The National Student Research Center

E-Journal of Student Research: Math

Volume 4, Number 1, June, 1997


The National Student Research Center is dedicated to promoting student research and the use of the scientific method in all subject areas across the curriculum, especially science and math.

For more information contact:

John I. Swang, Ph.D.
Founder/Director
National Student Research Center
2024 Livingston Street
Mandeville, Louisiana 70448
U.S.A.
E-Mail: nsrcmms@communique.net
http://youth.net/nsrc/nsrc.html



TABLE OF CONTENTS

 

  1. Does The Divisibility Rule For Three Work For Other Numbers?
  2. Measuring Fractal Dimension of Partial Aggregates
  3. Is the Formula For Finding the Area of a Triangle Always Accurate?
  4. Fractal Resistors
  5. Does The Size Of A Circle Affect The Value Of Pi?

TITLE:  Does The Divisibility Rule For Three Work For Other
        Numbers?

STUDENT RESEARCHER:  Amanda Senules
SCHOOL:  Mandeville Middle School
         Mandeville, Louisiana
GRADE:  6
TEACHER:  John I. Swang, Ph.D.


I.  STATEMENT OF PURPOSE AND HYPOTHESIS:

I would like to do a scientific research project to determine 
if the proven divisibility rule for three works with other 
numbers, like 2, 4, 5, 6, and 7.  The divisibility rule for 
three is that you add the digits of a number together, and if 
the sum is divisible by three, the original number is divisible 
by three.  My first hypothesis states that the divisibility 
rule for three works.  My second hypothesis states that the 
divisibility rule for three does not work with the number 2.  
My third hypothesis states that the divisibility rule for three 
does not work with the number 4.  My fourth hypothesis states 
that the divisibility rule for three does not work with the 
number 6.  My fifth hypothesis states that the divisibility 
rule for three does not work with the number 7.  My sixth 
hypothesis states that the divisibility rule for three does not 
work with the number 5.

II.  METHODOLOGY:

First, I wrote my statement of purpose and a review of 
literature on divisibility rules.  Next, I developed my 
hypothesis.  I then developed my methodology to test my 
hypothesis.  After that, I made my data collection sheet.  I 
then began my experiment.  

Step 1:  First, I added the digits of the dividend 1,234 until 
it made one digit.  1+2+3+4=10  1+0=1

Step 2:  Next, I divided the sum 1 by 3.

Step 3:  It did not divide evenly.  The proven divisibility 
rule for 3 stated that if three doesn't go evenly into the 
digit divided from, the original number (in this case 1,234) is 
not divisible by 3.  

Step 4:  I then performed long division to check.  I repeated 
the process with the divisors 2, 4, 5, 6, and 7.  After that, I 
randomly picked three other dividends and repeated the process.

My variable held constant was the process shown above.  My 
manipulated variables were the 6 different divisors and 4 
different dividends.  My responding variables were the answers 
and if the process works on all experimental numbers.

After the experiment, I completed my analysis of data, summary 
and conclusions, and application.  Finally, I published my 
research in the journal, The Student Researcher. 

III.  ANALYSIS OF DATA:

For my experiment, I randomly picked 4 dividends, 1,234, 8,046, 
2,469, and 4,733.  

Using the divisor 2 with the divisibility rule for three, I 
found that it worked for one of the dividends, but not for the 
other three.

Using the divisor 3 with the divisibility rule for 3, I found 
that it worked for all of the dividends I used.

Using the divisor 4 with the divisibility rule for 3, I found 
that it worked for 3 of the dividends, but not for the other 
one.

Using the divisor 5 with the divisibility rule for 3, I found 
that it worked with all of the dividends I used.

Using the divisor 6 with the divisibility rule for 3, I found 
that it worked for 3 of the dividends I used, but not for the 
other 1.

Using the divisor 7 with the divisibility rule for 3, I found 
that it worked with all of the dividends I used.

With the 4 randomly picked dividends, the divisibility rule for 
3 worked with the odd divisors I used and not the even.

V.  SUMMARY AND CONCLUSION:

After doing my experiment, I found that with the dividends I 
used, the divisibility rule for three works with the divisors 
3, 5, and 7, but not with the divisors 2, 4, and 6.

In conclusion, my experiment did not prove a lot.  For example, 
none of the dividends I used were divisible by 5, and the 
chances of the digits adding up to a digit divisible by 5 are 
one out of nine.  If I would have used a dividend divisible by 
5, it probably would have proven that the divisibility rule 
does not work with the divisor 5.  For example:  1,235 would 
become 1+2+3+5=11  1+1=2

Five does not go into two evenly, so, according to the 
divisibility rule for three, 5 won't go into the original 
number evenly.  Really, 5 does go into the number.  I have just 
proven that the divisibility rule does not work with the number 
5.

