Violence By Women, Finnish Style

A story about violence by women on the website of the Finnish state television company begins like this:

Vaikka naisten osuus nujakoinnista on kasvussa, valtaosa Suomessa tehdyistä väkivaltarikoksista on edelleenkin miesten tekemiä. Poliisin tietoon tuli vuonna 2012 kaikkiaan 29 000 miesten tekemää pahoinpitelyä. Samaan aikaan naisten tekemiä vastaavia rötöksiä kirjattiin runsaat 6000.
Miehet siis johtavat tilastoa, mutta kasvu on suurta nimen omaan naisten osalta. Nousua on jopa kolmannes vuoden 2005 tilastoihin nähden, analysoi apulaispoliisipäällikkö Erkki Kerola Keski-Pohjanmaan ja Pietarsaaren poliisilaitokselta.
- Naisten osalta nousu on kova. Miestenkin osalta lisäystä on, mutta ei ollenkaan niin rajua.

My translation (probably slightly off as I'm not sure about "nujakointi"):

Although women's share in assault and battery statistics is growing, the majority of Finnish violent crimes are still carried out by men.  In 2012 the police was informed about a total of 29 000 cases of assault and battery by men.  At the same time, equivalent crimes by women were recorded as somewhat over 6000.

Thus, men lead in the statistics, but growth has been especially large in the case of women.  Compared to statistics from 2005, the latter growth equals one third, states deputy police chief Erkki Kerola of the police in Keski-Pohjanmaa and Pietarsaari.
- The rise among women has been steep.  Men's statistics have increased, too, but nowhere near as rapidly.

Now for the actual statistics:  In 2005, men carried out 25 500 cases of assault and battery, women 4009.  In 2012, the comparable number for men was 29 000 and for women 6088.

Compare those statistics to the earlier translated text.  What do you spot?  The absolute increase in the numbers is actually larger for men than it is for women (an increase of 3500 and an increase of 2079).  That the latter is a much larger percentage is because the base for the percentage calculations is so much smaller.

To see why this matters, note that if some particular group had had one case of assault and battery in 2005 and two cases in 2012, the increase would have been 100%.

The lesson?  The base matters when calculating percentages, and sometimes comparing the percentages directly isn't terribly informative. 

The other thing I find troublesome in these data are the rounded numbers for men.  Now it could be the case that the total numbers of assault and battery by men just happened to be 25 500 and 29 000.  But it's more likely that those figures were rounded off.  Why was the same not done for women's numbers?

The title of this piece, by the way, translates to:  "The Caring Myth Crumbles:  Woman brawls and beats."  And the end of the piece tells us (in my translation):

Cat fights between inebriated women growing

A new phenomenon in recent years is also the brawling of young women in public spaces.  Drunken arrests of women are increasing.

Police reports and leaflets describe these events:  Women fought in a park.  A drunken woman tore at another woman in front of a restaurant.  The police arrested cat fight participants to sober up.

Hmm.  I wouldn't call that article exactly neutral, and I doubt very much that a reversal article on this topic would ever be written.  That's because men are the unmarked gender, so drunken men are individual drunks, whereas drunken women are a sign about all womankind turning more brutal and less caring.