Comment by jawns | Hacker News Reader

Comment by jawns | Hacker News Reader

The experiments this op-ed describes are really neat. I especially like the one that demonstrates the importance of "practice what you preach."

Here's another cool experiment related to social/moral development, which you can perform on your own kid:

Have two grown-up friends each show your baby a toy. The first friend should offer the toy to your kid but then retract the offer and keep the toy to himself. The second friend should offer the toy to your kid but "accidentally" drop it out of reach.

Now, give your baby a different toy, and then prompt her to share it with one of the friends.

Your baby is very likely to share it with the friend who dropped the toy, rather than the one who withdrew the toy.

The above experiment is based on a 2010 study involving 21-month-olds. It found that two-thirds of the kids shared with the person who dropped the toy -- and the other third kept the toy to themselves. Not a single one shared with the person who offered the toy then retracted the offer. (I talk about this more in "Experimenting With Babies" http://www.experimentingwithbabies.com)

Not only is this fascinating because it shows that babies act preferentially toward people who are kind to them ... but it holds true EVEN IF the person who initiates a kind act is unable to complete it, but demonstrates the proper intent. Think about the complexity of the cognitive processes that are going on in that evaluation. It's amazing!

That book has interesting experiments, but suffers psychology's common problem (especially baby psychology): massively over-extending the experimental data into unsupportable conclusions that ignore loads of confounding factors.

Example, quoted from the website:

""" The babies in the unjustified group spent more time looking at the adult's discordant expression and exhibited more checking behavior. Additionally, the babies' facial expressions only displayed empathy for the adult when a negative emotional response fit the scenario.

The takeaway:

Parents, take note: Your toddler can see through your "stiff upper lip" routine.

"""

The "takeaway" is completely unsupported. In no way does the experiment suggest that hiding pain is useless.

For one thing, "stiff upper lip" (aka "faked neutrality") is different from faked happiness.

For two, the experimental methodology only applies to cases where the baby knows that the experience's standard emotional response. That applies to hammer strikes on a finger, but not something like divorce or losing a job.

For three, "spent more time looking at the adult's discordant expression" shows that the babies learn from their parents reactions, as they expend effort to understand counterintuitive behavior. "Counterintuitive" is not the same as "fake" or "wrong".

So the "takeway" may in fact be the opposite of what the data suggests!

Thanks, my thoughts exactly.

Coming from a machine learning background, I am often shocked at the quality of experiments in psychology and even economics.

I think the reason may be because in machine learning, you describe your algorithm and experiment exactly. If someone can use the same algorithm and devise an experiment to disprove your results, you look like a fool. With softer sciences, maybe people just say, "then you obviously didn't do it right!" and there is no code to point at.

I like to think that in medicine they know how to do things properly, since there is so much regulation and money on the line.

In machine learning you are implementing known algorithms on fully understood systems.

Studies of humans (medical, psychological, economic, etc) are attempting to discover unknown algorithms on unknown, extremely complex systems. The quality will always seem lower than engineering studies because scientists don't know what they don't know.

I don't understand this at all. At 21 months, a toddler can walk and talk, and you can play keep away with them for fun. I don't know what it means to drop a toy out of reach. Do you drop it on a high counter or something?

In the original experiment -- see http://www.ncbi.nlm.nih.gov/pubmed/20424094 for the abstract -- the toddlers stood on one side of a slanted table, so that anything placed on the table would roll down to the other side, where the adult actresses were. Presumably, the toddlers were not able to walk around the table.

One way to replicate this at home, albeit imperfectly, is to place your toddler in a high chair, so that if you drop the toy on the floor, she is unable to reach it.

I wonder if they controlled the effect of the actresses own response. Although they might be excellent actresses, being mean to a baby is probably going to make you feel pretty rotten, at a level it might be hard to control, and the baby could be picking up on that.

Animals can pick up subtle cues that are not apparent to adults; I'd be amazed if babies weren't at least as good.

I'm thinking of Clever Hans http://en.m.wikipedia.org/wiki/Clever_Hans

Both actions seem to qualify as "mean to a baby" to me. Assuming the actresses agreed with me, this would cancel out the variable.

Point is, the whole gift giving/dropping/witholding might be a red herring, and perhaps the same results could be replicated by having both people give a toy to a baby, but one acting scary and mean. And perhaps if two mannequins went through the same giving/dropping/witholding experiment, the result would not be observed. And perhaps if a smiling actress stole the toy, while a scowling actress dropped the toy, the affect would be reversed.

