Denise  00:03

Denise, hi everyone. I am Denise, a longtime listener of your parenting Mojo. I love this podcast because it condenses all the scientific research on child development, compares it with anthropological studies, and puts it into context of how I can apply all of this to my daily parenting. Jen has a wealth of resources here, so if you’re new to the podcast, I suggest you scroll through all her episodes. I’m sure you’ll find one that will help you with whatever you’re going through, or one that just piques your interest if you’d like to get new episodes in your inbox, along with a free infographic on 13 Reasons Your child isn’t listening to you and what to do about each one. Sign up at your parenting mojo.com forward, slash, subscribe. Enjoy the show.

Jen Lumanlan  00:58

Hello and welcome to the your parenting Mojo podcast. Have you seen the Wonder weeks book, or did you download the app? The book has apparently sold over 2 million copies, and the app has been downloaded over 4 million times. So, the approach has certainly struck a chord with parents. It seems to help that the book was developed by two PhDs, Dr Franz ploy and his wife, Hetty vanderai ploy, whom I will refer to together as the ploys. And I know that when I see PhD following the neighbor of an author, I perceive the author as having some credibility. They’ve now been joined by their daughter, Xavier plus ploy, as the CEO of the Wonder weeks. Apparently, Hetty actually died quite young in 2003 so when you hear me refer to Dr ploy. Later in the episode, is Dr Franz ploy. So, in this upcoming pair of episodes, I want to ask two overarching questions. In this first episode, we’ll ask is the idea of Wonder weeks backed up by scientific research. And then in the upcoming episode, we’ll ask, okay, based on what we learned here in the first episode, what if anything should we do with the ideas in the book to help us and our baby? So, if you’re expecting or you have a child under the age of one, this episode is very much for you. Let’s get started with our first question and look at the scientific research behind the Wonder. Weeks. So, Franz ploy studied biology and psychology. His wife, Hetty, studied educational psychology and anthropology, and together, they observed chimpanzees with Jane Goodall in Gombe National Park in Tanzania between 1971 and 1973.

Jen Lumanlan  02:30

Franz obtained his PhD in the behavioral development of chimpanzee babies, and Hetty got hers in mother baby interaction in chimpanzees. Because these dissertations were submitted in the early 1980s and fortunately, I couldn’t find any copies online, possibly as part of their dissertations or in related work. The employees noticed that baby chimpanzees seemed to go through a series of what they termed regression periods, by which they meant a return to behavior like clinging to their mother and nursing often which they hadn’t done so much only a week or so previously. The ploys hypothesized that each of these regression periods was followed by a developmental leap, and they wondered whether similar periods might exist among human babies. Several previous researchers had generally coalesced around the idea that there are four major behavioral transitions in human babies. And in the late 1980s the ploys began a research project to see if these were all of the transitions, or if they could find any more. And they decided to do this by identifying periods of regressive behavior in babies, because there was agreement in the literature that these regressive periods accompanied developmental transitions. They say that this literature finds that some sort of transitions do exist, and that these occur at two, 712, and 18 to 21 months. Unfortunately, the literature here mostly consists of books rather than peer reviewed papers. It’s not an exaggeration to say you can pretty much publish anything you want in a book. That’s why we have the peer review process in journals, so other people look at papers before they’re published and check that they are grounded in previous literature. It isn’t a perfect process by any stretch of the imagination, but at least you know someone who knows the subject has checked it out. The ploys cite 13 authors or sets of authors in support of this claim, so I tried to track each of them down. Unfortunately, seven of them were book chapters or books themselves, and of course, these are books published in the 1970s and 80s, so I wasn’t able to get a hold of them. One paper was a discussion about relationships between peer infants, which isn’t really related to what we’re talking about. One was a newsletter. One was peer reviewed and talked generally about development, but not specifically about developmental leaps. Two papers by the same author did talk about shifts in development, but didn’t specifically say At what ages they occur. One of the 13 papers by McCall and colleagues, published in 1977 did specifically discuss finding shifts at two, 813, 21 And 30 to 36 months. So, they agree with the ploys that there’s a shift at two months. But other than that, there’s complete disagreement on the timing of the shifts. They cite other research indicating hypotheses about more regression periods, and conclude quote only after the regression periods are firmly established, can one address the quest for the number and the nature of the bio behavioral transitions. So, this is what they set out to do. The ploys. Were looking for the same kinds of regressive behavior that they had seen in chimps, which was seeking proximity to a preferred individual, usually the mother. When they worked with these chimps, they would primarily use direct observation, watching the chimps each month until they had 300 total minutes of observation time. They wanted to see if they could identify regression periods using a questionnaire and interview format with human mothers. So, they developed a questionnaire that they gave to mothers who agreed to participate in their study. In their paper, they said the questionnaire asks if the child was more fussy than usual, if they were looking for more body contact, if they were sleeping less, if they feared other people had more childish behaviors, ate less in a meal, had problems in changing and dressing. Seemed to decrease their activity. Wanted more cuddles from mom, and cuddling things other than mom, obviously not all of these things happened from the earliest ages. The first ones on the list happened at the earliest ages and continued, and the later ones on the list happened as more as the babies got older, as an additional source of information, a subset of the mothers were asked to record the infant’s location and the kind of interaction they were involved in for an entire day once a week, the so called check sheet was divided into half hour periods, and the mothers were asked to record whether the baby had spent time on her lap, in contact with her, within 1.5 meters of her in its playpen, in a chair or in another room. She also noted whether the baby had been awake, asleep, crying, playing on its own, been fed, been changed, bathed, been carried around andor been played with by either her or others. A further subset of these mothers were observed directly by the ploys who would code the interactions between the mother and the baby using the following criteria. Number one, the baby’s weight is supported by mother and their fronts are touching. Number two, fronts are touching, like in the first category, and also the mother’s arms supporting the baby’s back and head. Number three, on the lap, meaning the baby’s weight is supported, but there’s no front to front contact. The baby might be facing a mother or have their back toward the mother. Number four, they’re in contact. So baby is standing or sitting on its own, but its body is touching its mothers. Number five, the infant is within arm’s reach, which is defined as 1.5 meters, but they aren’t touching each other. Number six, they’re in the same room, but outside of arm’s reach. And number seven, they aren’t in the same room and they’re out of visual contact. These categories are considered to be mutually exclusive, so a mom and baby pair can only do one of them at once. The researchers noted that whenever the pair changed from one category to another, and the percentage of time was spent in each of the six categories was used along with the questionnaire data. The main finding of this study that took place over 20 months was that most mothers found that at a certain age, their babies were more tiring, difficult and demanding, and that at other periods, this was not the case. Difficult periods were accompanied by increased crying and body contact and a decrease in the amount of sleep. Deploys called these periods regressions to validate the questionnaire data, deployers used the observational data and calculated the time that mother and baby spent in physical contact and plotted this against the baby’s age in weeks. The observational data was pretty well correlated with regressive periods that the mothers reported in the questionnaires, which isn’t massively surprising, since they’re essentially getting at the same information from very same research subjects in two different ways. Then they tried to link the regression periods with information that was already known about developmental shifts that babies make in their first 20 months. Employees didn’t do any statistical analysis with the data from the questionnaires. They simply looked for a high percentage of participants reporting similar data, which means their data are more likely to find a result where there really isn’t one. They did do some statistical analysis for the subset of mothers they observed directly looking at the relationship between the time the baby spent in body contact with the mother and whether it fell inside or outside of a regression period that the mother identified in the questionnaire. The generally accepted p value is point 05 which means if you get a P value of point 05, or less in your analysis, then your evidence is strong enough to reject your null hypothesis, that there is no link between body contact, between baby and mother and mother’s report of fussiness, then we say your results are statistically significant. The ploys never explicitly state their null hypothesis, but we have to assume is that there was no relationship between the observed data and the data from the questionnaire. The probability of the distribution of frequencies under the null hypothesis yielded a p value of point 03, which does not reach significance because that number is greater than point 05, once they combine the observations instead of looking at them in. Individually, they did get very good significance with a P value of point 004, so much less than the p value we’re looking for. Okay, so far so good, right? They got statistically significant results when they did the analysis in a certain way. Well, let’s look at the information again and just see if there might be more here than meets the eye. So, the mothers, and they were all mothers in the study were not randomly selected. They were recruited by midwives or teachers of toddler play groups, and then the mothers wrote a letter to the researchers. The researchers interviewed the mothers and quote selected healthy mothers who had no previous medical and psychological problems, who looked forward to having the baby, intended to look after it themselves were financially secure and had an extended support system, which consisted of family living in the same town and free access to a Child Guidance clinic. End Quote, and everyone in the Netherlands could access those Child Guidance clinics at the time, all of the babies were healthy, full term newborns, except one who was born four weeks early, but after four weeks in an incubator because of possible prematurity of the lungs, but apparently the baby had no postnatal problems. Quote, there was some economic diversity in their sample, but all of families were Dutch and Caucasian. Maybe you’re already seeing some red flags here, because I know I am. Well, firstly, we got that highly homogenous group of mothers, which the researchers specifically selected because they were homogenous in the paper, they said that. Quote, we followed a very strict selection procedure to find absolutely problem free and healthy families. We wanted to understand normal development before venturing into understanding pathology. End quote, the unstated assumption here is that normal development is the way that white children who don’t have any major life stressors develop, which is an idea you may recall we’ve won in run into once or twice in the past, most notably in the development of the your x year old child books. There was some economic diversity in the sample, but the effects of that are going to be muted in a country where there’s a high degree of social support, even the mothers with lower incomes had regular access to the Child Guidance clinic. People in the US have access to that same kind of support that the mothers in this study had would definitely be middle class. The mothers in the study were repeatedly asked if they were experiencing major life stressors, and none of the mothers reported experiencing any major stresses during the entire 20 months of the research, which seems kind of wild here in the US, where half of adults worry that a major health event could lead to a bankruptcy, which would likely result in losing your home and food insecurity as well.

Jen Lumanlan  12:38

Moving on to the mental health implications of that highly homogenous group of mothers, we have to remember that pathology means the study of disease. So, the ploys see any of the stressors that are associated with poverty and a lack of social support and major life changes as likely to cause disease, and that good middle class white families won’t experience those stressors. So if we understand how those middle class white families operate first, then we can get everyone else to act more like them, and then we won’t have any more social problems, because children who are raised like middle class white children are raised don’t cause problems, and certainly having a mother with any sort of previous medical or psychological problem, and who might have been feeling ambivalent about having a baby, and who lives in a country where there’s no legally required paid medical leave, so she might not be able to stay home with a baby, even if she wants to, who doesn’t have a husband with a stable job and a family nearby to help, is pathological. The second major issue is that while the authors stated that the mothers didn’t know the study’s purpose, meaning they were blind to its purpose. There is no mention of anyone other than the authors doing any of the work in this study, which means the study authors were not blinded, and that means the people who were doing the coding knew what they were looking for and were more likely to make observations that fit with their hypothesis that regressive behavior appears in certain periods because the researchers weren’t blind to what data they were trying to collect. We can imagine a mother answering the door with a smile, and the researcher thinking, looks like we’re not in a regression week, and codes their observations accordingly. Maybe another morning, the mother opens the door with an eye roll, a sigh and clothing covered in vomit, and the researcher thinks, oh, it looks like a regression week to me and then codes their observations accordingly. The researchers certainly knew the exact age of the baby they were visiting, so they may well have had a hypothesis about whether they were visiting during a regression week, even before they got to the door, which could have biased their data collection. The visits weren’t recorded using video or even audio only, so we have no way to have a second researcher corroborate the coding. Thirdly, the primary data source was self reported by mothers, which means the mothers were the ones filling out these logs every half hour, once a week. I don’t know if you can cast your mind back to being the parent of a young baby, or maybe you don’t need to cast your mind back at all, and you’re in a. Right now and either remember or imagine what it would be like to mark down every half hour whether your baby had spent time on your lap, in contact with you within 1.5 meters of you in its playpen, in a chair, in a push chair in another room, as well as whether baby had been awake, asleep, crying, playing on its own, been fed, been changed, been bathed, been carried about and been played with either by you or others. Is there any chance you might forget whether the baby had done each of these things and which half hour period it had done them in? Is there any chance you might have forgotten to do the log entirely and done it in five minutes at the end of the day to the best of your recollection, relying generally on your impression of whether your baby seemed more or less fussy than usual at the moment, and then back filling the data to fit that. I’d say there’s a decent chance that this did happen on more than one observation for each child, which means our data may not be very reliable. A PhD thesis by Dr Ashley woolmore that we’ll look at in more depth later said that the largest source of data loss in their separate study was written questionnaires not being returned, and when the researcher interviewed participants, they would often learn about families moving house, changes in daycare arrangements and other stresses in the household that were rarely volunteered in the questionnaire responses and which undoubtedly had an impact on the infant’s Fauci Ness. The fourth issue is that we don’t know much about what kind of conversation the ploys had with the mothers. They observed in person. We know the ploys tried to blend in with the background for most of the time and not interact with the mothers, and they asked the mothers if they changed their behavior because the observers were present. And the mothers report did using a little more modesty with nursing and using the bathroom. We don’t know which of the ploys did these home visits and whether there were differences in the observations by Hetty and Franz for the reasons that mothers might have felt more comfortable with their bodily functions and conversation being observed by another woman. Issue five is the ploys adopted almost exactly the same set of criteria for the direct observations as they’d used with chimpanzees in Tanzania, with no discussion on the appropriateness of these criteria, the most obvious issue is the definition of within arm’s reach. Since chimpanzees have really long arms, the ploys use the same 1.5 meters to define within arm’s reach in their chimpanzee studies and with their human study. But a chimpanzee’s arm span is about one and a half times its height. So, a chimpanzee of average height may have a total arm span of about eight feet, which is four feet from midline to end of fingers. By comparison, I’m about five feet eight, which is reasonably tall for a human woman, and the distance from my midline to fingertips is three feet, which is a 25% difference. So should the difference that’s considered within arm’s reach really be the same for both species. A similar issue arises in the sixth criteria in the observational study, which is within the same room but out of arms. Reach. This criteria is a bit different from the categories that use with the chimps, which were within five meters, within 15 meters and visible but over 15 meters away. Since chimps move over a larger area than human mother baby pairs do in the observational data, the researchers coded time spent in the playpen or outside the playpen, and we think about both that and the amount of time mothers spend carrying their child, we start to realize how many cultural assumptions there are in this data collection. We don’t have any information on what parenting practices these parents followed. Maybe they were following Attachment Parenting, which, by the way, is different from attachment theory. We covered that in episode 138 on most of what you know about attachment is probably wrong. If they were doing Attachment Parenting, then they were probably carrying their babies a lot more than average, potentially most of the babies waking and even sleeping hours. And perhaps the babies fussed less because of it, because there’s no need to fuss when you’re seeking proximity to a caregiver and you’re already touching them. And lest you get excited about Attachment Parenting and think that it’s going to be the mythical parenting method that saves you from all crying, it’s pretty common for former attachment parents to find my work when they get burned out from carrying their baby all the time, because carrying baby all the time is feasible when we live in a culture where we have a whole bunch of other adults around who can help carry the baby, it can be a recipe for burnout when we’re trying to do all that by ourselves in our own house, with no other support, and potentially with one or multiple other children’s needs to consider. Or maybe these parents were following the resources for infant educar is approach known as rye, where we put babies on a mat and allow them to kind of entertain themselves for a good chunk of the day without needing as much interaction from us. Our family never had a playpen, and if they are used by parents practicing rye, it’s probably to help separate an older child from a younger child, rather than as a way to contain a baby. I’m not saying that either of these methods is necessarily better for all mothers or all babies, but the parents do have different ideas about how to raise their children, and to the extent that we ignore this, the data about how much time they spend in proximity to each other has an awful lot of cultural issues that affect it that the researchers aren’t. Ring issue six. As is that this study we’re discussing was published in 1992 which was the same year as the Dutch edition of the Wonder weeks book, which was published under a title in Dutch that translates to oops, I’m growing the first English translation wasn’t published until 2003 so the ploys data on human subjects from one study hadn’t even gone through full peer review by the time their instructional manual for parents, based on their data, was published, which seems a little bit forward to me. It’s not like their work was confirming decades of research that came before it. These were pretty new ideas at the time, employees state in their paper. We’ve been looking at that quote. Of course, it goes without saying that the phenomenon of regression periods should be firmly established in follow up studies. One of the first things to be done, in our opinion, is to measure directly or more precisely the various aspects of the regression periods as reported by the mothers. For instance, the increases in crying can be measured by using a voice activated tape recorder and calculating the number of minutes crying per day, or the decreases in sleep can be kept track of by keeping a sleep log and calculating the number of awakenings per night. The increases in Mother infant contact can be directly observed and recorded with an event recorder, as we did, but the observation should be carried out at least weekly, in order not to miss any period. Now that we have a lead from the maternal reports. This has become possible. End quote, and this leads us right into issue number seven, which is the number of participants and studies this work is based upon. I’m sure you might have imagined the Wonder weeks empire, with its multiple book editions and its ubiquitous app, and you might have thought this must be based on lots of studies of 1000s of babies, right? How else could they possibly know that these 10 predictable leaps work for so many babies, since the app is available in 16 countries and the book has been translated into 20 languages, well, okay, maybe not 1000s of babies. That sounds like a lot, but it must be hundreds of babies, right? Hopefully many hundreds? Well, no, it isn’t actually hundreds either. It isn’t even 100.

Jen Lumanlan  22:07

This initial study, which was the only study of the phenomenon of regression periods in human infants at the time, the first book was published in Dutch, looked at 15 mother infant pairs 15, and then the researchers did the in home observations, which were so important to understanding the regression periods with two of the pairs for initial study, 15 is not a terrible number. It can tell us if there’s something there worth investigating in greater depth, 15 does not give us statistically great results. Okay, we’ve already talked about some of the statistics here, and I’m going to go into this kind of at a slightly deeper level. We need to understand two statistical concepts, right? Power and p values. We already talked about p values a little bit. Power is related to a number of things, but one is the failure to reject a null hypothesis, and we can typically consider a power of 80% or higher to mean there’s a possibility of either correctly detecting an effect or accepting the null hypothesis. A null hypothesis is the opposite from what the researchers expect to find, right? They usually expect to find some interaction between two variables, so the null hypothesis is that there’s no interaction between the two variables they’re considering. And all the ploy statistical analysis tells us is that mothers who report their babies are fussy at the moment tend to hold their babies for longer in that period. That’s it. So, coming back to our statistical significance issue, I do think we can assume the null hypothesis is there’s no link between baby’s body contact with mother and mother’s report of fussiness, which would mean the mothers see their babies are fussy and are deliberately or non deliberately, spending less time holding the babies in those periods. So, this p value, right stands for probability value, the number with describing the likelihood of finding the data that you found in your study, which you think tells you something important when there’s actually no real effect. So actually the null hypothesis is true. In other words, it’s telling you how likely it is your results happen by chance. So, recall that the generally accepted p value is point 05 if we if your P value is less than that, we say your results are statistically significant. And so, power and p values interact with sample sizes, which is why this is important here. And the main way to make sure you get enough power and a statistically significant result is to use an equation to calculate what sample size you should use. When you test more people, there’s a greater chance that the result you’re finding is actually because there’s an effect to detect, not because all the people you happen to test, all by chance, have something in common. So if we want an 80% power level, which is also widely accepted in the psychology community, and a p value of point 05, this equation tells us we need a sample of 126 which is over eight times as many mother infant pairs as the ploys actually studied with a sample of 15 that they used if they have a power level of 80% Their P value is point one, five, which no serious researcher would ever say is significant. Or if we hold that P value of point 05 we’re looking at a power level of around 18% not the 80% we wanted to be reasonably sure we’re detecting an effect that’s there, or knowing that we should accept the null hypothesis. So, we should call this a pilot study, which is a study researchers do to find out if there’s more worth investigating, but that they would never describe as offering definitive results. And we hope that the ploys or other researchers did the follow up studies using the voice activated tape recorders or the sleep logs or weekly observations, right? Well, no, now we turn our attention to the topic of replication, which is the idea that if the results on a study are real, that other researchers using the same methods will be able to find them as well in 2003 Dr Franz ploy co-edited a book called regression periods in human infancy, which contain the results of three more studies on regression. The first was done by Dr Marta Sederni and Dr Carlos Rostran at the University of Girona in Spain, and they had an explicit goal of confirming Dr ploys results that he had obtained in the Netherlands. Once again, these researchers selected mothers and babies of, quote, middle socio economic and cultural class. End Quote, with sufficient social support and quote, without any relational problems of any kind. End quote, we aren’t told whether the mothers knew what the researchers were looking for, or whether the mothers were blinded to the research, even if the researchers themselves weren’t. Soderney and Rostran found regression periods with a mean length of two weeks and a range of one to four weeks, the distribution of the regression periods was much more diffuse than Floyd’s original study. The children were not always distressed to the same level within a regression period, and mothers were not always annoyed by the regression to the same degree. Overall, the picture is of data that is pretty fuzzy. There’s something there, but it’s much more difficult to see what it is in a separate paper, Dr Sidernie Dr Ploy and another colleague of Dr ploys looked at the regression periods they found. So, I compared these with the periods in the Wonder weeks book. Both of them found regression periods at around weeks four to five and eight to nine. Soderney found one at 12 to 14 weeks and then 16 to 20 weeks, while wonder week says there’s one at exactly 12 weeks, and then a smooth period, and then another regression at 15 to 19 weeks. There’s more agreement on a regression at what wonder week says is 23 to 26 weeks, and Soderney says is 24 to 27 weeks, wonder week says there’s a regression from 29 to 30 weeks, which is a clear, calm period for Soderney. And then they agree again a bit more with wonder weeks saying that there are regressions at 34 to 37 weeks, 42 to 46 weeks, and Soderney saying 32 to 38 and 42 to 48 which are both wider Windows than wonder week says Soderney’s Last regression is weeks 48 to 53 which only slightly overlaps with the Wonder weeks regression of 51 to 54 weeks. So, when we read in the studies abstract that quote a temporal relation between the regression periods found in earlier meaning in the ploys study and the transition periods reported here. End quote is sort of true. There is some kind of relationship, but there was no statistical analysis to understand the extent to which these two studies are describing the same regression periods or not. The authors note that. Quote, some scientists could prefer the use of standardized tests or scales of development in order to observe in a more systematic and precise way the process of change, end quote, rather than simply asking mothers if their children have learnt anything new this week. And yes, we might indeed prefer that approach, but we are not going to find it in this study. Oh, and their sample size was 18 mother infant pairs. So, let’s keep track. Shall we of the number of children that the entirety of the Wonder weeks method is based on so far we have the original 15, and now this new 18 for a subtotal of 33 the next chapter that we look at in the regression periods in human infancy. Book is a very interesting one, as it was authored by Dr Franz Ploy. It references a study by doctors de veer and vanguard in the Netherlands, and it seems as though the three of them had actually worked on the study together, but there was some kind of squabble over the results. Dr ploys chapter says that De Vere and Vanguard refused to allow him to publish the data, so it directs readers over to DEAVERE and Vanguard study to see the data before then describing ploys interpretation of it. So, let’s go to the data. First. The three researchers had observed four infant mother pairs. Now we’re up to a total of 37 pairs with weekly questionnaires and interviews, weekly observations of the pairs in their homes, in which the researchers had a timer going off in an earpiece every minute and they recorded. At that instant, what the mother and baby were doing, but they didn’t produce video recordings, so there was no way to double check their coding. They measured sleeping behavior automatically using equipment placed next to the bed, although they didn’t say what that equipment was or how it measured sleeping behavior, the mothers who participated were very much not blinded to the study’s purpose. In fact, the paper says the ploys book, which had been published in 1992 had been very successful with lots of newspaper and TV coverage. So because it was impossible to avoid the mothers hearing about the book or receiving it as a present, the researchers gave a copy of the book to each of the mothers to keep the study group as homogenous as possible. Strong regression weeks were defined as those in which as much as 80 to 100% of the mothers had reported that their infants showed regressive behaviors. Although it is hard to imagine how there could be 80% of four mothers, since three of them would be 75% and four of them would be 100%.

