Frustrated student and mentor!

[keratin]

I am mentoring a student who wants to compare Al to wood bat performance. The student, through practical knowledge and background research, is familiar with baseball bat dynamics. Both of us have reviewed the helpful materials in Sports Science related to comparing aluminum versus wood baseball bats. My student originally wanted to compare sweet spots (ss)-a clean, solid experiment.

His M.S. teacher insists he analyze a single, random point. His argument is that using multiple points introduces more than 1 independent variable. Any thoughts on how to convey that 1pt is not representative on any bat, let alone to make a comparison between Al and wood? Thanks much

[Craig_Bridge]

If I'm reading you correctly, the M.S. teacher involved does not believe that you and the student have a sound scientific basis for your experiement by only using the sweet spot. If I may translate your and the student's concern with the M.S. teacher's experiemental design:

The use of an arbitrary point on a bat will lead to unfair comparisons of bat performance.

You really have only one choice to prove your concerns and that is to use the scientific method construct a scientific experiment to prove it. If I were faced with this challenge, I would run an experiment on two different wooden bats (length/shape) and one aluminum bat using serveral impact spots measured from the handle end (being sure to pick the sweet spot on each bat as one of the common points).

From this, you should be able to prove that the sweet spot of each bat produces the best results for that bat and that you know how to determine what the sweet spot is. Once you have done this and it turns out that the sweet spot behaves as advertised, then you can run another experiment to do what you really want to do only using the sweet spot and be on a strong scientific footing because you have proved your methodology is valid.

Once you are into comparing wood vs aluminum using only the sweet spot performance, you are going to have to use multiple bats of both types with different profiles in order to draw any conclusions about the material effects.

[Louise]

I am mentoring a student who wants to compare Al to wood bat performance. The student, through practical knowledge and background research, is familiar with baseball bat dynamics. Both of us have reviewed the helpful materials in Sports Science related to comparing aluminum versus wood baseball bats. My student originally wanted to compare sweet spots (ss)-a clean, solid experiment.

His M.S. teacher insists he analyze a single, random point. His argument is that using multiple points introduces more than 1 independent variable. Any thoughts on how to convey that 1pt is not representative on any bat, let alone to make a comparison between Al and wood? Thanks much

I think you are talking about a variation on this project?
http://www.sciencebuddies.org/mentoring/project_ideas/Sports_p016.shtml?from=Home

If you are always counting hits from the sweet spot on the bat, I think that is a controlled variable. The manipulated variable would be bat composition and the responding variable is the distance.

I think what you are saying is that the teacher argues all positions on the bat should be used, and not just the sweet spot. [thus, position on the bat does not need to be a controlled variable] You could argue that this should be controlled variable because the physical response/vibrations are optimized BUT did you actually find the sweet spot of the two bats, or did you just assume it was 5-7 inches from the bat end? Supposedly they are different for wood vs. al... You could also make an arguement from a more general point of view that the more variables you can control, the more accurate the experiment will be, and then cite the data on 'sweet spots' for why bat position is important.

I think I understand the teachers concern though. Basically, you have a situation where the student is discarding a certain amount of data (depending on how good a batter he/she is, I guess that could be a pretty high percentage of 'non-sweet' hits!) It is not a great experimental design to throw out some (potentially large) fraction of data. How are you judging where the ball hits the bat? Photograph? Independent observer? Opinion of the batter? The more rigorous this part is defined, the better the experiment is. Otherwise, this step is very subjective and could lead to a skewing of the data.

I would recommend collecting the data for all hits. Don't discard any data. Then, you can analyze the data as a whole and specifically the sweet spot data. You can tell if the sweet spot data _is_ less variable than the data as a whole. You can also tell if the advantages of a specific bat are general, or only apply to a certain batting style (i.e. someone who connects at the sweet spot).

I hope this helps. If I misunderstood what the issue is, please let me know. I was a little unclear on the 'insists he analyze a single, random point' part.

Louise

[Craig_Bridge]

I should have stated that I'm assuming you are using some consistent pitching machine and bat swinging machine so that the impact points are consistant as well as the relative velocities of the ball and bat at impact.

If there are human factors involved, then you are going to have a very hard time saying anything about a bat material difference without a huge number of samples.

