VINCENT LIU

LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER

Free Fall Lab
Performed: August 28, 2014

Introduction
Objectives

The purpose of the Free Fall Lab was to determine how fast a classroom-provided
clicker clicked. The clicker was driven directly by the AC power (alternating
current) delivered from a wall socket, and thus the click frequency should
accurately reect the standard utility frequency of the United States, 60 Hertz.
Therefore, the lab not only measures the reliability of the clicker, but also the
accuracy of the 60Hz standard.

Data Collection

The following gure depicts the schematic diagram of the experimental setup.
Data on the refresh rates of the clicker was indirectly measured by pulling a long
piece of paper tape through a clicker. The paper tape was held at on a tall table
and attached to a heavy object. At start, one team member would turn on the
clicker, another team member would release the paper that he/she was holding,
and a third team member would gently push the heavy object off of the table
shortly thereafter. To record the movement of the paper tape, a sheet of carbon
paper was placed below the clicker. As the paper tape fell with the heavy object, the
clicker continued clicking and pressing the carbon paper against it, creating a
series of dots. We identied the dot that corresponded to the rst click after the
object started to fall (i.e. the rst dot that was separate from the initial cluster),
1 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
aligned it to the starting end of a meter stick to the initial point, and measured the
position of each dot in centimeters. The paper tape reached its end before the
heavy object hit the ground, we did not therefore have to worry about cropping out
the nal data points. Since the paper tape was subject to a constant gravitational
acceleration, by comparing the distance between the dots and the corresponding
theoretical values, it was possible to determine the click frequency.

2 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Data and Analysis

Raw Data

Note: position in meters was calculated by dividing the position in centimeters by 100.

Time (t clicks) Position (cm) Position (m)
0 0.00 0.0000
1 0.70 0.0070
2 1.80 0.0180
3 3.20 0.0320
4 5.00 0.0500
5 7.20 0.0720
6 9.70 0.0970
7 12.40 0.1240
8 15.40 0.1540
9 18.50 0.1850
10 22.00 0.2200
11 25.70 0.2570
12 29.70 0.2970
13 34.00 0.3400
14 38.60 0.3860
15 43.50 0.4350
16 48.70 0.4870
17 53.90 0.5390
18 59.40 0.5940
19 65.20 0.6520
20 71.40 0.7140
3 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Error Analysis

This experiment is a very crude experiment that, in turn, gives a very crude
estimate. A rst category of error is error in the design itself. Our simplied model
of the experiment takes into account only gravity, when, in reality, there are other
forces at work. The table exerted a normal and therefore a frictional force on the
paper tape, and so did the clicker. In addition, pushing the heavy object off of the
table resulted in a horizontal velocity. It could also have made the block rotate,
resulting in more complex motion of the heavy object and thus the paper tape.
Even our presumed constant, gravity, is not constant. As Newtons law of Universal
Gravitation states,


A change in r due to the height of the table would change, albeit ever so slightly,
the gravitational acceleration. Of course, our lab is too crude for this to make a
difference. If we were to take into account all forces at work, we would be able to
get a complex model that better predicted the outcome. Thus, this would be
categorized as systematic error.

The other type of error, random error, is also present in this experiment. The
smallest unit on our meter stick was the millimeter, so the measurements were only
accurate up to half a millimeter. The dots themselves where sometimes large
enough to make it hard to decide which half of the millimeter the measurement
belonged to. Furthermore, another source of error could have originated from the
equipment used. The clicker may not have a stable frequency, or there may have
been a slight variation in the utility frequency, which in our analysis were assumed
constant. This is just a brief list of possible sources; there may also be sources of
error that were not taken into account such as the curl of the paper tape, if any
existed.

F = G
m
earth
m
object
r
2
= m
object
a
object
a
object
= G
m
earth
r
2
4 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
The Random Error As Pertains to Raw Data

Note: percent error was calculated as follows:

As one can see, the error was not terrible. Although it contributed more than 7%
on the second click because the position was very small, the numbers steadily
decreased, down to less than a tenth of a percent in the last four clicks.