To prove a divisibility rule right, you must test it with 
thousands of dividends.

With the dividends I used, I accept my first hypothesis which 
stated that the divisibility rule for 3 works.  I accept my 
second hypothesis which stated that the divisibility rule for 3 
does not work with the divisor 2.  I accept my third hypothesis 
which stated that the divisibility rule for three does not work 
with the number 4.  I accept my fourth hypothesis which stated 
that the divisibility rule for three does not work with the 
number 6.  I reject my fifth hypothesis which stated that the 
divisibility rule for 3 does not work with the number 7.  I 
reject my sixth hypothesis which stated that the divisibility 
rule for three does not work with the number 5.  

V.  APPLICATION:

I can apply my findings to every day life by telling people not 
to use a divisibility rule they made up without thoroughly 
testing it first.  I can also tell people not to use a 
divisibility rule for one divisor with another divisor.



TITLE:   Measuring  Fractal  Dimension  of  Partial  Aggregates

STUDENT RESEARCHERS:  Noah Forbes, Katie Johnson, Brad Jones
SCHOOL ADDRESS: Belmont High School
                  221 Concord Ave.
                  Belmont, MA 02178
GRADE:  12
TEACHER:  Paul Hickman - phickman@copernicus.bbn.com


I.  STATEMENT OF PURPOSE AND HYPOTHESIS:

We wanted to investigate how much of an aggregate can be 
removed and still have the same fractal dimension.  This 
included a study of the limitations of the computer in 
measuring the fractal dimension of objects.  We thought that 
the computer measurement would lose accuracy as more of the 
fractal was removed.

II.  METHODOLOGY:

Using the five aggregates the class had grown and scanned into 
the computer, we measured their original fractal dimensions 
using the program "Fractal Dimension 5.1".  Then we measured 
subsequent fractal dimensions as we removed parts of the 
aggregate.  First, we removed half of the aggregate and 
measured the fractal dimension using both the box and the 
circle method.  The half of the aggregate we kept was 
representative of the whole aggregate.  Next, we cut only a 
branch from the aggregate and magnified it to 200% 
magnification.  We then measured the fractal dimension of the 
branch.  In two cases, we measured the fractal dimension of the 
branch at normal size to see if the magnification was what 
changed the fractal dimension.

All five of the aggregates were grown for around fifteen 
minutes from a .2 M CuSO4 solution.  We used Electrodeposition 
cells with a copper cathode and a circular copper anode about 
eight cm in diameter. 

III.  ANALYSIS OF DATA:

Our measurements of the other groups aggregates agreed well 
with their own (see Data Table fig. 1).  This consistency shows 
that our technique for measuring the aggregates was relatively 
accurate.  When measuring the half aggregates, we found that 
the fractal dimension measured by the box method was very close 
while the fractal dimension measured by the circle method 
differed significantly.  The measurements using the circle 
method were above the original in some cases and below in 
others.  When the measured fractal dimension was greater, the 
half aggregate was very full and would take up most of the 
circle when it was being measured.  When the resulting fractal 
dimension was lower than the original, the half aggregate 
tended to be very linear and would leave a lot of empty space 
when being measured. 

When we measured the branches at 200% magnification, the 
fractal dimensions were high for the box method and very low 
for the circle method.  The box method was too high because 
when the branch was magnified, the pixels were magnified also.  
As a result, the branch lost some of its original shape.  The 
shape became too square along the edges making the fractal 
dimension measured high.  We found that it was the 
magnification that made the box method too high by measuring a 
few of the aggregates at original size.  When the branches were 
at original size the box method measured the fractal dimension 
very closely to the original fractal dimension.   The circle 
method was very low because the branch was long and thin, and 
as a result left a lot of empty space when measured.  We found 
that the less linear the branch was the higher the fractal 
dimension using the circle method.  No branch, when measured 
with the circle method, had a fractal dimension very close to 
the original.

Data Tables:

Group                         Fractal Dimension
             Box Method                       Circle Method
         Class Results  Our Results  Class Results  Our Results

Au          -------        1.763        -------        1.750
C&J          1.60          1.720         1.67          1.658
Honey        1.780         1.780         1.842         1.734
Mush         1.7           1.726         1.8           1.661
PaLaTo       1.610         1.636         1.735         1.616

FD at 1/2 Aggregate

Au                         1.755                       1.539
C&J                        1.704                       1.854
Honey                      1.790                       1.485
Mush                       1.730                       1.852
PaLaTo                     1.622                       1.643

FD at Branch

Au                         1.829                       1.198
C&J                        1.820                       1.188
Honey                      1.827                       1.193
Mush                       1.779                       1.026
PaLaTo                     1.768                       1.491

Mush @100                  1.720                       1.028

IV.  SUMMARY AND CONCLUSION:

We found that, when using the computer program Fractal 
Dimension 5.1, the fractal dimension of an aggregate can be 
measured accurately using the box method, no matter how much of 
the aggregate is present.  We found the circle method to be 
affected more by the shape of the aggregate.  The circle method 
measures the whole aggregate well, but as pieces are removed it 
becomes less and less accurate.  The circle method lost 
accuracy as the aggregates became less and less circular.  
Since the circle method measures the Fractal Dimension by 
placing a circle over the object, it works best when the object 
that one is measuring is a circle.  Objects that are not as 
circular, such as the half aggregate and the branch, leave too 
much of the circle empty, so the resulting fractal dimension is 
not accurate.  We found that the computer was not perfect; 
however, it was more accurate than we hypothesized.

