# Which statistical treatment to use. **URL:** \ **Category:** Factual Questions\ **Created:** [August 25, 2013, 9:14am UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117 "2013-08-25T09:14:58Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![chowching](https://avatars.discourse-cdn.com/v4/letter/c/f14d63/32.png) [@chowching](https://boards.straightdope.com/u/chowching)\ **Post date:** [August 25, 2013, 9:14am UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117/1 "2013-08-25T09:14:58Z")
My niece is asking me help in doing her science investigative project. Well we are stuck in which statistical treatment to use. Here is the gist of her project. It’s like this. They plant a fortune plant in a soil with acid phosphate and egg shells and another one in a soil with acid phosphate and mussel shells. They observed for 7 days and recorded the height of the two plants. They want to compare the two. I just need some advice which statistical treatment should be used. Should I use ANOVA or t-test? I haven’t done these kind of things since high school but I think l can study it by myself if I’m pointed in the right direction. Thanks guys. ---
**Author:** ![Quartz](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/quartz/32/267_2.png) [@Quartz](https://boards.straightdope.com/u/Quartz)\ **Post date:** [August 25, 2013, 10:42am UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117/2 "2013-08-25T10:42:25Z")
Given that there are only two plants, isn’t statistical treatment irrelevant? ---
**Author:** ![Crafter\_Man](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/crafter_man/32/458_2.png) [@Crafter\_Man](https://boards.straightdope.com/u/Crafter_Man)\ **Post date:** [August 25, 2013, 12:32pm UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117/3 "2013-08-25T12:32:20Z")
Just two numbers – the height of each plant after seven days? You can’t do statistical analysis on two numbers. Now, if you measured the height of each plant every seven days over the course of (for example) 60 days, you can do something with the data. ---
**Author:** ![chowching](https://avatars.discourse-cdn.com/v4/letter/c/f14d63/32.png) [@chowching](https://boards.straightdope.com/u/chowching)\ **Post date:** [August 25, 2013, 1:31pm UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117/4 "2013-08-25T13:31:06Z")
They did record the height of the plants every day for 7 days. I told her to just describe the results they gathered since there’s no point using a statistical test. Thanks guys. ---
**Author:** ![Crafter\_Man](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/crafter_man/32/458_2.png) [@Crafter\_Man](https://boards.straightdope.com/u/Crafter_Man)\ **Post date:** [August 25, 2013, 2:50pm UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117/5 "2013-08-25T14:50:35Z")
You could plot both data on the same graph: time on the independent (x) axis, height on the dependent (y) axis. Is the growth fairly linear over time? If so, you could do a 1st-order curve fit on each (a.k.a. linear regression, least squares fit), and compare slopes. ---
**Author:** ![MsBatt](https://avatars.discourse-cdn.com/v4/letter/m/c68b51/32.png) [@MsBatt](https://boards.straightdope.com/u/MsBatt)\ **Post date:** [August 25, 2013, 4:48pm UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117/6 "2013-08-25T16:48:05Z")
I want to know where to get a fortune plant. ---
**Author:** ![bleach](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/bleach/32/3106_2.png) [@bleach](https://boards.straightdope.com/u/bleach)\ **Post date:** [August 26, 2013, 6:26pm UTC](https://boards.straightdope.com/t/which-statistical-treatment-to-use/667117/7 "2013-08-26T18:26:53Z")
> [@Crafter\_Man](#): > > You could plot both data on the same graph: time on the independent (x) axis, height on the dependent (y) axis. > > Is the growth fairly linear over time? If so, you could do a 1st-order curve fit on each (a.k.a. linear regression, least squares fit), and compare slopes. Following up on this idea, the statistical test to compare coefficients of two regression lines is called the [Chow test](http://en.wikipedia.org/wiki/Chow_test) (ironic, given the OP’s username).