how do i do a science experiment on Which thermometer is most accurate?

If you assume that all the thermometers together have a normal distribution of error, then take a bunch of readings at different temperatures. Average the readings at each temperature. The thermometer with the best overall performance (you can average the variance for each themometer at all temperatures) would be the most accurate.

The only problem is the assumption of normal distribution. You could make a histogram (bell curve chart thingy) to check for normality but you really ought to be using at least 30 or more thermometers to assume normal distribution.

Go with what you've got and then in the discussion section you can discuss the difficulties of project funding. Did the school give your sister a grant for this project? If not, they should accept your results and be happy.

Since she is doing a science project (for school, I imagine) then there will be a chem or physics lab. In one of these there should be a standard thermometer that is NBS certified. She should compare the 6 to that one for accuracy. A lot of the suggestions so far have been very reasonable and worthwhile, but can't reach an accuracy level unless there is a vessel that can be fitted with a gromet and another for a vacuum line and potentiometric thermistor. Place the thermometer through the gromet, pull a vacuum to about 250 mm (measured accurately with a regulator) and compare the thermistor with the thermometers.

You need to get some distilled water and freeze part of it and boil part of it. Before doing the boiling part, look up your altitude on line to get the actual boiling point.

http://www.csgnetwork.com/h2oboilcalc.html

If you want to be really super careful, allow for the barometric pressure.

Bring the water to a boil then back off the heat so the water continues to move from the heat, but there are no violent bubbles. Measure the temp with each thermometer.

Then take the frozen water and break it up into smaller pieces (chunks and chips) and use an insulated container like a thermal cup or ice chest and put distilled water and ice in the container and stir it gently. The temperature will be as close to 32F (0C) as you can reasonably make it, so measure with each therm.

Well, for something simple, the boiling point of pure water is a pretty good reference point, if you are at a location close to sea level, the temperature of boiling water will be very close to 100°C or 212°F If you at an altitude significantly different you'll have to do a little research to find the boiling point at your altitude. Also mixture of pure water ice and water will be at very close to 0°C or 32°F. Checking each thermometer to see how close they read to these values would be a fairly good indication of accuracy.

I think you can put ice in water and see if it registers near 33 degrees. 32 degrees turns water to ice. Any thermometer that is closet to that temperature is the most accurate.

theres traditional thermometers with mercury tube. then theres the digital kind. you may want to compare them. instance the digital kind show degress in tenths or hundreths place, while traditional kind only whole numbers.

that sound like an interesting project.

i think of it is a robust theory for a technology test. My suggestion is to maintain the onions in the freezer, in the refrigerator, and at room temperature for equivalent quantities of time. be sure you have a sparkling speculation and be sure you're making a prediction. Write down all your information, and evaluate it afterwards. evaluate no remember if or no longer your prediction develop into shown or no longer.