Is It All Over by Age Five?

The marshmallow delayed gratification test faces its greatest challenge yet.

Posted Jun 02, 2018

Dora Zett/Shutterstock
Source: Dora Zett/Shutterstock

Delayed gratification displayed by your child is one of the most sought-after prizes of parents with aspirations for their children. Whether or not you are able to control the urge at an early age to immediately eat a single marshmallow placed in front of you and wait a little longer to enjoy the second marshmallow, is a well-known psychological test.

The marshmallow test, is seen as the ultimate psychological litmus test, demonstrating a young child’s ability to exhibit delayed gratification. The reasoning behind delayed gratification and its longer-term benefits became part of many parents’ belief-set following the seminal study on child delayed gratification carried out in the 60’s, with the researchers going on to show in a follow-up study in the 90’s, that those children best able to resist temptation enjoying the fruits of their self-control with a second marshmallow, outperformed their less-patient co-respondents in many later aspects of cognition and behaviour1

Unfortunately, as recipients of these insights through books, social media, folklore and the advice of child psychologists, we rarely delve deeper to understand the nuances that influence the results of such research studies. Sample size, effect size and degree of statistical significance are all able to significantly influence the results. For example, only 50 respondents were used in the original research. For psychology research, this represents a relatively small sample size. Was the study sufficiently powered? How reliable were the results? Was the effect size such that the results could be reliably reproduced? Moreover, the sample of children was recruited from the siblings of the staff at Harvard. Hardly a representative sample, although one could argue that at least they reflected a relatively narrow range of social economic backgrounds.

Fast forward to earlier this week and a new study2 has been carried out that appears to indicate that the main claim of this longitudinal study, namely that the ability to apply delayed gratification in life correlates with substantial benefits, appears to be under a cloud. The international press is having a field day. In essence, the new study appears to show that the effect of delayed gratification has been overstated and that delayed gratification displayed by a young child may not be such a valuable trait for indicating future performance in later life as first thought.

However, it is important to understand how the covariates in both the original and replicated studies (those factors that also vary within the sample population) have been accounted for. This appears to have been carefully calculated in the second but less so the original study. Importantly, this latest study makes two significant claims. The first that their findings (made with a far larger sample size and arguably more reliable) showed that the strength of correlation between the ability to wait for the second marshmallow and achievement at aged 15 was only half that of the original study. This could well be reflective of the first study's smaller sample size and is perhaps not surprising in itself. However, more interesting are the findings that socioeconomic class defined by home environment, family background, and early cognitive ability was observed to reduce this correlation by a further two thirds.

What should one make of this? One could argue that rising inequality, a declining US education system for the disadvantaged or simply being born to smart and successful parents is now more a driver of future achievement and success than ever. Taking a philosophical stance one might also postulate that this study indicates that achieving the American dream is not just about what you are willing to forgo, but who your parents are and where you went to school. 


1. Mischel, W. (2014). The marshmallow test: understanding self-control and how to master it: Random House.

2. Watts, T. W., Duncan, G. J., & Quan, H. Revisiting the Marshmallow Test: A Conceptual Replication Investigating Links Between Early Delay of Gratification and Later Outcomes. Psychological science, 0956797618761661.

More Posts