I was thinking about that, and the thing is people are often deceptive to children for their own good, as a way to handle them without upsetting them (arguably not ideal parenting but let it pass). So that, if you convince the child you really did drop it, you won't feel bad. But if you are overtly mean - it's unmistakeable that you deliberately the toy back, then it's hard to kid yourself that the baby didn't realize. Now did the baby realize? That's the whole thing we're trying to determine! I'm just saying that the actresses' model of what the baby will understand and won't will affect how they feel about it.

This of course all depends on the actresses having a particular model of the baby's ability to model them.

Simple enough to swap the actresses around to remove the individuals as confounding variables.

On a related note. HN gets pretty obnoxious nit picking about experiments. Guaranteed half the posts about a social science experiment come from novices whining about some flaw only they can see in the experiment yet a hard science paper gets posted and surprisingly no one has any complaints about the experimental design.

How would swapping actresses help?

The issue is not a quality of the individual person, but of their action - and the relationship between their action and the baby's later response to that person is very thing being tested.

BTW: Whatever your intention, calling people "obnoxious", "nit picking", "novices" and "whining" because they question something does not actually raise the level of discourse.

>How would swapping actresses help?

Using many actresses would allow the investigators to see if an individual receives a different response compared to the overall group. But it looks like you actually meant say the babies knew the testers felt bad so refused to share, not because the baby had been slighted. I think you need to read more on what the study tested.

>BTW: Whatever your intention, calling people "obnoxious", "nit picking", "novices" and "whining" because they question something does not actually raise the level of discourse.

Reread your comment:

>I wonder if they controlled the effect of the actresses own response.

The level of disrespect you show is amazing. It would be one thing if you actually read the study and realized that "Hey, they used the baby's mom and a drug addict in withdrawal as the two testers! Clearly the study is flawed." Nope, your first comment is essentially "I wonder if these professionals are incompetent."

What did the study test then? (I don't have a subscription to "Psychological Science", so can't read the full text, at least, not from the above link).

From the abstract on the first expt (and from the story linked here on HN), it seems unambiguous. My understanding is: one actress offered then refused the toy (intentional); a second actress offered but accidently dropped the toy (unintentional). Later, the baby treated the two actresses differently.

My thinking was that it would be impossible to control for the effect I described (i.e. the Clever Hans Effect), given the social nature of what they were investigating. You probably already know about Clever Hans and so didn't read the wikipedia article I linked, but it's pretty entertaining/interesting, and addresses the problems of controlling for the effect in some social experiments (EDIT: what I'm referring to is that it says in some social experiments they can't control for it just have to accept the effect may be there).

It would be really cool if they had a clever way of controlling for it (I've come across a few very clever experimental designs in Psych studies), so I was wondering if they did or not, and if so how.

I didn't think the conductors of the experiment were incompetent.

Thanks for rereading my comment.

Yes, this was my first thought too. The results need to be taken with a healthy dose of scepticism when there is no double blind and the interpretation of the result is subjective.

As the parent of a 21-month-old, I can assure you that nothing is out of reach—especially the counter.

Mine used two boxes of different sizes as stairs to reach the counter and the retrieved chocolates from the cabinet above. Then she climbed back down and moved the boxes back to where they came from so we wouldn't know. Only then did she relax and eat the chocolates.

I thought my wife was stealing all the chocolates for days until I caught the toddler doing it.

Mine has figured out how to climb the dog to get at the cat's tail; something that the toddler seems to view as a nutritious and delicious food item.

I fear the dog may have been the mastermind in this plan.

Who would have thought that 21-month-olds are predisposed to Kantian ethics?

Philosophy: how does it work?

Hint: its true even if babies haven't figured it out yet.

Well, should you take a cultural relativist POV then there's nothing "true" about Kantian ethics—it's a choice of values. It's interesting to see 21mo behavior selecting for those values.

If you take the deontological view then this is still a fairly complex behavior to execute. To consider whether this sort of judgement is one that should be fairly placed on the 21mo if it did the same is a challenging mental task.

So, perhaps babies are predisposed to heuristics which favor the categorical imperative. That'd be interesting and certainly worth surprise.

Many contradictory views are all "philosophy." It's rarely obvious which view is true, if any are.

Or could you say they're enacting a tit-for-tat strategy?

Get back to traveling.

Inside jokes, or humor of any kind, have no place on HN, Ben. May you be stripped of all your imaginary internet fun points, so that you learn not to be such a light-hearted and well-adjusted individual.

This book looks awesome. Grabbing one tonight thanks for letting me know about it!

Reply on news.ycombinator.com