Jen Lumanlan  30:56

In the final paper, De Vere and bet van Geert said that their study, quote failed to replicate the ploys original findings, both for data obtained through Holmes observations and maternal reports. They said that quote, no conclusive evidence was found. That is evidence above mere chance level for the model of 10 regressions during infancy. End quote described by the ploys, they bring up a new issue as well. They found that some behaviors that looked like regression were actually due to teething and had to be Re Coded as not regressions. Okay, that makes sense, but it brings up another issue. What if a baby is teething on the weekly survey day and then the tooth breaks out after survey day, but the mother forgets to tell the researchers, then we have a fussy day coded as a regression period, when actually it was nothing of the sort. In addition, they cite research comparing daily maternal reports with observational data on the frequency and duration of breastfeeding, and found virtually no agreement between the two types of data, given that the mothers in De Vere and Vanguard study were being asked to recall a much greater constellation of behaviors than whether or not you’re breastfeeding on a weekly basis, not a daily basis. They say that, quote, the reliability of their reports cannot be directly assumed. They go on to say that, quote, this being said, it still does not explain why, reliably or unreliably, the mothers of the ploy study reported regression behaviors in the same weeks. End. Quote, maybe we can hypothesize This is an artifact of their tiny sample size, which had such weak power that the ploys accepted the test hypothesis when they should have accepted the null hypothesis, that there is no relationship between infant contact with the mother and Mother reports of fussiness. A New York Times article reported on some of the controversy around the publication of the data in this specific study. So, after the study was published in a peer reviewed journal, Dr Ploy publicly denigrated the findings due to the two babies in the study who were often left in playpens. Apparently, he also violated privacy ethics by saying publicly that one of the mothers was under psychiatric care, and that it could have influenced her infants behavior. The University of Groinian declined to renew his contract after the incident, although Dr Floy insisted he left by choice and that he did not interfere with the publication of Dr de Villa’s work, which had been alleged. So now that we kind of have the background on the study they started together and then separated on right, we’ve looked at the data on that. Let’s come back to Dr ploys chapter in the regression periods in human infancy book, which is basically a rebuttal of De Vere and Vanguard study. So, ploy identifies a number of reasons why De Vere and Vanguard may have failed to identify the same regression periods that he did. He restates the, quote, very strict selection procedure that was used in the original study, because this was an attempted replication, the same selection criteria should have been used, but apparently, wasn’t. In addition, two of the mothers followed a very strict sleeping, contact and feeding schedule of the type advocated in the 1920s and 30s, when mothers were explicitly told to ignore their baby’s crying. The mother of the baby cone, named infant F said she let her baby cry, quote, in order not to spoil it. If the baby woke up and started crying, the mother left it crying in bed. If it was not yet the schema time to feed if the infant cried, it was less likely to be held at four weeks during the first regression period, the crying graph shows a peak, and then the body contact decreases dramatically, from 85% to 33% infant F starts smiling much earlier than the other infants, and very quickly began smiling a lot more than crying, because its mother would reward smiling with attention. Infant J and her mother were separated for 10 days at birth, and the mother did not object to this, saying that she, quote, was not up to it yet anyway, and both the mothers of infants F and J said in their entry questionnaires they were planning to use these strict schedules. So, it’s sort of a mystery why the three researchers working together at that point would accept. These mothers into the study, the mother of infant e started child minding two other children in her home 34 weeks after the birth. And remember, in the original study, the mothers were supposed to be financially secure enough that they could plan to stay home with the baby and look forward to it without needing or wanting to work themselves because their husband’s salary was sufficient. Of course, we’re assuming they’re in a cis het relationship, and they could never be single mothers, because how could you possibly work and be so available to the baby that they can be in contact with you whenever they want, if you’re working what we see here is a picture of a mother whose job it is to be available to the baby at all times. I’m not arguing that we should use the wake your child up when it’s time to feed it approach, far from it. But I also think that if seeing the regression periods happen is so reliant on subsuming all of the mother’s needs so baby can be in contact exactly when baby wants to be in contact, then something isn’t right here. Another big thing that comes up in this chapter, for the first time, is the impact of mother’s depression on parent child interaction patterns. Other researchers have found two patterns of interaction between depressed mothers and their babies, so called intrusive mothers engaged in rough handling, spoke in an angry tone of voice, poked at their babies and actively interfered with their baby’s activities. Withdrawn mothers were disengaged and responsive and emotionally flat. Infants of intrusive mothers spent most of their time looking away from the mother and didn’t look at objects either, and they didn’t cry very much. Infants of withdrawn mothers are more likely to protest and be distressed still more, researchers have found that depression lasting longer than six months has this kind of effect, and that maternal postnatal depression can affect the length and timing of regression periods. One of the mothers in DeWitt and bangurit study was depressed, which means that three of the four mothers in the study should not have been selected under replication criteria. And all of this explains why DeWitt and bangurt study failed to replicate the ploys original study. And it makes sense these authors weren’t the only ones to find that the significant findings completely disappear when the mother is depressed. The Wonder weeks book does mention depression in that it exists and that you should quote, look for the help you need, and don’t be embarrassed to ask for it, as if everyone could actually afford to go to therapy when it isn’t usually covered by insurance. There’s a comment slid in there that quote, today’s men are just as big a part of parenting as women, and go through it all too. End quote, When, sure, men listening to this podcast probably are more involved than the average man, but they’re not going through it in exactly the same way that birthing parents go through it. And the biggest point here is that the book never mentions that the data on when regression weeks show up doesn’t fit the pattern. If the mother is depressed, Dr Wilmore found a delay of the regression week by about a week in infants whose mothers were depressed, but argues that it, quote, seems improbable that this would have a detectable impact on human development. End quote, but ploy goes on to draw five main conclusions from Davia and Vanguard’s data. Firstly, the age length regression periods are reflected in direct observations of body contact. Quote, under normal circumstances, thereby the phenomenon is once again validated by direct observations. End quote, so the model works under normal, white, middle class circumstances. Secondly, under special parental conditions like rigid schedule care or depression and phobias. Direct Observation methods are not chaotic and do reflect age linked regression periods, but you have to look for the right signals. The researchers observing infant F realized the baby would smile when he was distressed because crying was so ineffective at getting his mother’s attention. So, what if this was the case with some other parents too? What if some of the parents in the original study were using a slightly less strict version of infant F’s mom’s parenting methods, and sometimes a child smiles instead of crying, and in 13 of the 15 cases where the mothers filled out questionnaires but were never observed directly, the researchers probably would not have had any idea. And also, if you’re depressed and have had a hard time responding to your baby, and your baby smiles at you instead of crying, it’s possible that this indicates a regression week. And what if there are other issues that affect relationships, like depression and what if there are other signals than crying and smiling that we know nothing about thirdly?

