This is reply in part to a 16-year-old post (https://www.annabelleigh.net/messages/187669.htm) on a 24-year-old study, but still...

Puzzled asserted that for any combination, if you added sexual abuse it had a seemingly beneficial effect. At the simple, straightforward level this is true for most of the combinations (though see interpretation below). But I am mightily puzzled because the combinations are listed in tables 3 and 4 -- but the headings for table 4 are exactly the same as for table 3. My best guess is that this is a printing error. I am looking at what looks like a scanned copy of the paper journal. I don't know how Puzzled would reach the exact conclusions he did without having some solution to this problem. Was he looking at some corrected version of the paper?

Hints welcome.

Evaluation (given the limitation):

A Pearson Coefficient is between two somewhat-continuous variables. What is the variable when you combine two things like "physical neglect and verbal abuse"?

Puzzled follows the authors in attaching great importance to the ranking of the 20 combinations, but I saw no tests of statistical significance for that rank ordering (though there are tests for individual coefficients with outcomes).

There are some other strange things in the data. According to this, if you have physical neglect and verbal abuse, then if you add emotional neglect then things get better. If you have verbal abuse and emotional neglect, then if you add in physical neglect then things also get better. Just looking at the correlation coefficients they seem to be all over the map instead of suggesting any emerging regularities. For instance, compare physical neglect and verbal abuse with those two and sexual abuse added in. Here are the correlations on the individual measures:

.39 .43 Lack of enjoyment in living
.21 .18 Lack of purpose in life
.08 .26 Poor future for the world
.33 .08 Own future poor
.05 -.28 Poor chance of good job
.30 .14 Poor chance of happy marriage
.24 .17 Poor chance of having kids
.17 .29 Poor chance of being good parent

The -.28 means it's MORE likely you'll get a good job? This just makes no sense.

There is strange stuff going on here, and my tentative conclusion is that it's just not a reliable study that we should pay much attention to.