[keratin]

Craig, thanks for your quick reply. My student originally wanted to simply compare the sweet spots on an Al and wood bat of the same length. The M.S. teacher told him to just pick an arbitrary point, not even the sweet spots. Agreed it is not a fair test. A review of the lit shows multiple point comparisons, above, below and including sweet spot in order to acquire enough data to make a comparison. His argument is that testing at multiple points along the bat barrel introduces more than 1 independent variable.

My student plans on dropping the ball onto each bat's barrel along a series of sections and measuring how high it bounces. This is similar to an approach used by the University of Mass. at Lowell Baseball Research Center.

[keratin]

I am mentoring a student who wants to compare Al to wood bat performance. The student, through practical knowledge and background research, is familiar with baseball bat dynamics. Both of us have reviewed the helpful materials in Sports Science related to comparing aluminum versus wood baseball bats. My student originally wanted to compare sweet spots (ss)-a clean, solid experiment.

His M.S. teacher insists he analyze a single, random point. His argument is that using multiple points introduces more than 1 independent variable. Any thoughts on how to convey that 1pt is not representative on any bat, let alone to make a comparison between Al and wood? Thanks much

I think you are talking about a variation on this project?
http://www.sciencebuddies.org/mentoring/project_ideas/Sports_p016.shtml?from=Home

If you are always counting hits from the sweet spot on the bat, I think that is a controlled variable. The manipulated variable would be bat composition and the responding variable is the distance.

I think what you are saying is that the teacher argues all positions on the bat should be used, and not just the sweet spot. [thus, position on the bat does not need to be a controlled variable] You could argue that this should be controlled variable because the physical response/vibrations are optimized BUT did you actually find the sweet spot of the two bats, or did you just assume it was 5-7 inches from the bat end? Supposedly they are different for wood vs. al... You could also make an arguement from a more general point of view that the more variables you can control, the more accurate the experiment will be, and then cite the data on 'sweet spots' for why bat position is important.

I think I understand the teachers concern though. Basically, you have a situation where the student is discarding a certain amount of data (depending on how good a batter he/she is, I guess that could be a pretty high percentage of 'non-sweet' hits!) It is not a great experimental design to throw out some (potentially large) fraction of data. How are you judging where the ball hits the bat? Photograph? Independent observer? Opinion of the batter? The more rigorous this part is defined, the better the experiment is. Otherwise, this step is very subjective and could lead to a skewing of the data.

I would recommend collecting the data for all hits. Don't discard any data. Then, you can analyze the data as a whole and specifically the sweet spot data. You can tell if the sweet spot data _is_ less variable than the data as a whole. You can also tell if the advantages of a specific bat are general, or only apply to a certain batting style (i.e. someone who connects at the sweet spot).

I hope this helps. If I misunderstood what the issue is, please let me know. I was a little unclear on the 'insists he analyze a single, random point' part.

Louise

Actually, the teacher is arguing that all points on the bat are equivalent, and that dropping the ball arbitrarily on a point along the bat barrel is representative of the entire bat, which is simply not the case. It is a variation on the sciencebuddies.org experiment. My student is using a video camera to capture the height and a device to drop the ball consistently on a spot. What he wanted to do is drop the ball with repeated trials ~ every few centimeters along the bat barrel similar to a comparison done between Aluminum bats made by Japanese mfg and US.

[keratin]

Louise, Forgot to say thanks for your reply. My apologies. I've included the link to Al bat comparisons.
http://m-5.uml.edu/umlbrc/Publications/ ... EA2004.pdf

The problem, my student and I both know and understand that a comparison of a single point (sweet spot or not) on the bat barrel does not an experiment make. As Craig noted it is not a fair test and no conclusions can possibly be reached save for one about that one singular point on the bat and even that is suspect.

[Louise]

(snip first paragraph)

My student plans on dropping the ball onto each bat's barrel along a series of sections and measuring how high it bounces. This is similar to an approach used by the University of Mass. at Lowell Baseball Research Center.

If your experiment is this controlled, I recommend doing exactly what the teacher suggests, pick one point and do the evaluation of materials properties. Does this relate to the actual physics of baseball? No, but it is a well controlled scientific experiment of some basic principles.

