Guest end3 Posted June 17, 2009 Share Posted June 17, 2009 There are smart individuals here. I have a question please. Help would be most appreciated. I am about to start a test at our local university. I am trying to compare three groups of seed-starting pots, three groups made of natural materials, and the forth being the control, made of plastic. My question is: How many individuals would be needed per group to make a "significant number" in that group? Or is this even a good question? I don't know why, but 20-something rings a bell in my head from college many years ago.... Thanks, End Link to comment Share on other sites More sharing options...
Guest end3 Posted June 17, 2009 Share Posted June 17, 2009 My question is: How many individuals would be needed per group to make a "significant number" in that group? Or is this even a good question? Hi, End! I think you're asking what the minimum sample of seeds would have to be in each set to yield meaningful data. Is that right? Phanta Yes, probably one in the same Link to comment Share on other sites More sharing options...
Outback Jack Posted June 18, 2009 Share Posted June 18, 2009 End3, you might want to also try your question at the JREF forum. (The link is to their Science, Math, tech subforum.) They have a large number of members and many of them are scientists, mathematicians, doctors, ect. You might be able to get some info from people who work with stuff like this on a daily basis. Link to comment Share on other sites More sharing options...
Guest Davka Posted June 18, 2009 Share Posted June 18, 2009 Statistical significance is a function of sample size. In other words, it is not a question of whether your results will be statistically significant, but rather how significant they will be. The larger the sample size, the greater the statistical significance. Statistical significance really means "how likely is it that these same results could be obtained through pure chance?" If you work with a small number of seeds, you can calculate the chances of any given outcome as fairly high. When you have a larger number of seeds, the chances decrease that any given outcome would be the result of pure chance. A sample size of 20 seeds for each of the 4 containers might be sufficient for the kind of experiment you are describing. Any strong variation across the majority of the seedlings in a single pot could be safely attributed to the material the pot is made of. If you are expecting very slight differences, however, increase your sample size. 100 plants per container type, for example. Link to comment Share on other sites More sharing options...
Recommended Posts