fix splithalf error when two variables #281
Conversation
There was a problem hiding this comment.
Pull Request Overview
This PR fixes a split-half reliability calculation error that occurs when there are only two variables in the analysis. The primary issue was that rowSums() would drop dimensions when applied to single-column matrices, causing downstream errors.
- Adds
drop = FALSEparameter torowSums()calls to preserve matrix dimensions - Fixes conditional logic structure in the item-deleted split-half analysis section
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
|
|
||
|
|
There was a problem hiding this comment.
[nitpick] The added blank line at line 5 appears unnecessary and doesn't improve code readability.
|
|
||
|
|
||
| } | ||
|
|
There was a problem hiding this comment.
[nitpick] The added blank lines around the closing brace appear unnecessary and don't improve code readability.
| } | |
| } |
juliuspfadt
force-pushed
the
fixSplithalf
branch
from
41fcf45 to
a0d739b
Compare
juliuspfadt
force-pushed
the
fixSplithalf
branch
from
a0d739b to
9007c66
Compare
juliuspfadt
force-pushed
the
fixSplithalf
branch
from
9007c66 to
1f228d9
Compare
vandenman
left a comment
vandenman
left a comment
There was a problem hiding this comment.
I have no preference about the renaming, but perhaps we should check what other software does (SPSS?) and then just call it the same.
|
If I am not mistaken, they call it "Alpha" and the analysis "Reliability Analysis". We had some issues with people not finding Cronbach's alpha... so I thought we bite the bullet |
|
any idea why the tests fail on macos? Do they need a higher tolerance? |
|
same thing happened in my SEM tests with bootstrapping... |
juliuspfadt
force-pushed
the
fixSplithalf
branch
from
1f228d9 to
ee12f17
Compare
juliuspfadt
force-pushed
the
fixSplithalf
branch
from
c31520f to
7a94c9d
Compare
2 participants
fixes jasp-stats/jasp-issues#3727
renames "Coefficient" alpha and omega to "Cronbach's" and "McDonald's" as per @EJWagenmakers request. Initially, I had named them "Coefficient" because it is the correct way. Since people cannot find the coefficients though and I dont feel as strongly anymore, this is what it is.
Also changing to "Reliability" from "Unidimensional Reliability" makes this easier to find. The helpfile still specifies that it should be used for unidimensional data.
By now I have added some changes for the standard error of measurement analysis. Sorry about that. I need them for a tutorial :). If you wish to test there is an example in the data library.