Obviously, what happens when a batter hits a ball is much more complicated than this. As you point out, people who study this problem do many more experiments in many different variations to try and understand the 'real world problem'. This is beyond the scope of a middle school science fair! That doesn't mean that the simpler experiment is worthless. It is more controlled, and targetted towards answering a simpler question.

It sounds like you have a pretty controlled setup; I had read the instructions to mean that there was an actual batter and pitcher. You could easily do a more advanced experiment. You could of course, collect data at multiple points and 'prove' the teacher wrong. Just be careful how you present it. [As in, don't present it that way! But, you could perhaps have a second hypothesis about whether the sweet spot is important; with sufficient data you could either show if there is a difference in bounce distance OR variability in distance] There is nothing wrong with having more than one manipulated variable; teachers at lower grade levels generally discourage it, since they are trying to teach the scientific methods to students who (mostly) have never been exposed to it. So, the recommendations are pretty 'cookie cutter'. My feeling is, as long as you satisfy the requirements of the project (i.e. have one complete set of data at a single point), you can safely go beyond the requirements/bend the rules (evaluating multiple points/comparision of sweet spots). But, you should have the single variable data broken out so the teacher understands that the project complies with the instructions.

Lastly, let me say again that what the teacher proposed isn't a bad experiment. It is a good beginner science project. It tests two materials, has a connection to a real world problem, and while it doesn't capture the scope of the real problem, it is an appropriately simplified problem for a young scientist. So, use your judgement here. Is the student you are mentoring going to be bored doing a simplified problem? Or is he or she going to get bogged down and frustrated by the more complicated multi-variable problem? Many middle school students struggle with the scientific method; adding extra variables would be too complicated. You know your mentee and his/her abilities/ scientific experience best.

Hope this helps!
Louise

[Craig_Bridge]

His argument is that testing at multiple points along the bat barrel introduces more than 1 independent variable.

The M.S. teacher either doesn't understand variables and variation as it applies to your area of investigation or is not expressing a sufficient understanding.

If one defines the sweet spot as that point on a given bat where the maximum energy transfer occurs, then if one determines where the sweet spot is on each bat under test, one has a controlled variable (e.g. offset from the sweet spot of the contact point going to zero), then the distance from the handle to the sweet spot is NOT an independent varible, it is a dependent variable!

No two wooden bats are identical even when they are the same shape/length/weight/balance point so you have a multi-variant problem to start with. Even wood from the same tree at different points is not identical even if you shape and balance it identically! Rotate the same wooden bat and you likely will get some variations even when you keep the distance from the handle consistent because of the grain and growth ring orientations affect the dynamics. The choice of the location where the logo is burnt into a major league bat is NOT arbitrary but is an indication of the rotation where the sweet spot is.

The whole argument for using aluminum bats in college baseball is based on being able to obtain much more consistent bats.

It sounds like you and your student have a reasonable grasp of the subject and just need to put together the experimental methodology to prove the existence of the sweet spot and get some college professors and/or experts from a company like Louisville Slugger to review your methods and write an opinion in order to help convince the M.S. teacher that you know what you are doing (along with the data involved in determining the sweet spot). The bury them in paperwork approach!

[keratin]

No two wooden bats are identical even when they are the same shape/length/weight/balance point so you have a multi-variant problem to start with. Even wood from the same tree at different points is not identical even if you shape and balance it identically!

Thanks for the fantastic insights into wood characteristics as it relates to bat performance, impact and design. I will pass on said "Think abouts" to my student. More importantly, thank you for your perspective on indy variable assignments. I agree it's not an independent variable. It's like swabbing multiple areas of a drinking fountain to ensure you are getting a representative sample. For most reasonable people it would be a winning argument. Unfortunately, it would only serve to antagonize said teacher. Multi-variant analysis is a good thing, being able to have enough data to make an analysis is a great thing!

[keratin]

There is nothing wrong with having more than one manipulated variable; teachers at lower grade levels generally discourage it, since they are trying to teach the scientific methods to students who (mostly) have never been exposed to it. So, the recommendations are pretty 'cookie cutter'. Louise

Thanks for your suggestions they were most helpful. Ironically, my student also suggested just doing what the teacher wanted but slipping in some additional points data in under the table for the judges to review!

My goal is to support the student and to help him stretch his talents...not limit them. I try to put myself in his shoes and consider how he will confront questions from judges and other folks (parents and other students) and can only shudder at what the finite data point graph might look like and how he's going to make a reasonable conclusion!

