I am trying to make a trend line that crosses the origin (0,0)

Use the LINEST function:

LINES (known-y-values, known-x-values, nonzero-y-intercept, more-stats)

nonzero-y-intercept:

• An optional value specifying how the y intercept (constant b) should be calculated.
• normal (1, TRUE, or omitted): The value of the y intercept (constant b) should be calculated normally.
• force 0 value (0, FALSE): The value of the y intercept (constant b) should be forced to be 0.

Details in the iWork Formulas and Functions User Guide, which may be downloaded via the Help menu in Numbers '09.

You will need a column on your table to hold the calculated values if you want to show the line on your graph.

Regards,

Barry

Just my two cents... I would check with your Chemistry professor first..

Have you manually proven that your data points actually do create a line of best fit that should fall on or close to 0,0 when plotted? If not then you should not (my prof would say cannot) report that it would in your results.

If you force your experiments results to show a 0,0 intercept and your observations dont support it, my chemistry professor would fail you for that lab.

I was taught you should report what you found, and then explain why you dont think the experiment worked properly, where your defects might have been, and how to avoid such errors in the future. i.e. how can i learn from what just happened, not how can I make my results show what the teacher wants to see.

Jason

Jason's point is well taken.

Regards,

Barry

Joshuwa,

Adding to Jason's points, experimental errors are expected. It's good to examine the results for causes, but in the end, all observations are subject to variation. If you run the experiment numerous times and the observations deviate from expectations in only one direction, you have a bias that should be explainable if you understand all the influences on your measurement equipment and the system you are measuring.

Jerry

Barry & Jerry

Thanks, I was expecting to get reemed for giving the advice I gave instead of an answer.

It has always been a pet peeve of mine when students think they have to give the teacher the perfect paper/lab report. My proff's seemed to like it better that we learned to observe and account for the error.

I am even trying to get my 5 y/o son to learn to explain what he thinks will happen, try it and observe, now figure out why it did or did not do what he thought. Like things that float for example. (unbenownst to him, he's learning scientific method). He calls it "spare-i-ments" and loves doing it so much he has dreams of things he wants to do.

Thanks again guys,

Jason

Other things to think about with regards to experiments and measurements:

An experiment that shows a linear trend over a certain range may not show a linear trend over another. Probably not the case for this particular experiment.

Measurements may be precise but not accurate. Think about measuring something with a ruler that is precise to 1/100ths of an inch but it is a little worn on the end. All measurements will be off by a little bit (inaccurate) but they will still be precise. Plot your linear trendline and it won't go through zero. Forcing it to cross at zero will create the wrong trendline; the correct trendline would have the slope of the non-forced trendline but an intercept of zero. For this density experiment, if you were measuring density by weighing objects then placing them in a liquid, perhaps the miniscus was affecting the accuracy of your measurements or the scale was not well calibrated.

Edit: this is in reply to the OP, not Jerry. I need to be more careful as to whose post I am reading when I hit reply.

The question is, what is the proper way to force the trendline to go through zero? The easiest way is with LINEST but it might not be the correct way. The next easiest way would be a visual look at your data. If the forced trendline does not fit your data (as seen in the chart below), you probably do not want to use it. You might be better off using the correct slope and a constant offset to the intercept so the line goes through zero (also shown in the chart below and explained previously in the example of a worn ruler). In this simple example, if you use the forced trendline to predict the result at 50, your result will be far from correct. If you use the non-forced trendline to predict the result at 50, it will be off by a small amount. If you use the constant offset trendline to predict the result at 50, it will be on the mark.

Whatever you choose to do, explain it well in the text.