Oceanography Student In Need of Statistics Help

[Oceanography]

I need to analyze my data statistically and am not sure how to go about it.

My hypothesis is that population A has increased significantly compared to population B over the past 20 years due to the increase in water temperature attributed to climate change.

What I am showing:

Twenty years ago population B was greater than population A.

Today population A has begun to greatly outnumber B in population.

How do I analyze this statistically?

I really have no clue where to start as I do not take Statistics, but will be working with professors to find an appropriate method. Though a head start in the right direction would be appreciated. I've been researching different methods, but none seem to be perfectly appropriate.

[Dr N S Gandhi Prasad]

Hello
Try the following
1. Do you have Time Series Data of the following Variables for a Period of 20 Years ?
(I) Population (A) (II) Population (B) (III) Water Temperture
2. If you feel changes in the sizes of population are due to natural causes, Examine their trends by fitting appropriate trend curves also supplement with a suitable statistical Test for testing the equality of population sizes after adjusting the data for the trends.

3. If you feel water Temparature is a contributing factor, run appropriate Linear or Non Linear Regression of each population on Temperature. If both the regressions are significant, test the equality of population sizes after adjusting the data for Water Temperature. In the event of non significant relations, carry out ususal statistical tests of Significance ( such as t - test for testing the equality of means

I hope this may serve as a good starting point

Good Luck
Dr N S Gandhi Prasad
Professor of Statistics ( Retd. )
Hyderabad India

[jerryb]

I'll add one thing,

HOW do you know that the population today is bigger than the population 20 years ago? if you have one data point from 20 years ago (say some population study) and a more recent one then i would suggest that you back up one step and verify that there IS a difference in the two populations, and not just a difference in your two measures due to the normal distribution of samples.

cheers
jerry