The lab write-up will usually include:

• name, partners's name, date
• number and title of experiment
• abstract or purpose (this is NOT procedure)
• The purpose of the lab is not "to better learn". Or the purpose of everything you do here is to learn, we all know, but WHAT IS THE PURPOSE OF THE LABORATORY EXPERIMENT??? What are you trying to measure? Do not put procedure in the purpose.  The purpose should stand alone: it should be understandable to someone who has not read the experiment.  That does not require procedure in the purpose, it just requires being specific about what you are doing.  For example, saying "we are determining the acceleration of the system" is not enough.  What system?
• apparatus schematics and method ONLY if they differ from the handout
• data in tabulated form
• You do not NEED correct significant digits in your EXCEL sheet as long as you have them correctly in the analysis/conclusion sections.
• plots (if any)
• Data and plots should be input/done/calculated in the Excel spreadsheet.  For informal reports, print them out and glue or tape them into your notebook.  For formal labs you should just attach the appropriate files to your document.  Tables and plots should be properly labeled (axes and titles with appropriate units).
• results/error analysis (use correct significant digits)
• Your analysis section should include a discussion of error. What are the possible sources of error? Are they random or systematic? Do you really think they are significant? If so, HOW DO YOU THINK THAT ERROR WOULD AFFECT YOUR DATA - that is, would it explain any discrepancies you saw? e.g., if you said that air friction was an error, you would conclude that air friction would slow things down, make the actual distance of your projectile less than predicted. Was this observed in the data?  Systematic means that every data point should have been affected in the same way (e.g. "always measured too small a time"), while random means that each data point may not be precise, but it might be higher than it should be for one measurement, and lower than it should be for another.  For example, if you are measuring the acceleration of a body on a level surface, if the surface is sloped forward, then it will systematically affect all of your data and make all of your accelerations too high.  On the other hand, if you are placing a measuring instrument on the surface, and moving it each time, one time it might be too far to one side of where it should be, and another time it might be too far to the other side, so it would be a random error.
• In coming up with sources of errors, it is essential to think about the assumptions made in any theoretical calculations.  For example, if vectors were used, and it was therefore necessary for one force to be perfectly horizontal, was it really horizontal?  Or if certain things were ignored, do you think that they should have been?
• If you did a straight-line fit, in your analysis you must always discuss both the slope and y-intercept to 90% confidence.  What is the physical meaning of each?  In addition, if you are giving a result that has error associated with it, do not give the result alone, give the 90% confidence interval only.  For example, do not say "we measured a = 0.1234 m/s^2" and then later give the error.  Simply give the 90% confidence interval right off.  Be sure to understand the physical meaning of the slope and y-intercept by comparing them to theoretical relationships.  Do NOT assume that the slope of something you have arbitrarily plotted is equal to the value you are interested in.  For example, if you plot v^2 vs x for a body moving with constant acceleration, the slope is TWICE the acceleration, not equal to the acceleration.
• A 90% confidence interval is your answer+/-2*standard_error.  You must round the SE to one significant digit, and round your answer to the same decimal place.  For example, 0.01234+/-0.00923 should be rounded to 0.012+/-0.009.
• Be sure to give units for ALL of your answers, slopes, y-intercepts, any other measurements you mention in your lab.
• When you give results or calculations, all of the data had better be somewhere to back it up.
• If your data is just plain wrong - i.e. makes not sense rather than just giving lots of error, you should make an effort to discuss that.  For example if your velocity vs. time graph shows a decrease in velocity as time increased, discuss whether you actually saw the cart slow as it moved down the track.
• All final results should be compared using the null hypothesis (if you have different measures of the same quantity).  If you have a theoretical result, you must compare that to the measured result using the null hypothesis.  Remember that the null hypothesis asks the question, is the difference between two results equal to zero?  If they are, then the two results are statistically equal.  An example is as follows:
  if result1 = 0.023+/-0.002 and result2 = 0.026+/-0.002 then
  is result1-result2 = 0? 
  result1-result2 = [0.023+/0.002] - [0.026+/-0.002]
  = [0.023-0.026]+/-[SQRT(0.002²+0.002²)]
  = 0.003+/-0.003. 
  Since zero is in this interval, then the null hypothesis is successful, and the two values are equal.
• conclusions (address the purpose)
• Your conclusion must address the purpose. Did you fulfil the purpose? You must restate your pertinent results in the conclusion. e.g., even if you had it in another section, state your predicted range vs. your actual range, and briefly state how they compared. Be sure to use ninety percent confidence intervals and be sure to have correct significant digits.
• Please be sure to answer any additional questions posed in the online lab manual (as questions or postlabs) in either the conclusions or an additional section at the end: whichever is most fitting.

PROOFREAD YOUR DOCUMENT. This is meant to be a well-written, concise document. You should consider it no different than a paper you have written for a humanities class.