Jen Lumanlan  39:27

Quote, it has proven important to filter out large peaks in body contact that can be explained by obvious external circumstances such as chaos travel and training crawling or bipedal walking. End, quote, some parents train their babies to crawl by physically moving the baby’s limbs. So of course, mother and child are going to be in physical contact if that’s happening, and also if they’re training the child to walk by holding the child’s hands. It’s the chaos that most interests me in this statement, because chaos is a word that can describe all of the things that the. Employees don’t want to see in their study participants. Chaos could mean the uncertainty that comes with not being white and middle class and financially stable and having lots of family support and being in perfect mental and physical health. All of these could lead to the child seeking increased contact with the mom, or maybe even with the dad, if he was the only person available. But these things are called life. They happen to all of us, not all at the same time, and to some of us more than others. But what good is a model that can only be applied under so apparently, perfect circumstances. The fourth of ploys conclusions is that the context, parenting, regime and attitude and the age of the child have to be taken into account in the analysis. Well, obviously yes on the age, obviously equally obviously yes on the context parenting regime and attitude, but this was never considered in the Floyd’s original study. Finally, quote, data of individual Mother infant pairs have to be analyzed separately and cannot be lumped. End. Quote. This is interesting and sort of strange, because if the regression periods are so universal that they are applicable to all children, why do we have to analyze the data of individual pairs separately? Plus, you might remember from our description of the analysis done in the ploys very first study that they only got statistically significant results once they combined all of the mother’s results together. Ploy concludes that, quote, the discussion now needs to move on from the existence of the phenomenon of age-linked regression periods to looking at how an infant adapts to special parental conditions. End, quote, but I think we’re pretty far from there. At this point, we thought we had 37 mother infant pairs contributing data that corroborates the fact that regression periods exist, but actually we now have 34 pairs that do this and four that seem to show the opposite. There are two more studies that I want to look at fairly closely, because Dr ploy also cites these as examples of replications of the ploys original study, meaning that they tried to use all the same criteria, and they found results that corroborate the findings of the original study. The first was conducted by doctors Willmore and richer in the UK, and has the not incredibly promising title of detecting infant regression periods, weak signals in a noisy environment. This study reported data from 30 mothers involved in a larger study that turned out once I’d done some digging to be Dr Woolman’s PhD thesis. So, we’re now up to 64 children across all of our studies supporting the hypothesis. Dr Wolmer’s thesis looked at the difference in regression periods among depressed and non-depressed mothers, but in this book chapter, they only report data on the non-depressed mothers without explaining any more about the larger study or the ways that depression was shown to impact the results. So, in this chapter, we see the actual questionnaire that the researchers use in these studies for the first time, where they first ask if the baby cried or fussed more easily or had more mood swings. And then to continue with the rest of the questionnaire, if not, it’s not a regression week. The second question asks about whether the child wanted more physical closeness or was more demanding of mother’s attention. Again, continue if these are present, it’s not a regression week, if they aren’t, finally, if at least two of the following are present. It’s a regression week. The child had sleeping problems or nightmares. Had eating problems, resisted being changed, was shy with strangers, was less vocal, was less active, sucked their thumb more often, behaved more babyishly. Was jealous, wanting mother also themselves, was very naughty, was very friendly, or through more temper tantrums. The questionnaire doesn’t specify what eating problems mean. So, who knows what the mothers interpreted that as meaning. The Floyds say at various places in the Wonder weeks book that quote, both breastfed and bottle fed babies can temporarily have smaller appetites. End quote when they approach a leap, and that quote during a fussy phase. Breastfeeding babies often want to nurse more. It may seem as though they’re nursing all day long. End quote, so both feeding more and feeding less are potential signs of a leap in the book, although how they know that from a questionnaire that asks for a yes or no response to a question on whether baby has eating problems remains a mystery. Okay, so let’s look at that list of behaviors in the questionnaire, because there are some interesting items on it. I imagine sleeping problems is somewhat self-explanatory for most parents, although I’m not sure how we can tell if a child who’s less than two years old has nightmares on resisting being changed, being very naughty and throwing more temper tantrums, what I see is a child whose needs we don’t understand. Why are they resisting being changed? Why are they being naughty? We have no idea. Maybe they have a need for autonomy. Maybe they don’t want to stop playing. Maybe they don’t like being cold when they’re stripped off naked. We have no idea. All we’re being trained to see is resistance, which is apparently indicative of a regression period. So. But what if it isn’t my daughter was shy with strangers every time she met them, and she always wants me to alter herself. So, who knows how that would have been coded? Plus being very shy with strangers and being very friendly are both indicators of regression periods, which seem like they can be in opposition. Being less vocal is considered a regression item, but being fussy is a primary regression item. So how could you make it through the first question about more fussiness and then also say yes to was less vocal? We do know that not all of the babies in this particular study were first borns. We don’t know about any additional pregnancies or births that happened during the study. As you may well know, acting babyishly can be a response to sibling dynamics, and it can also happen in response to finding out that another baby is on the way. Although the researchers say that quote in signal detection theory terms these meaning their results are weak signals within a lot of noise, so there is an increased danger of making both false negative and false positive findings. They go on to say that their results provide, quote, strong support for the ploys findings for existence and timing of regression periods at weeks 1217, and 26 and they call this strong support, even though, during the peaks, regression periods are only reported by 40 to 75% of participants. They didn’t start collecting data until until 10 weeks so as to make sure to have data from children who were overdue. So, they missed the week four and the week eight leaps. They don’t say why. They stopped at week 26 they noted they had been able to identify regression periods in children of depressed mothers, but that the timing and duration was different, and when they combined all the data, then all significant findings disappeared. This is sort of getting worrying, because between 13 to 17% of mothers in the US do experience depression, with an especially high proportion of Latina mothers having it. But the Wonder weeks book doesn’t ask mothers if they’re experiencing depression and if they are that this separate set of patterns applies to you, it also begs the question, what other patterns are we missing by looking at these highly homogenous samples in the studies, what patterns would we find among children of mothers who work or single parents, or neurodivergent parents, or neurodivergent children who are obviously way too young to be diagnosed? How can we possibly know if this book is relevant to us or not? Wilmore and Richard introduce yet more reasons why this data is difficult to understand. This is the first time I’ve seen that the timing of regression periods is based on gestational age and not from age of date of birth. I was so surprised by this that I had to look closely at the book to find the single mention that the method is based on due dates. So I ran a survey in the your parenting Mojo Facebook group to see if anyone else had missed this, and I found that 70% of the people who responded said they did follow the Wonder week system with their baby. That was actually a lot more than I expected, especially since there are a lot of new apps on the market now to track every aspect of your baby’s development. Of those, 70% of the people who use the Wonder weeks, a whopping 18% of parents hadn’t realized the correct date to use is the due date rather than the birth date. Of course, in the studies, the researchers are the ones making all these calculations, but if all you have is the book and you miss the single time this is mentioned, you might be forgiven for wondering, why on earth this system doesn’t seem to apply to your child. But it turns out we weren’t the only ones to miss this in his PhD thesis, Dr woolmore says Vander Wright and ploy do not make it absolutely clear that the gestational age of the infant is used to determine the timing of the regression periods. However, their use of gestational data was later confirmed during the planning phase of this project, the issue was not considered all the timings for this project were based on infants age from date of birth. This resulted in discrepancies in the number of weeks that data were collected. For example, when prospective parents were referred slightly late, ie during the infant’s 12th week, and the infant was also two weeks overdue, the study could only start during the infant’s 15th week based on gestational age, oops.

