One way to compare the effects of income vs. daycare on children’s behavior is to look at the effect sizes of both inputs. The effect size is how much of an effect the specific input variable has on the output variable.
It is a little tricky to do this though because you need to look at standardized measurements to compare across models/papers (basically so you’re comparing apples to apples). Another issue is that smaller studies tend to have inflated effect sizes and are less reliable. The third issue in comparing effect sizes is publication bias because positive results (aka results showing there is a difference) are easier to publish than null results (results showing there is no difference).
The best way to get around these issues is a meta-analysis. These look at (1) the strength of the evidence weighted by sample size, (2) the combined effect sizes across studies, and (3) test for publication bias in the literature (usually).
SES Meta-Analysis
For instance, when looking at the effect of income on antisocial behaviors, Piotrowska et al. (2015) identified 133 studies containing data suitable for effect size calculation, and 139 independent effect sizes were analysed (total N = 339 868). Their global meta-analysis showed that lower family socioeconomic status was associated with higher levels of antisocial behaviour.” The effect size was around 0.1. They did not find evidence of publication bias. ((income and antisocial behavior meta-analysis))
Peverill et al. similarly showed the following: “Among 26,715 participants aged 3–19 years, we observed small to moderate associations of low family income (g = 0.19), low Hollingshead index (g = 0.21), low subjective SES (g = 0.24), low parental education (g = 0.25), poverty status (g = 0.25), and receipt of public assistance (g = 0.32) with higher levels of childhood psychopathology. Moderator testing revealed that receipt of public assistance showed an especially strong association with psychopathology and that SES was more strongly related to externalizing than internalizing psychopathology.” They did not find evidence of publication bias. (SES and child psychopathology))
Letourneau et al. (2013) used a smaller number of papers in their meta-analysis, but still found a small significant effect low SES on children’s language skills, aggression, and internalizing behaviors/depression. ((SES and child development smaller meta-analysis))
Korous et al. did a systematic review (ie. gathered all the papers) of meta-analyses that looked at SES and cognition. Across these studies, “We identified 14 meta-analyses published between 1982 and 2019. These meta-analyses consistently reported positive associations of small to medium magnitude, indicating that SES is a meaningful contributor to the development of cognitive ability and achievement.” ((SES and cognition))
Fowler et al. (2009) looked at how neighborhood safety affected mental health in a meta-analysis. They found that exposure to community violence was associated with Externalizing behaviors, PTSD, and internalizing behaviors. ((crime and mental health meta-analysis).)
There are several more high quality studies not quoted, but in summary, there is a strong and consistent effect of income & SES more broadly on child development, especially cognition and aggression.
Daycare Meta-Analysis
Unfortunately, the data on daycare’s effects is MUCH weaker and sparser than the literature on income. There are no meta-analyses (at least to my knowledge) of the effects of daycare on child behavior.
There is one meta-analysis on how daycare effects cortisol levels. This meta-analysis only included 9 studies and the combined number of children was 303. This is extremely small (smaller than many single studies in the income literature), which is unfortunate. Their effect size was r=.18 and they did not check for publication bias using a funnel plot, like many of the income studies did. ((daycare and cortisol))
While this is the best evidence to date, it is not definitive nor is it particularly strong. Also, it is important to note that high cortisol does not always translate to behavior, though it has been shown to have ill effects in mice. However, chronically stressed individuals are actually shown to have a blunting of their cortisol response system (basically they cannot appropriately respond to a stressor because they are burnt out) in many human studies. I’d suggest looking at studies with the Trier social stress test if you’re curious about hormonal blunting.
Conclusions:
It’s too bad that the daycare literature as a whole is not as reliable or as well studied. I would say we definitively know that having a lower income is related to behavior problems and lowered cognition in children. We have some evidence that daycare may be stressful as measured by increased cortisol. In weighing these two thing, I think it’s important to remember the relative weight of the evidence. The daycare evidence is weak, so I wouldn’t base my decisions on it personally. It is way more important to do what is right for your family!! If one parent wants to stay home with a child and you can live a comfortable lifestyle doing that, then stay home with your child. If both parents want to work, then work.
If you want to hedge your bets with the daycare literature, having a grandparent or other relative care for your child may be best (assuming they are competent - I didn’t even get into all the parenting quality studies, but caregiving quality matters). That way you don’t need to spend money on childcare and the baby gets one on one attention. A nanny may be similar in terms of 1:1 attention, but comes with a steeper price tag than daycare. There isn’t a ton of literature on whether that’s worth it, but based on the “quality of daycare mattering” studies that show a low child to caregiver ratio is better, it may be another great option.
Edit: Since this has come up on other posts, I want to note that the income literature shows effects on child behavior across multiple reporters (parents, teachers, observers, clinicians). This is not true of the daycare literature, which fails to show a parent reporter effect.
