate your results.
If we were to test the whole senior class I think we would get just the same results as we did with our statistics class especially since the senior group is not that large. One of the only things I would change is they would roll more dice so it could simulate all of them not just a few of them.
I agree that histograms are very useful for data analysis. The way they show the lines and bell curve really makes life a lot easier.
DeleteI think that the experiment that was done in class was valid. It had almost exactly the right amount of people that we suspected based on pure guessing. We could go test a larger population of people to increase the validity of our results. This way their is more data to use and draw conclusions from.
ReplyDeleteI didn't have issues with any part of the experiment. I felt that the simulations of the dice was explained extremely well, as was the aspects of the actually taste testing.
I believe the simulation resulted in extremely good results. We got just what was expected. If we had only done 25 simulations than the results might be skewed because of the small amount of data that was collected. Doing 50,000 would show us a even better trend in the results. The more data that is collected the more likely the probability that is estimated is going to show through.
Histograms are a great way to display data because it is visual. The bell curve allows you to actually visualize the standard deviations along with the bars representing each number rolled. This visual representation makes drawing conclusions easier.
If we were to test the senior class the number of dice would need to be raised greatly. The experiment would be extremely time consuming by hand, and I would suggest that we get possibly a computer program to run simulations instead. We would still expect about 1/3 of the seniors to guess correctly, but it still might be skew because of the small number of seniors.
I agree wi you that the more test trials the more valid the test would turn out
Delete-Yes, I think the experiment was valid. The histogram would have probably stayed relatively the same no matter how many trials we did, so I don't think it would have made too much of a difference if we did more. I can't think of a way we could have improved it.
ReplyDelete-No, I think I pretty much understood what we were doing. The instructions and procedure was pretty easy to follow.
-Yes. Like I said, I don't think there would have been much of a change in the overall results no matter how many trials we did, so I think the results were quite legitimate.
-The histograms are useful to the visual learner, as the curved line shows the distribution of numerical data.
-If we were to taste test the entire senior class, we would have to increase the number of dice being rolled; however, I think the results would basically stay the same.
I agree that the results would stay the same if we were to test the entire senior class, and I agree that the procedure was easy to follow.
DeleteI think the experiment was valid, but I think we should have done more trials for each person. I think it proved that most people don’t have that ability to actually differentiate between the two drinks.
ReplyDeleteNo, I thought the activity was very straight forward, and I understood it all.
I think the results of the simulation were very good because there were so many trials. Also, at the end, the results were exactly as we had predicted. I don’t think that it would have changed too much for however many taste-test trials we did. I believe that the percentage of right guesses would always be about the same for any population. Of course, with the fewer trials, the answers are a little more random because it really isn’t a lot of trials. By the time we got to 50,000 trials, the results would be following a certain shape already and all the new trials would just make the numbers higher.
Histograms are great ways to visualize our data because they help us to actually see our results in about 1 second whereas we could have stared at a chart for hours and not gotten that same understanding of what we were seeing.
I think that we would have to roll more than just 19 dice for the senior class if we were to simulate them. We would need to have one die per person in the class. The results would be the same except that instead of rolling 6 of the “correct” numbers every time, it would be standard to roll a number equal to 33% of the senior class.
I agree that doing 50,000 trials would just end up making the numbers higher.
DeleteI do think that the experiment was overall valid. I'm not sure if this would increase the validity, but having people choose the option of saying they can't tell the difference would allow us to see the percentage of people who actually can but it wouldn't show the percentage of right answers compared to wrong guesses. I think I understood the experiment and wasn't lost. I do think the simulation had good results and I think that whether we did 25 or 50,000 taste test, the results would be very similar if plotted in a histogram and the percentage of wrong guesses versus right would probably be similar too. The histogram helped put results into a visual format and showed how the simulation and taste test compared and related to each other. If we tested the whole senior class, then we would probably use one or two groups of dice and have two groups roll them and count. I think that the results would be still similar to our class with a smaller percentage that could actually tell the difference between coke and Pepsi.
