How to address experimental error in conclusion

1. Dec 12, 2015

24forChromium

I am writing an essay which includes the experimental "verification" of a theory. As it turns out, the theory was able to predict a certain variable F's direction of change (i.e. increase or decrease across two data points) "correctly" for 90% of the time, and predicts the value of F at each point with an average error of 0.5Newtons.

How should I say the theory is "proven" and to what extent in the experiment's summary?

2. Dec 12, 2015

Student100

Proven? What about the 10% of the time or the $\pm$ .5N? That's what you want to talk about, not that they theory was "proven." What were sources of error, ways to design a better experiment? Limits on results?

How did you analyze your data?

3. Dec 12, 2015

24forChromium

The method I used is kind of inappropriate, I monitored two independent variables and used the theory to get the dependent, the dependent is also measured experimentally. After that, the predicted values of the dependent F is compared to the empirically observed F.

4. Dec 13, 2015

micromass

Staff Emeritus
Then you might want to do a statistical analysis on your data set.

5. Dec 13, 2015

24forChromium

I am afraid that this reply is not specific enough. What kind of results should the analysis yield and how are they related to the verification/falsification of the theory?

6. Dec 13, 2015

micromass

Staff Emeritus
I'm afraid that's for you to figure out. But generally, if you have a theory that you want to verify experimentally, then using a statistical analysis is crucial. Just saying "it predicts it correctly 90% of the time" is not saying much. You need to study the errors and whether the errors are significant enough to reject the theory. I suggest you study confidence intervals and hypothesis testing. Maybe even regression analysis.