For his sake, we'll be elaborating on the experiment and as such give him the opportunity to conduct a more sophisticated analysis that isn't boring and provides him good talking points!

[bradleyshanrock-solberg]

There are indeed whole classes of problems that require several manipulated variables. People trained in modern engineering, process control or even management techniques often get very frustrated with science fair encouragement of only one variable.

The problem is that doing so properly requires different training than a middle school or high school student is going to have, including some grounding in statistics.

I think that it's impossible to always hit the same part of the bat on every swing, barring a "T" instead of a pitching machine and a robot swinging. You have to work with what you have.

To do this experiment so the "Sweet spot" is just one variable, all you need to do is have a lot of swings per bat. Given the same pitching machine and the same batter (with a way of discounting hits outside of the sweet spot), the variation based on where in the "good part" of the bat you hit will be damped out....each experimental "run" will have "samples" from all over the "sweet spot" and thus you can compare the total results from each bat to the total results of each other bat.

My recommendation would be something like this.

Pitching machine (so all pitches come in at same speed/angle, pitcher doesn't get tired)

Batter starts fresh every day with a new bat. (don't do 100 swings on one bat, then expect the 101-200 on a different bat to be fair to the new bat) At a minimum, batter needs to be able to do a fullset of swings without feeling "tired" and needs enough rest between bats to not have long term fatigue issues.

Do as many swings as you can per bat, without ruining the results with fatigue. Only count the results of the "sweet spot" hits.

Most importantly, document what you're doing to control the variation, and recognize that you'll need a pretty big difference between two bats to show that one is really "better". There is enough you can't control that your margin of error will be moderately high. Any bats that are reasonably close to each other in performance should be considered "the same within experimental error" and only those that are a lot better or worse can be considered significant enough to cause a valid conclusion.

In brief...if someone else repeats your experiment, they should get the same results. Present your approach and conclusions from data with that in mind.

[bradleyshanrock-solberg]

If the teacher is being very rigid about this....all I can think to do is set up some kind of mechanical experiment.

Put the ball on a "T" so no pitcher is involved.

Work out some kind of mechanical contraption to swing the bat.

Fix the "T" and the "Bat swinger" so that repeated swings don't change where the bat is hit.

test it that way.

What annoys me about this though is that it doesn't tell you anything useful. How a bat performs when you hit it in one spot is not useful to the batter. He wants to know how it will perform when HE swings it against a real pitcher. The statistical "sweet spot" test is something
any batter could use to find the right bat FOR HIM.

(after all, your "swing" has a big influence on a bat's performance,
and not all bats are suitable for all swings)

A test like you want to do is actually useful for any batter, even if he won't get the same results you did (because his swing is different).

[keratin]

What annoys me about this though is that it doesn't tell you anything useful. How a bat performs when you hit it in one spot is not useful to the batter. He wants to know how it will perform when HE swings it against a real pitcher.

Bradley, thanks for your helpful comments. I couldn't agree with you more about the utility of the experiment, or lack of same. While the dropping of the ball on the bat is not applicable to a real world scenario, it does make a comparison independent of a set of other variables (the batter's swing, power, force, etc.)

Unfortunately, changing anything at this juncture, even arguing for multiple variables, is nearly impossible with his teacher's mindset. My goal, nonetheless is to maximize the value of the endeavor for my student. He may not get a heck of a lot out of the actual experiment or be able to draw any conclusions of value using only one point, but he'll extend his study and research skills, focusing more on the theoreticals this go round.

P.S. It might be my process background in play, but us regular life scientists also find it frustrating dealing with one variable for science fair.

[bradleyshanrock-solberg]

If it was my student, I'd probably have him do it both ways.

One way to make his teacher happy (rig a mechanical contraption to test the bats with only one variable)

One way to make HIM happy - doing science with real world applications, using the statistical "sweet spot" approach.

Assuming the student had the time and energy.

Students do tend to get a lot out of these science projects that have nothing to do with science. Your kid is learning what it's like to deal with an unreasonable boss that has no knowledge of the thing he's making decisions about

Recognizing that someone is going nonrational and doing your best to work around it is a life skill, it's just not a very fun one.