Since the distribution of ratings of each item are both negatively skewed (mean skewness across all items = -1.52) and kurtotic (mean kurtosis across all items = 3.96). I first reversed the ratings for each item and then took the square root of each to reduce skewness and kurtosis. I then reversed the transformed ratings back to the original direction so that higher ratings would me that the item was rated as more important in their view of love and that lower ratings would me that the item was rated as less important in their view of love. This resulted in the distance between response options changing from 1 point between response options for the original scale to a gradually increasing interval between response options (see table below).
 

Anchor	Original Metric	Distance	Transformed Metric	Distance	Average Distribution of Responses
Not Important	0	NA	0	NA	2%
	1	1	0.21	0.21	2%
	2	1	0.45	0.24	3%
	3	1	0.72	0.27	7%
	4	1	1.04	0.32	15%
	5	1	1.45	0.41	28%
Very Important!	6	1	2.45	1.00	43%

These transformations resulted in a reduction in kurtosis across all items from 3.96 to -.57 and a reduction in skewness across all items from -1.52 to -.24.

Summary Measures	Mean	SD	Kurtosis	Skewness	N
Before Transformation	4.86	1.13	3.96	-1.52	8,813
After Transformation	1.68	0.63	-0.57	-0.24	8,813

Return to the Concept of Love Dimensions Page.

Return to the Love Test Findings Page.