Who's likely to get a longer prison sentence for killing their spouse? Men or women?

I came across an article entitled John Welden sentenced to 13 years, eight months for slipping his girlfriend an abortion pill and was a bit curious how that compared to sentences for killing a spouse. When doing so I found two almost opposite claims.

Claim 1 (which seems to appear more frequently) as stated here:

The average prison sentence for men who kill their intimate partners is 2 to 6 years. Women who kill their partners are sentenced, on average, to 15 years.

Claim 2 (originally found here):

Wives who kill their husbands were acquitted in 12.9 percent of the cases studied, while husbands who kill their wives were acquitted in only 1.4 percent of the cases. Women who were convicted of killing their husbands were sentenced to an average of six years in prison, while men received an average of seventeen years for killing their wives

Claim 1 generally references some unnamed 1989 work from the National Coalition Against Domestic Violence which I've been unable to track down. A number reference a no longer existing page at cybergrrl.com but the Internet Archive's copy again fails to actually cite a specific document showing source data to support the claim. Claim 2 fairly closely aligns with a publication by the US DOJ's Bureau of Justice Statistics which specifies a data set of the 75 largest US counties in 1988 - i.e. it appears to trace down to actual court data.

It seems to me that the second claim - that men receive harsher sentences - is the one supported by the data though the first claim seems to be made more frequently. There are certainly instances of discrimination against women which are right to point out - one of the more interesting examples I think is the use of blind auditions seeming to change orchestra hiring results - but all too often the underlying scholarship is dubious at best. As I noted before, for example, the research that people seemed to find most intriguing in Sheryl Sandberg's book Lean In either seems to have failed the reproducibility test or come into conflict with other research findings.