This
is the most difficult section to get right.
Credibility as an experimenter is hard-won here and so easily lost
here. There are key elements of a good
discussion that must be there. There are
other elements that may be added which enhance the experimenter’s
credibility. However, respect must be
paid to the reader by not rambling on too long.
Finally, the organization of the points is important.
·
Errors made in executing
the procedure
·
Innovations used
during the procedure of this experiment
o
Innovations that
worked
o
Innovations that
proved unfruitful
·
The largest
contributor to the uncertainty of this experiment
·
Improvements to
the experiment that will decrease the largest contributor to the uncertainty
·
Secondary
significant contributors to the uncertainty
·
Improvements that
will improve the efficiency of the experiment
·
Improvements that
will reduce the cost of the experiment
·
Possible
directions for further investigation that you would find interesting, and why
·
The procedure for
this experiment was not followed correctly, leading to results that are
meaningless. The oven was not switched
on, so the temperature was not controlled or recorded. The ambient temperature inside the oven when
not switched on was some 120 degrees C below the temperature of interest for
the phenomenon under investigation.
·
The results of
this experiment substantially agree with the expected results. Most of
the calculated values for g agree within uncertainty with the expected value,
and those that don't are within 2 times the uncertainties away from the
expected results.
· The results of this experiment are within 10% of the expected results, but are are more than 5 uncertainties away from expected results. This disagreement is systematic. If the results were all scaled lower by 10%, they would agree within uncertainty with the expected results. No particular element of the procedure or equipment has yet been identified as the cause of this discrepancy, it is suspected that the discrepancy is not substantial. The procedure will be repeated with fixed parameters, and each element of the apparatus and procedure will be investigated thoroughly to identify the source of this discrepancy.
·
In an attempt to
stabilize the bob and keep it from spinning, we added a second string beside
the first one. Unfortunately, although
it did stabilize the bob, it also prevented the bob from entering a circular
orbit.
·
The largest
source of uncertainty for this experiment is clearly the measurement of the
temperature, which contributed approximately 10 % to the final
uncertainty. By using a digital
thermometer, and better thermal insulation, the contribution from this source
could be reduced to 1%. If this is done,
however, then the uncertainty in the timing will become significant, given that
it currently contributes about 2% to the uncertainty.
·
The cost of this
procedure could be greatly reduced by the use of the new plastic pipettes
instead of the glass ones. The cost
savings per pipette would be substantial without impacting the results at all.
·
The experiment
could easily be automated, allowing for the procedure to be repeated 10,000
times without intervention of the experimenter.
This would require a capital outlay of about $3000, and a week to set
up. 10,000 trials would take 3 weeks to
complete, but not involve any significant man-hours, and would result in the
uncertainty of the measurement being reduced by a factor of 100.
·
This experiment
can easily be continued to the micrometer domain. Theoretically, this process should scale
linearly with length, but clearly at some small-scale limit the Thompson effect
should become discernable. It would be
interesting to look for the Thompson effect at this scale.
If
you made mistakes in analyzing the data, correct these errors. If data analysis is incomplete or if you are
not confident of your analysis, then complete the analysis, and check your
work.
If you simply CANNOT complete the analysis, then as a
desperate last resort you must state that the analysis is preliminary and unreliable,
and you must state this absolutely first. You are not entitled to
conclude anything except that this is a work in progress.
Frankly,
anyone who does this sounds completely incompetent.
·
I was feeling
sick when I was doing this experiment, so I don’t know how good it is.
·
I am under-paid.
·
Some sources of
error are: The moon was in half-phase.
This might have affected my value of ‘g’. ….
·
I think I
calculated the slope of the graph wrong.
For
each of the experiments that you have done, write any items that you think
might belong in a discussion. Write
these in point form and in any order.
Now mark the ones that, on second look, do NOT belong in a good
discussion.
Once
the elements of the discussion are identified, they need to be organized into a
coherent work (NOT A POINT-WISE LIST).
This is a literary skill. However,
there is an over-riding principle that should be followed. Statements that enhance the credibility of
the experimenter the most should go first.
First impressions are important.
Some
people immediately jump to the wrong conclusion based on the previous
statement, and put statements like “The procedure used was incorrect”
last. The last thing one wants to admit
is that something was done wrong.
However, this is a SERIOUS mistake when it comes to these types of
reports. It is the FIRST responsibility
of investigators to route out mistakes, and most especially their own. Clearly, a report that is being issued based
on work that was done wrong is an interim report. Such reports are expected to uncover possible
mistakes, and ways of reducing uncertainties.
With
these thoughts in mind, organize the elements you have identified in the above
exercise into an order that puts the items that enhance your credibility the
most first.