By Jack Auty 20/12/2016

New year’s resolutions- the Babylonians did it, the Romans did it and now we do it; with one survey suggesting that this ancient tradition is still popular reporting that over 40%  of Americans take part. Evidence can also be found in trends of Google searches, there is an annual cycle of Google searchers for the word “diet” nearly doubling between the beginning of December to the beginning of January. But do news years resolutions work?

Fig 1. The annual cycle of failed aspirations.

One group of University of Scranton researchers would argue that resolutions definitely do work (Norcross et al. 2002). They recruited 434 people into a study in December 2001. These people were asked are you planning on making a new year’s resolution this year? 159 said yes and the rest said no. But of the people who said no; 123 said that they had a problem behaviour that they were seriously thinking about changing.  The researchers then followed these people up by phone for the next 6 months. What they found is that of those who made a new year’s resolution 46% self reported that they have succeeded in changing behaviour. However, the success rate of those who wanted to change but didn’t want to make a resolution was only 4%! Now there are a myriad of limitations to this study for example the comparison shouldn’t be between people who want to change and haven’t set a date to people who set new year’s resolution; the comparison should be between resolutioners and people who commit to changing a behaviour at a specific date throughout the year. But then again who cares? Certainly not the funding bodies, hmmm should I fund research into curing cancer or whether or not new year’s resolution work?
Now you might be thinking that one reason new year’s resolutions might work is because we do it as a group and we are supported by peers? Buuuut maybe not. A group of University of Michigan researchers wanted to know if a peer network or a buddy system would help older adults achieve their goal of walking regularly (Kullgren et al. 2014). They randomly allocated 92 participants into a control group, a peer network group where they were to encourage each other to walk and arrange times through an online message board, or a financial reward group where the participants who achieved their goal would have the chance to win $200. After 24 weeks of studying these participants the researchers found that the control group achieved their goals the best with about 35% of people achieving their walking goal. The financial incentive group performed about the same, but the peer support group was significantly worse with only 18% achieving their walking goal. What might be causing this is that if any of the peer support group didn’t want to or couldn’t do the walk that day the rest of the group was more likely to cancel. The group became as strong as the weakest link (fig 2.). This suggests that doing resolutions as a group may not be the best idea.

Fig. 2. Obscure reference to English game show. 

OK but what are some good suggestions for resolutions? Well most of the research isn’t to do with resolutions but is about goals in general. This research is still not that fundable, hence the small sample sizes, and probably shouldn’t be relied on heavily. But I’ve committed to writing a blog about it, so here you go:

  1. Make the resolution measurable and measure it regularly. One study coming out of the University of North Carolina recruited 47 people (small sample size) who wished to lose weight (Steinberg et al. 2015). They randomly allocated the people into two groups, one group was asked to record weights everyday and the other group was told to weigh themselves when they felt like it. The group that recorded their weights everyday lost 6.1kg more than those who didn’t (this grouped average a once-a-week weighing schedule).
  2. Make the resolution specific? Possibly not. One study from the University of Sydney analysed records of pharmacist-patient interactions (n=~1700) (Smith et al. 2013). They extracted goals set by the pharmacist and patient, and rated them as specific, moderately specific and not specific based on set criteria. They found that success rates were relatively equal across the specificity groups. So maybe specificity isn’t that important in goal achievement? When you look in detail into this study you see a trend where specific goals weren’t achieved unless the goal was method based. So be specific in your method not in the achievement….. might be the take-away from this paper. E.g. Good resolution: go to the gym three times a week. Bad resolution: lose 10.8kgs. Both resolutions are specific but one is method based.
  3. Make the resolution achievable? This too may be a motivational myth. One study from the University of Birmingham on 126 participants (small sample) looked into whether failure of goal affected your ability to achieve a second goal (Healy et al. 2015). They did this by rigging up exercycles with dodgy speedometers so they could induce failure or success in the goal of biking a set distance in a set time. They then set the participants another biking goal and looked to see if the rate of achievement of the goal differed between the groups that had been made to succeed or fail at the first task. They found absolutely no effect of induced failure or success. So go ahead and reach for the stars (#BeatConorMcGregor)?

This might be a good time to mention that science, and in particular psychology, is going through a repeatability crisis. A large number of studies have been repeated and the results have not been the same. We are calmly dealing with the situation (fig. 3.). One group attempted to repeat 98 papers, 97% of which found a significant result (Open science collaboration 2015). When these were repeated, often with the help of the original authors, only 36% had a significant result. So everything I just said about making a good or bad resolution comes with a grain of salt. A grain of salt 50km wide, similar to those shiny patches of salt found on the dwarf planet Ceres

My new year’s resolution is to write one blog a month on science and not science. This resolution doesn’t involve a peer support network, is specific (in a non-specific way), and is measurable. So I’ve got all the potentially unrepeatable science backing me!


Fig. 3. Calmly dealing with the induction fallacy. 


Norcross JC, Mrykalo MS, Blagys MD (2002) Auld lang syne: success predictors, change processes, and self-reported outcomes of New Year’s resolvers and nonresolvers. Journal of clinical psychology 58 (4):397-405

Kullgren JT, Harkins KA, Bellamy SL, Gonzales A, Tao Y, Zhu J, Volpp KG, Asch DA, Heisler M, Karlawish J (2014) A Mixed Methods Randomized Controlled Trial of Financial Incentives and Peer Networks to Promote Walking among Older Adults. Health education & behavior : the official publication of the Society for Public Health Education 41 (1 0):43S-50S. doi:10.1177/1090198114540464

Steinberg DM, Bennett GG, Askew S, Tate DF (2015) Weighing every day matters: daily weighing improves weight loss and adoption of weight control behaviors. Journal of the Academy of Nutrition and Dietetics 115 (4):511-518. doi:10.1016/j.jand.2014.12.011

Smith L, Alles C, Lemay K, Reddel H, Saini B, Bosnic-Anticevich S, Emmerton L, Stewart K, Burton D, Krass I, Armour C (2013) The contribution of goal specificity to goal achievement in collaborative goal setting for the management of asthma. Research in social & administrative pharmacy : RSAP 9 (6):918-929. doi:10.1016/j.sapharm.2013.02.002

Healy LC, Ntoumanis N, Stewart BD, Duda JL (2015) Predicting subsequent task performance from goal motivation and goal failure. Frontiers in Psychology 6:926. doi:10.3389/fpsyg.2015.00926

Open science collaboration. Estimating the reproducibility of psychological science  (2015). Science (New York, NY) 349 (6251)