Time (t clicks) Position (cm) Absolute Error (cm) Percent Error (%)
0 0.00 0 0%
1 0.70 0.05 7.143%
2 1.80 0.05 2.778%
3 3.20 0.05 1.563
4 5.00 0.05 1.000%
5 7.20 0.05 0.694%
6 9.70 0.05 0.515%
7 12.40 0.05 0.403%
8 15.40 0.05 0.325%
9 18.50 0.05 0.270%
10 22.00 0.05 0.227%
11 25.70 0.05 0.195%
12 29.70 0.05 0.168%
13 34.00 0.05 0.147%
14 38.60 0.05 0.130%
15 43.50 0.05 0.115%
16 48.70 0.05 0.103%
17 53.90 0.05 0.093%
18 59.40 0.05 0.084%
19 65.20 0.05 0.077%
20 71.40 0.05 0.070%
Absolute error ( )cm
Position ( )cm
!100%
5 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Graph of Raw Data
As one can see, the graph looks like a nice, smooth curve. As gravity is a constant
force, we hypothesize that this set of data is approximately quadratic. What this
means in terms of the click speed, however, is yet to be determined. The position
can only go so far in calculating acceleration, and a quadratic curve is harder to
analyze than a linear. To help with this, we can plot velocity as a function of time,
which can relate its slope, acceleration, to gravity and will presumably be linear
assuming gravitational acceleration is constant.
6 AP PHYSICS 1+
P
o
s
i
t
i
o
n

(
m
)
0
0.2
0.4
0.6
0.8
Time (Clicks)
0 5 10 15 20
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Calculating Velocity and Time

Calculating velocity is straightforward. You take one measurement, subtract it from
the last, and divide it by the time interval. I used central difference to approximate
the derivative and obtain the velocity. This is surprisingly accurate. For simplicity,
we will denote the larger time and the smaller time .

As it turns out, the error in the estimation of the velocity by the central difference,
is on the order of . Our estimate of is about 1/120 of a second, and we
know for a fact that is far less than a second, should be very small. Now,
here are the results using the smaller and larger values of time:

In comparison, both the non-central difference at the larger and smaller time
estimates have an error on the order of , so they are less accurate estimates. We
could use even more data points to get a better estimate, but that would add too
much complexity and decrease the number of data points.
x + !x x ! "x
f x + !x ( ) =
f x ( )!x
0
0!
+
f
'
x ( )!x
1
1!
+
f
''
x ( )!x
2
2!
+ o !x
3
( )
f x " !x ( ) =
f x ( )!x
0
0!
"
f
'
x ( )!x
1
1!
+
f
''
x ( )!x
2
2!
+ o !x
3
( )
f x + !x ( ) " f x " !x ( ) = 2 f
'
x ( )!x + o !x
3
( )
f x + !x ( ) " f x " !x ( )
2!x
= f
'
x ( ) + o !x
2
( )
!x
2
!x
!x !x
2
f x + !x ( ) " f x ( )
!x
= f
'
x ( ) + o !x ( )
f x ( ) " f x " !x ( )
!x
= f
'
x ( ) + o !x ( )
!x
7 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Velocity vs Time Table

Note: Displacement was calculated subtracting two adjacent measurements, "t was 1 for each
time, and speed was calculated by dividing "x by "t.

Time (Average Clicks) Time (!t Clicks) Displacement (!x) Speed(m/click)
0.5 1 0.0070 0.0070
1.5 1 0.0110 0.0110
2.5 1 0.0140 0.0140
3.5 1 0.0180 0.0180
4.5 1 0.0220 0.0220
5.5 1 0.0250 0.0250
6.5 1 0.0270 0.0270
7.5 1 0.0300 0.0300
8.5 1 0.0310 0.0310
9.5 1 0.0350 0.0350
10.5 1 0.0370 0.0370
11.5 1 0.0400 0.0400
12.5 1 0.0430 0.0430
13.5 1 0.0460 0.0460
14.5 1 0.0490 0.0490
15.5 1 0.0520 0.0520
16.5 1 0.0520 0.0520
17.5 1 0.0550 0.0550
18.5 1 0.0580 0.0580
19.5 1 0.0620 0.0620
8 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Velocity vs. Time Graph
Overall, this graph conforms fairly well to our expectations. As we expected, it was
fairly linear. The coefcient of determination is 0.995. This means that about
99.5% of the error in the velocity can be attributed to the least squares regression
line of velocity on time. Overall, it is in the writers opinion that this line looks so
linear that we can assume constant acceleration. However, the imprecision of this
data set will disallow any analysis of the residuals, as it is clear just from the graph
that they are quite random. We will do a regression analysis. Here are the
conditions:
Linear: this graph looks very linear.
Independent: yes, the velocity at one time can be viewed as independent from the
previous.
Normal: Below is a histogram of the residuals:



9 AP PHYSICS 1+
V
e
l
o
c
i
t
y

(
m
/
s
)
0
0.018
0.035
0.053
0.07
Time (Clicks)
0 5 10 15 20
y = 0.0028x + 0.0081
R# = 0.995
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Due to the small number of data points, the histogram does not look perfectly
normal, but there is no evidence against a normal residual distribution. Besides, the
regression t-test and t-interval are robust with regard to normality.
Equal variance: the data points do not seem to converge or diverge as time
increases.
Random: these events can be seen as random. (Randomized experiment)
The conditions are met. We can proceed.
Here is the Minitab output:
10 AP PHYSICS 1+
C
o
u
n
t
0
1
2
3
4
Value
-0.0025 to -0.0020 -0.0015 to -0.0010 -0.0005 to 0 0.0005 to 0.0010 0.0015 to 0.0020
Residuals
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Using 20-2=18 degrees of freedom, we get a critical t-value of 2.100922 for a 95%
condence interval.
Working it out,

In other words, we can be 95% condent that, for every click, the velocity increases
by from 0.001640m/click to 0.003884m/click. In other words, we can be 95%
condent that the acceleration in is between 0.001640m/click
2
to
0.003884m/click
2.
Now, we can nd the time per click by relating this acceleration
by gravity, g=9.81m/s
2
:

Not bad at all. We get a value of 59.587 clicks, compared to the 60 clicks per
second presumed. Here is the percent error calculation:

The percent error is only 0.688%, meaning our experiment was very accurate.
Now, lets repeat the calculation of the clicks per second with the upper and lower
bounds:

This is interesting. We can be 95% condent that there are between 50.248 and
77.328 clicks per second, and our center is 59.587 clicks. Although the
b t
*
SE
b
= 0.002762 2.100922 0.000534 ( ) = 0.002762 0.001122 = 0.001640, 0.003884 ( )
m/ click
2
a
grav
= 9.80665
m
s
2
= 0.002762
m
click
2
1s
2
=
9.80665
0.002762
click
2
1s
2
=
9.80665
0.002762
click
2
1s = 59.587clicks
Observed ! Accepted
Accepted
"100%=
59.587Hz ! 60.000Hz
60.000Hz
"100%= 0.688%
a
grav
= 9.80665
m
s
2
= 0.001640
m
click
2
1s
2
=
9.80665
0.001640
click
2
1s
2
=
9.80665
0.001640
click
2
1s = 77.328clicks
a
grav
= 9.80665
m
s
2
= 0.003884
m
click
2
1s
2
=
9.80665
0.003884
click
2
1s
2
=
9.80665
0.002762
click
2
1s = 50.248clicks
11 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
measurements were very accurate, they were not very precise. As one can see, as far
as we know, the true clicks per second could very well be as low as 50 or as high as
77. This is interesting, because it would seem unlikely to get such a close estimate
despite the wide spread.
Now, lets wrap it back up and analyze our original data: position. To do this, we
must develop a model that describes motion with constant acceleration.
We start with deriving velocity with acceleration:

This intuitively makes sense. You gain a velocity of a every second, and you start at
v
0
. Now, lets derive position.

In this case, is x
0
is 0 due to the denition of the 0 point. v
0
is the intercept of the
velocity-vs time graph, which happens to be 0.008076m/click. This is not zero; by
the time the rst click was measured, the weight (and paper tape) had already fallen
a small distance. We already have a, which is 0.002762m/click
2
. Thus, the equation
we have is:
a =
dv
dt
dv = adt
dv
v
0
v
!
= adt
0
t
!
v " v
0
= at
v = v
0
+ at
v =
dx
dt
dx = vdt = v
0
+ at ( )dt
dv
x
0
x
!
= v
0
+ at ( )dt
0
t
!
x " x
0
= v
0
t +
1
2
at
2
x = x
0
+ v
0
t +
1
2
at
2
x = 0.008076 ! clicks + 0.001381! clicks
2
12 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Theoretical Position Table