V.  APPLICATION:

Our research with fractals taught us that there exists 
limitations in the measurement of self-similarity in accordance 
to the medium in which it is measured.  The same result is seen 
in the magnification of a picture.  One can only magnify a 
picture so many times before the detail in the picture is lost.  
The same loss of detail occurs as the fractal is enlarged and 
as more of the original fractal is removed.


  
TITLE:  Is the Formula For Finding the Area of a Triangle
        Always Accurate?

STUDENT RESEARCHER:  Michael Phillips
SCHOOL:  Mandeville Middle School
         Mandeville, Louisiana
GRADE:  6
TEACHER:  John I. Swang, Ph.D.


I.  STATEMENT OF PURPOSE AND HYPOTHESIS:

I would like to do a scientific research project to see if the 
formula for finding the area of a triangle is always accurate.  
The formula for finding the area of a triangle is A = 1/2 bh.  
My hypothesis states that the formula for finding the area of a 
triangle will always be accurate on all triangles that I test.

II.  METHODOLOGY:

First, I wrote my statement of purpose and hypothesis.  Then I 
reviewed the literature on triangles and area.  Next, I wrote a 
methodology to test my hypothesis.

My manipulated variable was the size of the triangles and the 
angles of the triangles.  My responding variable was the answer 
to the formula.  My variable held constant was the formula.

I drew six triangles of different size on 1 cm graph paper.  
Then I computed the area of each triangle using the formula.  
Next, I found the actual area of each triangle by counting the 
number of square centimeters within it.

I then analyzed the data, wrote my summary and conclusion, and 
applied my findings to the world outside of my classroom.  Then 
I published my findings in a national journal.

III.  ANALYSIS OF DATA:

For the right triangle, the measurement of the area with the 
formula was 16 sq. cm. and the actual area was 16 sq. cm. For 
the acute triangle, both measurements of the area were 10 sq. 
cm.  For the obtuse triangle, both measurements of the area 
were 10 sq. cm.  For the scalene triangle, both measurements  
of the area were 12 sq.cm.  For the equilateral triangle, both 
areas were 1.5 sq. cm.  For the isosceles triangle both areas 
were 12 sq. cm.

IV.  SUMMARY AND CONCLUSION:

In my research, I discovered that the formula for finding the 
area of a triangle worked on all kinds of triangles.  
Therefore, I accept my hypothesis which stated that the formula 
would always be accurate.

V.  APPLICATION:

I could apply my findings to the world by using the formula for 
finding the area of triangles, because it is easier and always 
accurate.



TITLE:  Fractal Resistors

STUDENT RESEARCHERS:  Anna Hutchinson, Keith Waters
SCHOOL ADDRESS: Belmont High School
                  221 Concord Ave.
                  Belmont, MA 02178
GRADE:  12
TEACHER:  Paul Hickman - phickman@copernicus.bbn.com


I.  STATEMENT OF PURPOSE AND HYPOTHESIS:

We began by looking at Sierpinski's triangle (fig 1) and carpet 
(fig 2) on the computer program J. Gasket Meister.  This 
program allows one to generate a pattern with a few simple 
rules.  The gaskets were interesting to us and when our teacher 
mentioned an experiment building a triangle out of resistors, 
we decided to imitate the experiment with the carpet.  We 
wanted to find out more about the relationship between the 
triangle experiment and the carpet experiment.

The theory behind the resistor network is that there is less 
resistance to current with many resistors in parallel.  We 
tried to find a relationship between the resistance of the 
square and triangle networks at iterations 1, 2, and 3, and 
their Fractal Dimensions.  Our hypothesis stated that the slope 
of the ln-ln graph of the carpet experiment would be less than 
that of the triangle experiment because the carpet has a higher 
fractal dimension, and therefore there would be more parallel 
resistors and less resistance.

II.  METHODOLOGY:

We had the first three iterations of the triangle already built 
on circuit boards with twenty-seven 1000 ohm resistors; we 
built the first three iterations of the carpet on eight circuit 
boards.  Then we measured the resistance for the first, second, 
and the third iteration of the triangle with an ohmmeter.  In 
order to get accurate results, we disconnected the iterations 
from the rest of the network to measure them, and then 
reconnected them to measure the whole network.  We measured 
each side of each iteration and took the average resistance to 
get a better reading.  We used the same procedure to measure 
three iterations of the carpet. 

