I have three parameters, 1. Attachment level with phone (categorical data eg. Highly attached, detached, highly detached etc) 2. Hours spent using phone before covid (continuous) 3. Hours spent using phone after/during covid (continuous). Now what or how to correlate these three parameters to draw a meaningful conclusion?
Correlation - need help to understand which method nto use for before after scenario
- Thread starter mghost
- Start date
- Tags before and after correlation analysis relationship
So you would ask whether the association between attachment and hours of usage
was stronger (or weaker) after Corona than it was before Corona?
Do you have a theory which explains why the strength of the association should change?
By the way, perhaps one could imagine that phone usage generally increased from before
to after, and how much it increased was partly influenced by level of attachment. But
seemingy this is not of interest here?
With kind regards
Karabiner
was stronger (or weaker) after Corona than it was before Corona?
Do you have a theory which explains why the strength of the association should change?
By the way, perhaps one could imagine that phone usage generally increased from before
to after, and how much it increased was partly influenced by level of attachment. But
seemingy this is not of interest here?
With kind regards
Karabiner
There is no theory as such. Its based on a general understanding.
"By the way, perhaps one could imagine that phone usage generally increased from before
to after, and how much it increased was partly influenced by level of attachment. But
seemingy this is not of interest here?"
Yes this is of interest here, is there a way to find?
Basically trying to find out if there was a change in hours of use of phone, and if this change has something to do with attachment levels.
"By the way, perhaps one could imagine that phone usage generally increased from before
to after, and how much it increased was partly influenced by level of attachment. But
seemingy this is not of interest here?"
Yes this is of interest here, is there a way to find?
Basically trying to find out if there was a change in hours of use of phone, and if this change has something to do with attachment levels.
So at last we know what you actually want to find out.
The simplest way would be a calculation of the individual differences hours after - hours before.
You can correlate this differences with attachment, using Spearman's rank correlation. This
will tell you whether there is an association between attachment and behavior change.
For interpretation, you should additionally check whether hours of phone use at baseline was
closely associated with attachment (Spearman correlation). Possibly, those with high attachment
were already using their phone extremely often, so that this could not increase much more
(ceiling effect).
With kind regards
Karabiner
The simplest way would be a calculation of the individual differences hours after - hours before.
You can correlate this differences with attachment, using Spearman's rank correlation. This
will tell you whether there is an association between attachment and behavior change.
For interpretation, you should additionally check whether hours of phone use at baseline was
closely associated with attachment (Spearman correlation). Possibly, those with high attachment
were already using their phone extremely often, so that this could not increase much more
(ceiling effect).
With kind regards
Karabiner
Thank you so much for your suggestion. Yes I believe Spearman's rank correlation, would be perfect for my case. I would certainly use this. Is there any way I can cite or acknowledge you in my paper?
I appreciate your hint towards ceiling effect and your presumption is correct. Should I reject data points with high attachment and high 'before hours'? In other words how to tackle with the ceiling effect? Any suggestion?
I appreciate your hint towards ceiling effect and your presumption is correct. Should I reject data points with high attachment and high 'before hours'? In other words how to tackle with the ceiling effect? Any suggestion?
I would not exclude data. And whether there's a ceiling effect in your sample, is currently only a hypothesis.
You could display hours before and hours after for each of the attachment level separately, for example
using box-and-whisker-plots. This could give you further insight
With kind regards
Karabiner
You could display hours before and hours after for each of the attachment level separately, for example
using box-and-whisker-plots. This could give you further insight
With kind regards
Karabiner