Time (t clicks) Position (cm) Position (m) Theoretical Position
(m)
0 0.00 0.0000 0
1 0.70 0.0070 0.009457
2 1.80 0.0180 0.021676
3 3.20 0.0320 0.036657
4 5.00 0.0500 0.0544
5 7.20 0.0720 0.074905
6 9.70 0.0970 0.098172
7 12.40 0.1240 0.124201
8 15.40 0.1540 0.152992
9 18.50 0.1850 0.184545
10 22.00 0.2200 0.21886
11 25.70 0.2570 0.255937
12 29.70 0.2970 0.295776
13 34.00 0.3400 0.338377
14 38.60 0.3860 0.38374
15 43.50 0.4350 0.431865
16 48.70 0.4870 0.482752
17 53.90 0.5390 0.536401
18 59.40 0.5940 0.592812
19 65.20 0.6520 0.651985
20 71.40 0.7140 0.71392
13 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Theoretical Position Graph
14 AP PHYSICS 1+
P
o
s
i
t
i
o
n

(
m
)
0
0.2
0.4
0.6
0.8
Time (Clicks)
0 5 10 15 20
Position (m)
y = 0.0014x
2
+0.0081x
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER

Wow. The theoretical position lines up almost exactly to the observed position.
There is some random error, but the line hits all of the points. Now, we look at a
residual plot:
This plot is interesting. Note that the central line is not x=0: the average residual
is slightly below it. This is not surprising because we did not use a best t line on
the quadratic formula. Nevertheless, there seems to be a trend. Although the writer
does not know exactly why that is, a hypothesis can be made: because this line was
not obtained by best t of the data itself but by best t of the central derivative,
there is obviously error (and slight bias) in the estimation. Thus, it must have
missed the curve slightly and consistently overestimated or underestimated
certain sections of the curve. This was not apparent in the original graph (the
graph looked like it t perfectly), but because this is such a zoomed-in view of the
residuals, much smaller differences can be discerned. It is the writers opinion that
this is simply due to error on the side of the model, but more research is needed to
verify this. Little can be further analyzed from the residuals, as the error in part of
the model seems to obscure any more subtle but still important errors that
pertain to the experiment.

15 AP PHYSICS 1+
P
o
s
i
t
i
o
n

(
m
)
-0.008
-0.005
-0.002
0.002
0.005
Time (Clicks)
0 5 10 15 20
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Theoretical Position % Error Table

Percent error determined by |observed-accepted|/accepted*100%

Although there are low percent errors across the board, the percent errors do follow
a slight trend, at least in magnitude. This reects the previous observation in the
analysis of the residuals.
Time (t clicks) Percent Error
1 25.98
2 16.96
3 12.70
4 8.09
5 3.88
6 1.19
7 0.16
8 0.66
9 0.25
10 0.52
11 0.42
12 0.41
13 0.48
14 0.58
15 0.72
16 0.87
17 0.48
18 0.20
19 0.00
20 0.01
16 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Alternative Analysis: Quadratic
Quadratic Regression Interpretation

Quadratic regression is quite a different beast than linear regression. It employs a
similar least-squares method that is very difcult to analyze, and unlike power
regression, it is not simply a transformed version of linear regression. Note that this
regression assumes constant acceleration, so if that condition is violated, all of the
following falls apart. Here is the Minitab output:







We know that , so 2Theta1 should be the acceleration. This gives
an estimate of 0.002724, very similar to our previous estimate, 0.002762.
Continuing with the calculation,

The percent error is shown below:

x = x
0
+ v
0
t +
1
2
at
2
a
grav
= 9.80665
m
s
2
= 0.002724
m
click
2
1s
2
=
9.80665
0.002724
click
2
1s
2
=
9.80665
0.002724
click
2
1s = 60.00076clicks
Observed ! Accepted
Accepted
"100%=
60.000 ! 60.00076
60.000
"100%= 0.00127%
17 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Thats right: the error is a tenth of a percent of a percent. That is insane.
Here is the condence interval:
Using 21-3=18 degrees of freedom, we get a critical t-value of 2.100922 for a 95%
condence interval.

