# Spearman Rank + Quadratic Regression (spps) HELP! x

Watch
Announcements
Thread starter 1 year ago
#1
Hi,
I'm trying to find out if there is a significant correlation between salinity and no. of Bivalves found on a shore. I did a Spearman Rank and this was my result:
"There was a statistically significant correlation between Salinity and the number of Bivalvia individuals (Spearman Rank Correlation Coefficient, rs = 0.63, n = 40, P < 0.05)."
But then I did quadratic regression and it's saying it's not significant, with a P value of 0.267 !!
How can this be when my rs value is pretty close to +1?
Am I using the wrong tests?? If so, what do I use??
Many thanks.
0
reply
1 year ago
#2
(Original post by Aknwsfg)
Hi,
I'm trying to find out if there is a significant correlation between salinity and no. of Bivalves found on a shore. I did a Spearman Rank and this was my result:
"There was a statistically significant correlation between Salinity and the number of Bivalvia individuals (Spearman Rank Correlation Coefficient, rs = 0.63, n = 40, P < 0.05)."
But then I did quadratic regression and it's saying it's not significant, with a P value of 0.267 !!
How can this be when my rs value is pretty close to +1?
Am I using the wrong tests?? If so, what do I use??
Many thanks.
Whenever you get a discrepancy like this, it's likely to mean that the assumptions that underlie at least one of the tests do not hold. The starting point for any bivariate analysis like this is to plot the data. What do you see when you do this?
1
reply
Thread starter 1 year ago
#3
(Original post by Gregorius)
Whenever you get a discrepancy like this, it's likely to mean that the assumptions that underlie at least one of the tests do not hold. The starting point for any bivariate analysis like this is to plot the data. What do you see when you do this?
The relationship looked quite weak...
Although, I also performed other tests that said Salinity was significantly high at a certain site and significantly low at another, then another test said that Bivalvia was significantly high and low at the same sites so I was trying to conclude with a strong correlation... I guess it's all fell apart now
0
reply
1 year ago
#4
(Original post by Aknwsfg)
The relationship looked quite weak...
Although, I also performed other tests that said Salinity was significantly high at a certain site and significantly low at another, then another test said that Bivalvia was significantly high and low at the same sites so I was trying to conclude with a strong correlation... I guess it's all fell apart now
If you post a plot of your data, I can make some more suggestions.
0
reply
Thread starter 1 year ago
#5
(Original post by Gregorius)
If you post a plot of your data, I can make some more suggestions.
Ok, I have attached it!
0
reply
1 year ago
#6
(Original post by Aknwsfg)
Ok, I have attached it!
So looking at that plot, I would be very dubious about a quadratic regression. (a) you don’t have all that much data to estimate the three coefficients (constant, linear and quadratic terms) and (b) I can see by eye, That the residuals (the differences between the actual values and the values predicted by the quadratic model) are far from evenly divided above and below the fit, let alone normally distributed (which is the assumption of the method). I would trust the non parametric test, which shows a weak, but significant, correlation.
0
reply
X

Write a reply...
Reply
new posts
Back
to top
Latest
My Feed

### Oops, nobody has postedin the last few hours.

Why not re-start the conversation?

see more

### See more of what you like onThe Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

### Poll

Join the discussion

Yes (56)
29.02%
No (137)
70.98%

View All
Latest
My Feed

### Oops, nobody has postedin the last few hours.

Why not re-start the conversation?

### See more of what you like onThe Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.