ReplyDelete-Yes I do think our experiment was valid. We made it so that it was very unlikely to tell the difference between the two sodas based on any visual inference that could be made. I believe that a way we could make it more valid would be to just repeat it. That is a major part of any experiment whether it be in a math class such as statistics or biology, a science class.
ReplyDelete-There was nothing about the experiment I didn't understand.
-Yes I believe the simulation resulted in good results. If we had only done 25 then the results could have possibly been different because there would be less chances to get different numbers. 50,000 test trials could have given us a clearer idea as to what the average would be since there would be more chances to guess right or wrong.
-Histograms are a good way to visualize test results because it shows the average or median number of people that guessed it right. It also can show us the extremes of the results.
-The whole process would take a very long time to do since there would be 50-60 test subjects. And I agree with the earlier posts that's computers should be used to count the dice so that there would be less error. The expected outcome would probably me most of the class guessing incorrectly expect for some random lucky guesses mixed in.
-Yes i feel like the experiment was valid. I feel like the experiment would of been more valid if the sodas were cold instead of at room temperature.
ReplyDelete-There wasn't anything I didn't understand.
-Yes I believe the simulation resulted in good results. If we had only done 25 test trials the outcome probably would have been slightly different. 50,000 test trials i feel like would be way too many, you wouldn't be able tell the difference after that much taste testing.
-Histograms are easier for people who are visual learners.
-The entire senior class would take more time than it took to just do 19 students, but I think the outcome would be the same.
I do think this was a very valid experiment. However, I think if the drinks had been cold rather than warm it would have made a difference.
ReplyDeleteI think the instructions were pretty clear and the experiment was easy to follow.
With 25 trials, I think the results may have been somewhat similar but slightly off since that's only a few compared to what we did. 50,000 would have ended with very similar results, except thre would have been much more data.
Histograms are a great way of visually showing the distribution of data.
When testing the seniors,you'd need to add a die for each student in the grade. I feel that the results would still be similar to our class alone. Even still it would not be enough information to determine that most people in general cannot tell the difference.
I thought the test was valid, but if I could change something, it might have been the temperature of the drinks, like what Rachael mentioned earlier in class.
ReplyDeleteI understood the taste test because it was very well explained. Every part of the test was clear to why we would do that in such a situation. The steps made a lot of sense.
Since the histogram is peaked at the mean of correct guesses, it was a 'perfect' bell shaped histogram. It showed the distribution of data accurately, and in this case, our prediction was correct.
For the senior class I feel like the results would be the same just more
I believe that the test was valid, however the results would have altered some, had people took smaller sips as well as re-took sips if unsure about the 3 options. We could have also chilled the drinks, that would have made it easier to distinguish the difference of the two options.
ReplyDeleteI do believe that you made each part of this experiment clear.
Since the bell curve was on point I believe that in other situations the results would be about the same. The bell curve may alter a little but mostly stay about the same. Since I am more of a visual learner it was easier for me to determine the results.
I believe that if we added the whole senior class the results will not change drastically.
I believe that the test was valid because the results hit the 1/3 chance of people guessing it right. The chances and the probability seemed spot on in the way that it was presented in the results. I think that histograms are a great way to present the data in these type of trials, and show the results in a way that is easy to see and understand. I wouldn't change much about the experiment except expand the numbers of the people that tried the taste test. I feel that if you get more people to do it, then you might have more people that guesses right and that it would make it more accurate. But, I don't think it would change much.
ReplyDeleteThe test results for the taste test for the sodas I feel was valid although it will vary depending on how people drank the soda and the temperature of the soda. If people drank it slowly then they might have had a better taste of what the drink was than the one who drank quickly.
ReplyDeleteThe activity itself was very easy to understand and I had no problems understanding it.
I feel that the test was valid for the senior class as our class is fairly small. Although if there were more of the junior class in stats it may be more valid for them. If we increased the number of trials we would have much more valid results for a larger group of people.
The curve on the histogram makes the graph type very good for analyzing a lot of trials for an experiment.
The senior class because it is small would probably not change the results very much if at all.