Multiplying the whole thing by 2 to get acceleration, we get (0.0026724,0.0027748)
The condence interval looks more promising, too. We can be 95% condent that
the acceleration is from 0.0026724m/click
2
to 0.0027748m/click
2
.

Wow. This time, not only are we accurate, but we are also precise. We can be 95%
condent that the true frequency of the clicker is between 59.449 clicks/second
and 60.577 clicks/second. Basically, we know the clicks per second within 1 Hz.
Here is a quick calculation of the percent error for the upper and lower bounds:


This means that the experiment could not have been more than one percent off.
The accuracy and precision of this analysis explains the puzzle we encountered in
the velocity analysis: why were we so accurate (at 59.6) when we could have been
from 50 to 77? This is caused by the inherent imprecision of using the central
difference to approximate the derivative, scattering results and causing a greater
spread in points. Thus, we underestimated the precision of the results: in reality,
they were quite precise, as shown by the quadratic analysis. Though imprecise, the
central difference is generally accurate and thus produced a similar result to
quadratic analysis.
b t
*
SE
b
= 0.0013618 2.100922 0.0000122 ( ) = 0.0013618 0.0000256 = 0.0013362, 0.0013874 ( )
a
grav
= 9.80665
m
s
2
= 0.0026724
m
click
2
1s
2
=
9.80665
0.0026724
click
2
1s
2
=
9.80665
0.0026724
click
2
1s = 60.577clicks
a
grav
= 9.80665
m
s
2
= 0.0027748
m
click
2
1s
2
=
9.80665
0.0027748
click
2
1s
2
=
9.80665
0.0027748
click
2
1s = 59.449clicks
Observed ! Accepted
Accepted
"100%=
59.449Hz ! 60.000Hz
60.000Hz
"100%= 0.918%
Observed ! Accepted
Accepted
"100%=
60.577Hz ! 60.000Hz
60.000Hz
"100%= 0.962%
18 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Theoretical Position Table

Note: Theoretical position obtained by -0.0039797+0.0087386t+0.0013618t
2
Time (t clicks) Position (cm) Position (m) Theoretical Position
(m)
0 0.00 0.0000 -0.0039797
1 0.70 0.0070 0.0061207
2 1.80 0.0180 0.0189447
3 3.20 0.0320 0.0344923
4 5.00 0.0500 0.0527635
5 7.20 0.0720 0.0737583
6 9.70 0.0970 0.0974767
7 12.40 0.1240 0.1239187
8 15.40 0.1540 0.1530843
9 18.50 0.1850 0.1849735
10 22.00 0.2200 0.2195863
11 25.70 0.2570 0.2569227
12 29.70 0.2970 0.2969827
13 34.00 0.3400 0.3397663
14 38.60 0.3860 0.3852735
15 43.50 0.4350 0.4335043
16 48.70 0.4870 0.4844587
17 53.90 0.5390 0.5381367
18 59.40 0.5940 0.5945383
19 65.20 0.6520 0.6536635
20 71.40 0.7140 0.7155123
19 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Theoretical Position Graph
Again, the theoretical position lines up almost exactly with the points. This time,
because the line was a calculator line, we can show the R
2
value, almost 1. This
means that almost 100% of the error in the position can be attributed to the
quadratic least squares regression line of position on time. More important is the
fact that the quadratic regression looks exactly right for the curve. Other methods,
such as exponential or logistic, dont t quite as well.
20 AP PHYSICS 1+
P
o
s
i
t
i
o
n

(
m
)
0
0.2
0.4
0.6
0.8
Time (Clicks)
0 5 10 15 20
y = 0.0014x
2
+ 0.0087x - 0.004
R# = 1
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER

A look at the residual plot shows a similar, albeit less pronounced, pattern as the
previous residual plot. This time, there is no inaccuracy of the model to blame: the
best t line was calculated directly from the data. Again, there is no denite answer
to why there seems to be a trend. However, with a more accurate model in hand,
we can start making hypotheses. The largest residual was at 0. An explanation for
this is the act of pushing the heavy object off of the table, giving it a boost
compared to the ideal situation. Thus, the rst few points would have been
higher than predicted, exactly what shows here. After that, other than the
unexplained spike at 16, the graph seems to go slowly but surely downwards. This
may be the effect of friction and air resistance on the heavy object and the paper,
but it may be simple coincidence. Again, one must keep in perspective the
crudeness innate to the free fall lab and the tiny residuals, so all of this might be
fussing over nothing. Still, it is the writers opinion that this phenomenon can be
and should be investigated in more depth with a more precise experiment.


