February 7, 2018 by Robert Franklin, Esq, Member, National Board of Directors, National Parents Organization
The Warshak Consensus Report observed that none of the four significant outcomes reported by McIntosh et al. were derived from measures that met basic scientific standards, a point also noted by Linda Nielsen in greater detail.
You’d think that would be about as bad as a supposedly scientific paper could be, but I’m not so sure. The effort turned in by Samantha Tornello give’s McIntosh’s a run for its money.
Similarly, Tornello et al. used an instrument with no established reliability or validity to assess the child’s attachment to the mother. The instrument was abbreviated and modified from an established instrument, but there is no evidence of the validity of the modified version instrument.
That’s on a par with McIntosh’s approach, but Tornello’s gets worse, much worse.
Tornello, et al … reported that children who at age one had frequent overnights (1 to 5 overnights per week) were more likely than those with some overnights to be insecurely attached to their mothers at age three.
Bad news for advocates of shared parenting? Nope, bad news for Tornello, et al.
[T]he Warshak Consensus Report and other scholars have questioned the meaning of the attachment findings because the instrument was completed by mothers rather than by trained professional raters.
When parents provide the relevant information, it’s often skewed toward their view of how things ought to be instead of how things are. It’s an understandable human tendency which is why researchers need to avoid that method of gathering data. Tornello, et al failed to do so.
It gets still worse.
Tornello, et al issued a press release when their study was published. Many researchers do the same, so no problem, right? Not exactly. Tornello’s team took the remarkable approach of stating in its press release facts that were contradicted by its own data. Stated another way, they lied.
The press release issued by the lead investigators’ university, while failing to mention the unreliability of the attachment measures, incorrectly claimed that infants who spent at least one night per week away from their mothers had more insecure attachments than babies who saw their fathers only during the day. In fact they did not.
For example, Fabricius, et al reported that,
Karina Sokol, conducted a test for linear relations in the Tornello et al. data and found no correlation in these data between the absolute number of overnights with father and insecurity with mother.
Anti-father advocacy doesn’t get much clearer than that. Frankly misstating one’s own findings in order to further one’s anti-dad narrative should be beyond the pale. But when it comes to that very narrative, apparently anything is acceptable.
From there, the Tornello work goes from bad to weird.
More than half of the children classified as frequent overnighters lived predominantly with their fathers. But the data were reported and interpreted as if the mother was always the “resident” parent and the children were overnighting with a “nonresident” father. Thus the “resident” and “nonresident” parents were mislabeled.
Yes, that actually happened. The simple fact is that no one (I suspect including Tornello, et al) can tell which parent is the resident parent and which the non-resident one. More than half of the children who were classified as frequent overnighters weren’t overnighters at all, at least not with their fathers. Did they have overnights with their mothers? If so, how frequently? That renders the study a murky muddle whose results are impossible to decipher in any way. The Tornello study means nothing except that its authors need some lessons in how to conduct social science inquiries.
And let’s not forget that, since the sample used in the study came from the Fragile Families and Child Well-Being data, it in no way is representative of people generally. The Fragile Families data have never been intended to provide information about the general population. That study is a very valuable effort to tell us more about people who tend, far more than most, to be poor, black or Hispanic, unmarried, etc. It’s given us a huge database of information that countless researchers have utilized to delve into a welter of different questions.
But Tornello’s pretense that those data are in some way representative of the general population is just one more example of their agenda-driven narrative masquerading as science.
Sadly, I can’t cover all the many deficiencies of McIntosh and Tornello’s work, but other topics demand attention. I’ll finish this up tomorrow.
#children'swell-being, #fathers'rights, #sharedparenting