0

I'm doing a stationary sport in which I use different colored equipment for different weather. I am recording my scores along with the color of the equipment and the weather in which I am using it.

I would like to plot the score averages for each type of color on a bar graph and add the weather conditions as filters. This way, I want to compare the equipment in different contexts and thus see which one is the best for, say, cloudy weather.

The problem: I don't have equal data sets for each type of color in each type of weather. What can I do to have an accurate comparison?

Andrei Erdoss
  • 141
  • 3

1 Answers1

-1

You can compare averages of unequal sample sizes. There's no problem with that.

Germaniawerks
  • 1,027
  • 1
  • 10
  • 15
  • 1
    "no problem" is ambiguous. You *can*, but you lose power; for more, see my answer here:[How should one interpret the comparison of means from different sample sizes?](http://stats.stackexchange.com/questions/31326//31330#31330) – gung - Reinstate Monica Dec 28 '13 at 02:05
  • 5
    @gung "lose power" would assume that some particular alternative was available, as if one could choose to reallocate the available sample sizes in some other proportion. If that's the case, then yes, uneven allocation will have lower power than an even allocation. – Glen_b Dec 28 '13 at 03:10
  • @Germaniawerks the question was *what* could be done rather then *if* it is possible. Even if the question was about if it is possible then your answer still would be lacking of explanation and details. (-1) – Tim Feb 24 '15 at 14:52