21 AP PHYSICS 1+
P
o
s
i
t
i
o
n

(
m
)
-0.008
-0.005
-0.002
0.001
0.004
Time (Clicks)
0 5 10 15 20
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Theoretical Position % Error Table
Percent error determined by |observed-accepted|/accepted*100%

This table conrms the accuracy of the model. There are low percent errors across
the board (with the exception of the rst point), just like the other model.
Time (t clicks) Percent Error
0 100.00
1 14.37
2 4.97
3 7.23
4 5.24
5 2.38
6 0.49
7 0.07
8 0.60
9 0.01
10 0.19
11 0.03
12 0.01
13 0.07
14 0.19
15 0.34
16 0.52
17 0.16
18 0.09
19 0.25
20 0.21
22 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
Conclusion

Results

The results were extraordinarily good and just as expected, perhaps surprisingly so.
We investigated two methods, one dealing with the calculation of velocity and then
acceleration and one dealing with extracting acceleration directly from position. In
our rst method, we found an acceleration of 0.001381 m/click
2
, an initial velocity
of 0.008076 m/click, and a frequency of 59.587Hz. In our second method, we
found an acceleration of 0.0013618 m/click
2
, an initial velocity of 0.0087386 m/
click and a frequency of 60.00076Hz. The percent errors were 0.688% and
0.00127% respectively. The greatest difference between the two method, however,
was not their accuracy but their precision. The rst model gave plausible values of
50 to 77 Hz, while the second narrowed the range down to a mere 59.5 to 60.5.
Overall, it is the authors opinion that quadratic analysis is more reliable than
velocity analysis, as it provides a more accurate view of the true precision due to its
direct nature (especially when compared to the inherent imprecision of the central
difference). Even with all these imperfections, it is safe to say that the clicker really
was operating at 60Hz or something very close to it and that the US standard was
enforced relatively well, at least in the environment tested.

Error Analysis

Overall, this lab found no evidence of systematic error. Both methods gave a 95%
condence interval that included 60Hz, even the 1Hz-narrow interval of the
second model. This is surprising: there are physical forces at work that would
contribute to systematic error such as friction, air resistance, and the constant
horizontal velocity after pushing the heavy object off of the table. They might be
small, but I did not predict that the results would be so close to the prediction.
However, there might be just a trace of systematic error as shown by the residual
plot. We will use the second model because it better approximates the curve and is
a more direct analysis of the data. First is the point (0,0). This point was a perfectly
23 AP PHYSICS 1+
VINCENT LIU
LAB PARTNERS: JOSE MANUEL CORICHI GOMEZ, MIA CAVENER
normal point in the rst model, but in the second, it was a bit of an outlier with
respect to regression, especially compared with the others. I believe that this error
can be attributed to the act of pushing the heavy object off of the table, as this
would cause the rst few points to be a bit farther than it would otherwise. Indeed,
this is somewhat true, but it could be a coincidence. Next, it seemed that the model
started out under predicting, but, after a while, started over predicting. The under
predicting part could have been due to the pushing off of the table, but as friction
slows the object, it may have caused the over predicting of the model. Again, this is
just splitting hairs, but it may have some signicance. There is also the question
about the spike at 16, but it is the authors opinion that this was random error and
is not a repeatable outcome.
This does not mean that there was a lack of random error. Random error was still
present, whether from the size of the dot made, the inaccuracy of the meter stick
used, or something along the lines of the paper tape curling in the experiment.
However, random error did not really affect the results because the number of
points was in excess of 20.

Future Research

Although this lab had remarkably accurate and precise measurements taking into
account the crudeness of the experiment, more precise experiments should still be
done. These involve more precise equipment and taking into account frictional
forces, as well as a better-controlled environment. Experiments of that nature could
truly investigate the tiny discrepancies and trends that were discussed in the analysis
of the residual plots. A different look at this scenario, consistency, is also possible.
Instead of analyzing how fast the clicker clicks, one could analyze how consistent it
was in speed. Another extension could be doing the same experiment in a region
with a different utility frequency such as Europe.
24 AP PHYSICS 1+