Jen Lumanlan  49:16

They also say there’s a problem inherent in using a timing unit as long as one week, which meant the researchers would call the mothers on a weekly basis. The weeks were rounded, so an infant that was seven weeks and three days old is in week seven, while one that is seven weeks and four days old is in week eight. And while we’re at it, why do we consider the week an appropriate resolution to study this kind of issue at all, other than the fact that it’s convenient for the researchers, and it’s convenient to communicate to parents, the ploys have not presented any evidence whatsoever to show that a week is a duration of time that gives us an accurate picture of infants behavior. We could look at four-day cycles or six day cycles, or any other duration we like. Oh. Obviously they have to be cut off somewhere, but it seems as though if you’re looking for weak signals in noisy data, it might be a good idea to study children who actually share a due date. If you think gestational age is the primary issue, your data is organized around, but it turns out gestational age is not very accurate either. These dates are usually calculated using naglers Rule, which says we should determine the first day of the last menstrual period. Subtract three months, add one year and seven days. That produces a due date. There’s approximately 280 days from the start of the last menstrual period. Oh, and that rule was developed around the year 1850 and it has not been updated since one study found that last menstrual period based gestational age estimates didn’t agree with expert obstetric estimates. In 39% of over 600,000 live birth records in Ohio between 2006 and 2009 it also assumes a 28 day menstrual cycle and ovulation on day 14, when 1% of people who menstruate in one study of over 600,000 cycles have cycles of 15 to 20 days, 7% have cycle lengths between 36 and 50 days. Another study found that 46% of people in a smaller study of 786 cycles had differences in their cycle length of seven days or more, and 20% had differences of 14 days or more. And the differences are not randomly distributed. One study found the gestational length for white first time mothers is seven days shorter than would be expected using naglers rule. There’s also no one perfect method of determining the ovulation date, which varies from one person to the next. So, this whole idea is based on understanding the exact gestational age of your baby, which is not an exact science itself at all. Dr Woolner study introduced even more suggestions for reasons why children might exhibit regressive behavior that aren’t actually regressions, things like changes in the family routine, the arrival of a new caretaker, as well as separation from the mother, family stress we’ve been discussing. But the survey doesn’t ask about any of these things. It doesn’t say, Is your child showing any signs of regressive behavior that aren’t explained by changes in the family routine, the arrival of a new caretaker, separation from the mother, family stress, or any of the other issues we’ve been talking about in this episode. So, what if those things are happening? And we attribute it to a regression, but it isn’t really. The researchers say there were hints of these ideas in the data, but it was not possible to test them systematically. And in any event, it is unlikely that all sources of noise can ever be eliminated. Which brings up the question that underpins this episode, if we can’t eliminate the noise, what’s the point of understanding the regressions? They might have developmental interest for researchers, but if they don’t actually tell parents much, that’s useful, what is the point? And I thought of another issue related to all these family stressors. What if the baby was essentially the same from one week to the next, but the mother was feeling more tired or stressed out or even sick, and the very same behavior that seemed fine last week suddenly seems clingy and overwhelming, and Mother reports regression like behavior when it isn’t even there. Some of the mothers in this study return to work toward the end of the study period, which the researchers say did introduce more noise into the data. But they caution that, quote, studies that fail to find regression periods must be carefully scrutinized to ensure they have not simply failed to see through the noise in their data, such, they have mistaken their own failure to find the phenomenon for the phenomenon’s nonexistence. End quote. The final conclusion they draw about their data is that researchers should, quote, be careful about heterogeneity in the sample of participants, especially the existence of maternal depression. End quote. But the Wonder weeks book presents this information as if it’s relevant to everybody from every background in every country, with no mention of physical or mental challenges or any other variables we’ve discussed. The depressed mothers in these studies had behaved babies with clear, fussy periods at weeks 15 and 17, not the long peak between 14 and 18 that Dr Wilmore was expecting to find. He found a peak at week 25 instead of the expected 26 but each of these findings from a different data set, data that was corrected for gestational age and removing children with illnesses and data without these corrections, no single set of data from depressed mothers match the ploys data, and yet the ploys never mention in the book that depressed mothers can expect very different regression, period, timing. The final study that I want to look at from this book was a study in Sweden of 17 first born infants. So now we have a grand total of 82 children across these four studies, the children were divided into two groups with eight mothers responding to the usual weekly questionnaire and interview guide, translation. Into Swedish nine dyads when the observation group, which was further divided into two cohorts, one was observed from two weeks to seven or nine months and the second from seven or nine months to 15 months, no explanation given as to why of all this subdivision happened, or what if any impact it might have had on the results. During the period when the observation group children were not being observed. They were followed by questionnaires that the mothers answered. The observations consist