It is a little tricky to do this though because you need to look at standardized measurements to compare across models/papers (basically so you’re comparing apples to apples). Another issue is that smaller studies tend to have inflated effect sizes and are less reliable. The third issue in comparing effect sizes is publication bias because positive results (aka results showing there is a difference) are easier to publish than null results (results showing there is no difference).
The best way to get around these issues is a meta-analysis. These look at (1) the strength of the evidence weighted by sample size, (2) the combined effect sizes across studies, and (3) test for publication bias in the literature (usually).
SES Meta-Analysis
For instance, when looking at the effect of income on antisocial behaviors, Piotrowska et al. (2015) identified 133 studies containing data suitable for effect size calculation, and 139 independent effect sizes were analysed (total N = 339 868). Their global meta-analysis showed that lower family socioeconomic status was associated with higher levels of antisocial behaviour.” The effect size was around 0.1. They did not find evidence of publication bias. ((income and antisocial behavior meta-analysis))
Peverill et al. similarly showed the following: “Among 26,715 participants aged 3–19 years, we observed small to moderate associations of low family income (g = 0.19), low Hollingshead index (g = 0.21), low subjective SES (g = 0.24), low parental education (g = 0.25), poverty status (g = 0.25), and receipt of public assistance (g = 0.32) with higher levels of childhood psychopathology. Moderator testing revealed that receipt of public assistance showed an especially strong association with psychopathology and that SES was more strongly related to externalizing than internalizing psychopathology.” They did not find evidence of publication bias. (SES and child psychopathology))
Letourneau et al. (2013) used a smaller number of papers in their meta-analysis, but still found a small significant effect low SES on children’s language skills, aggression, and internalizing behaviors/depression. ((SES and child development smaller meta-analysis))
Korous et al. did a systematic review (ie. gathered all the papers) of meta-analyses that looked at SES and cognition. Across these studies, “We identified 14 meta-analyses published between 1982 and 2019. These meta-analyses consistently reported positive associations of small to medium magnitude, indicating that SES is a meaningful contributor to the development of cognitive ability and achievement.” ((SES and cognition))
Fowler et al. (2009) looked at how neighborhood safety affected mental health in a meta-analysis. They found that exposure to community violence was associated with Externalizing behaviors, PTSD, and internalizing behaviors. ((crime and mental health meta-analysis).)
There are several more high quality studies not quoted, but in summary, there is a strong and consistent effect of income & SES more broadly on child development, especially cognition and aggression.
Daycare Meta-Analysis
Unfortunately, the data on daycare’s effects is MUCH weaker and sparser than the literature on income. There are no meta-analyses (at least to my knowledge) of the effects of daycare on child behavior.
There is one meta-analysis on how daycare effects cortisol levels. This meta-analysis only included 9 studies and the combined number of children was 303. This is extremely small (smaller than many single studies in the income literature), which is unfortunate. Their effect size was r=.18 and they did not check for publication bias using a funnel plot, like many of the income studies did. ((daycare and cortisol))
While this is the best evidence to date, it is not definitive nor is it particularly strong. Also, it is important to note that high cortisol does not always translate to behavior, though it has been shown to have ill effects in mice. However, chronically stressed individuals are actually shown to have a blunting of their cortisol response system (basically they cannot appropriately respond to a stressor because they are burnt out) in many human studies. I’d suggest looking at studies with the Trier social stress test if you’re curious about hormonal blunting.
Conclusions:
It’s too bad that the daycare literature as a whole is not as reliable or as well studied. I would say we definitively know that having a lower income is related to behavior problems and lowered cognition in children. We have some evidence that daycare may be stressful as measured by increased cortisol. In weighing these two thing, I think it’s important to remember the relative weight of the evidence. The daycare evidence is weak, so I wouldn’t base my decisions on it personally. It is way more important to do what is right for your family!! If one parent wants to stay home with a child and you can live a comfortable lifestyle doing that, then stay home with your child. If both parents want to work, then work.
If you want to hedge your bets with the daycare literature, having a grandparent or other relative care for your child may be best (assuming they are competent - I didn’t even get into all the parenting quality studies, but caregiving quality matters). That way you don’t need to spend money on childcare and the baby gets one on one attention. A nanny may be similar in terms of 1:1 attention, but comes with a steeper price tag than daycare. There isn’t a ton of literature on whether that’s worth it, but based on the “quality of daycare mattering” studies that show a low child to caregiver ratio is better, it may be another great option.
Edit: Since this has come up on other posts, I want to note that the income literature shows effects on child behavior across multiple reporters (parents, teachers, observers, clinicians). This is not true of the daycare literature, which fails to show a parent reporter effect.