Sponsored by

Here is a summary of tips to write better lab reports.  This is not a complete list but is comprehensive enough to improve your reports.

This collection is based on the past experience in reading and grading students'  lab reports for more than 3 semesters. To my knowledge, none of the following is an "unreasonable" expectation or "too-much".

NOTE: If you are from a different class/section not taught by Anbu, your instructor may have different expectations/opinions. Also your instructor may have different penalties.

Better lab reports

• Units
• The importance of writing units in reports can not be stressed just once or twice.  Without units, a number is meaningless in reports.  (eg. If you say the distance between your home and school is '500', what do I understand ? Is it 500 meters or kilometers - is it close or far ? But if you say, the distance is '500 meters', then I know that you stay very close to school.)
• Missing units will make you loose 1 point in the lab report, at each place.  But if you were in NASA, missing units make a difference of $125 million and loosing the Mars orbiter altogether.  See these news articles in 1999 at CNN and Space.
• Also keep in mind that ratios of same quantities do not have units (eg.  in a graph of 'X vs. X', the slope has no units, because both X have the same units.)


• Comparing two values
• Do not use qualititative words ('agree somewhat', 'fairly close', 'very close', 'acceptable' may make sense to you, but not to others who read your report.  How close is 'very close' ?)
• Instead use quantitative comparisons ('5% error', '+/- 7.3% ' are much better and are also informative.)
• Human error
• 'Human error' is neither 'systematic' nor 'random'.  Why ?  If a human makes an error occasionally in an experiment, it is a 'random' error.  If the error is made consistently, it is a 'systematic' error.
• So when do humans make an error ? The distinction is hard to make (can lead to a philosophical debate). If you know that you commit errors consistently, then why do the experiment ?  So never write 'human error' in lab reports. It is not a valid error. For this class, 1 point will be deducted if you write 'human error'.
• Graphs
• Never draw jagged lines in graphs.
• Graphs have to be single, smooth, continuous lines.  Use pencils and draw neatly.
• When there are more than two lines in a graph, label the plots to say which graph is for which experiment.
• Graphs must have their axes clearly labeled, with appropriate units. Axes with no labels and units will also loose 1 point.
• Mark the significant values/points in your graph (like maximum, minimum points on your graph). Reading a graph should be accurate and is not to be left as a guess work.  Failure to mark these values on your graph has a penalty of 1 point.
• Error Values
• Giving error in values (eg. an error of 5 meters) although looks acceptable, is not a good way to understand the system/experiment.
• Instead find errors in percentage values (eg. '1.53 % error' gives an idea how small or large the error is, depending on your experiment.)
• In this class, we shall define percentage error as,
( Measured - Theoretical ) * 100  / Theoretical
• Data
• Never fudge the data.
• Report the actual values/numbers you get in experiments.
• Fudging the data, if caught will be heavily penalized.
• Conclusions
• You should make conclusions based ONLY on your experiments or observations. Do not write what you like to get or what your textbook says.
• At times, the experiment may not work as expected.  In those cases, write what you actually got and also what you expected to get.  This clearly shows discrepancies, if any and helps you understand better.
• Yes/No answers
• Never write 'yes' or 'no' or one-word answers in reports.
• For one who reads the report, there is no way to know if you guessed (flipped a coin ?) or had a valid reason to write that answer.
• Instead show that you thought about the question and so have some reasoning to write that answer.
• Incomplete answers
• Never miss to answer a question in the report.
• Each question carries some point and so missing a question reduces your total attempted points.
• Predictions
• Your predictions need not be same as what you expect to get (after all it is only a prediction). So do not think that predictions have to be correct.
• Understand the difference - a prediction is not a guess.  Therefore, you must have a valid reason for your prediction.

Thanks to Reuben Johnson & Prof. Mizutani for their valuable comments.

Why should you write lab reports at all ?

• Lab reports are never meant only for grading purposes. They are meant for you to understand the experiment and help you to document your observations.
• Reports are the main (or only ?) way, information is passed between individuals in real world - between departments in a company, to your supervisors.  So your reports speak for you (offline or in your absence).
• Employers do not like when you write sloppy reports.  If you have to explain what you meant in a report, then that report is useless. Again your report has to speak everything you wanted to say or observed.
• Everyone learns by practise.  Lab reports help you to writing better reports - later to keep your job, get pay rises and promotions too :-).
• Note that advances in computer technology will not stop report writing - at least in the near future.  Instead you only learn new software or new technology, that help you write better reports.  So you might as well use the free practise opportunity you get in laboratory courses.

Pay attention to writing better lab reports from now on.  Put in your best.

Feb. 12, 2003 / anbumani@vt.edu