We then made a ln-ln graph of the average resistance vs. the 
number of resistors on a side (length) for the triangle and 
carpet: the slope of these graphs gave us a figure related to 
the fractal dimensions of the resistor networks.

III.  ANALYSIS OF DATA:

Our results for the triangle matched those of previous 
researchers; in fact, it matched exactly the mathematically 
predicted value, which was 0.737.  The slope of the graph for 
the results for the carpet was 0.613.  

The fact that we were able to get a straight line in the ln-ln 
graph for the carpet indicates that it is possible to replicate 
the triangle experiment with the carpet and get a slope.  The 
results show that the slope for the carpet was less than that 
of the triangle; therefore the rate of increase of the 
resistance was lower for the carpet.

IV.  SUMMARY AND CONCLUSION:

These results agree with our hypothesis.  We came to the 
conclusion that as the fractal dimension goes up, the 
resistance of that network goes down.  This conclusion makes 
sense because, if an object has a higher fractal dimension,  it 
would take more resistors to build a network of that object, 
and therefore the resistance would be lower.  If we were to 
measure a solid sheet of resistor, (which would have an FD of 
two) the resistance would be really low.  

V.  APPLICATION:

The research we did relates to resistance on the atomic and 
molecular scale, which scientists are trying to learn more 
about now.  It is the structure of these networks that 
scientists can use as a model for the imperfect patterns found 
in natural electronic materials.



TITLE:  Does The Size Of A Circle Affect The Value Of Pi?

STUDENT RESEARCHER:  Dana Blount
SCHOOL:  Mandeville Middle School
         Mandeville, Louisiana
GRADE:  6
TEACHER:  John I. Swang, Ph.D.


I.  STATEMENT OF PURPOSE AND HYPOTHESIS:

I am doing a mathematical research project to see if the Pi is 
affected by the size of a circle.  Pi is the ratio of the 
circumference of a circle to its diameter.  Pi is always equal 
to 3.14.  My null hypothesis states that the value of Pi will 
not be affected by the size of the circle.

II.  METHODOLOGY:

First, I wrote my statement of purpose and did my review of the 
literature on Pi, circumference, diameter, and circles.  Then I 
developed my hypothesis.  Then I wrote the following 
methodology to test my hypothesis.  

First, I gathered my materials, which included ten circular 
objects of different sizes, paper, pencil, ruler, calculator, 
and data collection form.  Then I took one of my circular 
objects and made a mark on the edge.  I placed the mark on a 
piece of paper and marked where the mark, on the circular 
objects, was on the paper.  I carefully rolled the circular 
object on the paper until the marked point touched the paper 
again. I marked this spot on the paper.   Then I measured the 
distance between the two marks on the paper to find the 
circumference of the circular object.  Next, I measured the 
width of the circular object to find the diameter.  Then I 
divided the circumference by the diameter to compute the value 
for Pi.  I recorded my data on my data collection form.  I 
repeated this process with all of my circular objects.  Next, I 
listed my variables as shown below. 

My variable held constant is the formula for Pi.  My 
manipulated variable is the size of the circular objects.  My 
responding variable is the value of Pi for the circles.

Then I analyzed my data using the information on my data 
collection form.  Next, I wrote my summary and conclusion where 
I accepted or rejected my hypothesis.  Then I applied my 
findings to everyday life.  Finally, I published my findings in 
The Student Researcher.

III.  ANALYSIS OF DATA:

I found that two out of the ten circular objects I used for my 
experiment had a Pi value of 3.14.  Two of the circular objects 
had a Pi value above 3.14.  Six circular objects had a Pi value 
below 3.14.  My average Pi value was 3.06.  These differences 
were due to the fact that my instrument of measurement was not 
accurate enough to find the true Pi value of 3.14.

IV.  SUMMARY AND CONCLUSION:

After I analyzed mt data, I found out with my instrument of 
measurement that the majority of the circular objects had a Pi 
value below 3.14.  Two out of my ten circular objects had a Pi 
value of 3.14.  Two of the ten circular objects had a Pi value 
above 3.14.  I therefore, accept my hypothesis which stated 
that the size of the circles would not affect the value of Pi.  
If my instrument measurement had been more accurate the value 
of Pi for the circular objects would have always been 3.14.  
This research should be repeated and more exact measurements 
should be taken to ensure that an accurate value of Pi is 
computed.

V.  APPLICATION:

I can apply my findings to everyday life by telling teachers 
and students that Pi always equals 3.14 if you use the right 
instruments that allow you to accurately measure a circular 
object's circumference and diameter.

© 1997 John I. Swang, Ph.D.