# Thread: Identifying Effect size in Power Analysis

1. ## Identifying Effect size in Power Analysis

I have a set of data for a water quality parameter on which is used to compute Rank Spearman Trend. I am interested to know whether the number of sample is sufficient from the power for the sample size of this parameter. However I am a bit confused as after doing some study I found that I need two group where the difference in mean of the two groups are subtracted and divided by the standard deviation to find the effect size. But in this case there is only one group i.e. the data which is used for the computation of trend. Therefore my question is given the sample size and the data from which I can compute the mean and standard deviation, how do I find the effect size and compute the power to identify whether the number of sample is sufficient for the trend analysis. Can anyone help?

2. ## Re: Identifying Effect size in Power Analysis

To have a trend you need 2 variables, so how don't you have two variables?

Water variable against time?

3. ## Re: Identifying Effect size in Power Analysis

Hi,
for a Spearman correlation the effect size is the smallest value of the correlation coefficient that you want to detect with a given probability, so you just need to decide on that, you do not need two groups etc.

regards

4. ## Re: Identifying Effect size in Power Analysis

So you don't need two groups, but you need two variables, this is where I was getting confused by the OP?

5. ## Re: Identifying Effect size in Power Analysis

Originally Posted by hlsmith
To have a trend you need 2 variables, so how don't you have two variables?

Water variable against time?
Yes in that sense there are two groups. However comparing the average of time vs the average of water quality variable does not make much sense...

6. ## Re: Identifying Effect size in Power Analysis

Originally Posted by rogojel
Hi,
for a Spearman correlation the effect size is the smallest value of the correlation coefficient that you want to detect with a given probability, so you just need to decide on that, you do not need two groups etc.

regards
Re: Identifying Effect size in Power Analysis

Can you please elaborate a little further with an example. Lets say you have five set of data for Aluminum collected from 5 water samples:
Date Aluminum (mg/L)
Jan 1, 2016 0.0013
Feb 1, 2016 0.0015
Mar 1, 2016 0.0012
April 1, 2016 0.05
May 1, 2016 0.003

So is this about the smallest detectable size which I am deciding on aluminum perhaps based on historic maximum and minimum?

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts