(0.01 cm) it is obviously reasonable to call the y-intercept zero, Other cases may not be so clear cut, The
{rule of order when trying to determine whether or not a direct linear relationship is indeed a direct,
to ask yourself what would happen to the depenclent variable ifthe independent variable were
zero. In many cases you can reason from the physical situation being investigated whether or not the graph
should logically go through the origin. Sometimes, however, it might not be so obvious. In these cases we
Will employ a rule of thumb which is used only for this course and has no real statistical validity. Itis called
Csimply says that i the y-intereept does not exceed 5% of the maximum value of your y-axis
wwe will assume that the y-intercept is negligible and will call itzero. IF the y-intercept exceeds this
arbitrarily chosen value of 5% of the maximum y data, we will assume that it has some physical significance
‘and will go about trying to determine that significance. Apply the 5% rule to the data of the previous two
spring experiments and see what it yields.
‘Square Relations~Top Opening Parabolas
‘When the data you collect is non-linear and looks like it might be parabolic like the graphs illustrated in
graphs 3 and 4, you will employ a powerful technique called linearizing the graph to determine whether or
not the graph is truly a parabola with its vertex at the origin, Consider the data and graph below for an
experiment in which a student releases a marble from ay
rest allowing it to down an inclined track. ‘The Position vs. Time for a Rolling Ball
student notes its position every 1.0 seconds along the
track. a
180
Data Table - Rolling Ball Experiment] i¢5,
[time | position ‘140
$r20
position
0.0 1.0 20 3.0 40 50 60 70 60
time(s)
Since the graph is not linear for this experiment, you cannot determine the equation that fits the data using
only the techniques shown for the previous graphs. The slope ofthis graph, for instance, is constantly
changing. Notice, however, that the trend of the graph shows that as time increases in constant increments,
position increases in greater and greater increments. This is why the graph bends away from the time axis.
‘Since position increases ata greater rate than time, is there any mathematical manipulation of the data that
could be performed which would allow us to plot a linear graph? Notice that the graph looks like it might be
a parabola. Since the quadratic equations yield parabolas when plotted, and take the form y = ax? + bx + cy
Wwe might look to this form for a hin, Firsi of all if tis is a parabola, i appears to have its vertex atthe
origin, When the vertex of « parabola is at the origin, the values of b and ¢ are zero and the equation reduces
toy = ax. If we think that this graph is parabolic with a (0,0) vertex, we might try fo manipulate the data
based on the form y = ax. Think about it. If position is increasing at a greater rate than time, isnt it possible
that squaring time, which will make it increase at a greater rate, might make it keep up with position? This
reasoning is the basis of creating a test plot.
Basically, a test plot is a graph made with mathematically manipulated data for the purpose of testing
‘whether of not our guess about a mathematical relation might hold true. Since we think that the graph above
may represent a square relationship (parabola), we will make a new table in which we will square every time
value. We square time because itis the variable which is not keeping up in this case. We will then piot a
‘graph of position vs. time? to determine whether or not our hunch was right. If we are correct, our test plot
Experimental Design and Graphical Analysis of Data Rex P. Rice-2000 Poge 10