of three-hour home visits once a week, when the researcher recorded the amount of time mother and infant had body contact, and the amount of time baby cried. Fretting and fussing were recorded once every 60 seconds due to confounding factors like illness. The number of infants included in the analysis varies from 13 to 16 per week, which we have to imagine happened in the other studies as well. But no other authors explicitly state this or told us how they accounted for it in their analysis, the relationship between the number of identified regression periods and the number of expected periods was above chance level. But only 55% of the regression periods were observed within the expected weeks. Over 60% of the children were identified as regressive during weeks 14 to 1932, to 37 and 40 to 46 weeks periods not directly supported by this data were four to five, seven to 911, to 1222, to 2649 to 52 and the 61 to 64 week periods, so two thirds of the regression periods in the Wonder weeks book. The researchers hypothesized some reasons for this finding, including less homogenous development after age nine months, cultural differences between the Netherlands and Sweden, and individual variability between infants due to factors not controlled for, like temperament. And noted that, quote, all of these factors need to be addressed in future studies. End quote, it’s also possible there were some translation issues. Dr Wilmore realized in the discussion section of his PhD thesis that a question on the ploys questionnaire in Dutch asks whether the infant is quote, tiresome, which in in Dutch means that the child is feeling out of sorts. But in Walmart’s study, the English translation meant the mother was having a hard time that week because of the child’s behavior relatedly, Dr Woolner wonders whether mothers perceive infants are more challenging during regression periods, which occur on approximately a four-week cycle due to factors unrelated to infant development, such as menstrual cycles, the missed translation of tiresome certainly could have contributed to this. In Dr Wilmer thesis, he says that he did not generate evidence to support regression at weeks 12 or 26 but this cannot be interpreted as failing to provide support for their existence because of the translation issue, but it also can’t be used as support for their existence either, which is exactly what Dr ploy goes on to do, as we will see very soon. Further, Dr Wilmer said that when he asked things like, quote, have you been able to make any time for yourself this week, and have you picked your baby up when he or she has wanted you to more often this week, mothers often asked how other participants had responded to these questions. Some mothers seemed reluctant to give responses that could be interpreted as them struggling to look after their baby, an issue which would be totally missed in the yes or no responses to written questionnaires.

Jen Lumanlan  58:21

So, the upshot of all of this is that we are looking at studies of a grand total of 82 children, three of which are described as replications of the original study, but which come nowhere close to finding the same outcome. And in his book chapter summarizing the findings of these so-called replications, Dr ploy states that quote the evidence presented herein provides good support for the claim that age related regression periods exist and for the conclusion that their existence should be taken into account whenever developmental processes in infancy are discussed. Admittedly, so many regression periods were not expected meaning in his original study, and the findings came as a surprise. However, the data were firm and they have been replicated. End quote, he does note some of the confounding stressors that can generate regression, like behaviors, including parental divorce, going to daycare for the first time, holidays, switching between caretakers and when the baby is ill. He also wonders whether the differences in lifestyle and culture of infant care across countries might affect the results, and describes this as a topic for further study. Yet no further studies have ever been conducted. So the ploys published their only study on the timing of regression periods in 1992 and as far as I can tell, they didn’t publish any other papers after this point that establishes the timing of the regression periods more firmly by 2003 they’re saying we should stop debating whether the periods exist or not, with only 82 infants worth of evidence, much of which is not entirely aligned with the ploys ideas to back up that claim. Yet in the Wonder weeks book, The ploys confidently state. Date, all babies go through a leap in their mental development around the same age calculated from the due date, when the evidence we’ve seen does not really support this assertion. So, this is where we leave it for this episode. So, the upshot of all of this is what’s presented as settled science in the Wonder weeks book is really anything, but it seems likely that babies cry in some periods more than others over their first couple of years. But when those periods are is very much contested, or at least it was in the early 1990s nobody else has done research on this topic since then, although other researchers who commented for the New York Times article were divided on whether they had seen anecdotal evidence for leaps or whether they thought it was a load of rubbish. In our next episode on this topic, we’ll take a closer look at the evidence for what the ploys say is happening in the developmental leaps that are supposed to follow the fussy periods, as well as the implications of the advice that they offer to parents for what they can do to support those developmental leaps. If you would like to look at any of the 46 references I looked at for this episode, 37 or 80% of which are peer reviewed, which is a much higher proportion than you will find of the references in the Wonder weeks book. You can find them at your parentingmojo.com/wonderweekspartone.

Denise  01:01:19

I’m a your parenting Mojo fan, and I hope you enjoy the show as much as I do. If you found this episode especially enlightening or useful, you can donate to help Jen produce more content like this. Just go to the episode page that Jen mentioned. Thanks for listening.