Largest quantitative synthesis to date reveals what predicts human behavior and how to change it

Pandemics, global warming, and rampant gun violence are all clear lessons in the need to move large groups of people to change their behavior. When a crisis hits, researchers, policymakers, health officials, and community leaders have to know how best to encourage people to change en masse and quickly. Each crisis leads to rehashing the same strategies, even those that have not worked in the past, due to the lack of definitive science of what interventions work across the board combined with well intended but erroneous intuitions. 

To produce evidence on what determines and changes behavior, Professor Dolores Albarracín and her colleagues from the Social Action Lab at the University of Pennsylvania undertook a review of all of the available meta-analyses — a synthesis of the results from multiple studies — to determine what interventions work best when trying to change people’s behavior. What results is a new classification of predictors of behavior and a novel, empirical model for understanding the different ways to change behavior by targeting either individual or social/structural factors.

A paper published today in Nature Reviews Psychology reports that the strategies that people assume will work — like giving people accurate information or trying to change their beliefs — do not. At the same time, others like providing social support and tapping into individuals’ behavioral skills and habits as well as removing practical obstacles to behavior (e.g., providing health insurance to encourage health behaviors) can have more sizable impacts.

“Interventions targeting knowledge, general attitudes, beliefs, administrative and legal sanctions, and trustworthiness — these factors researchers and policymakers put so much weight on — are actually quite ineffective,” says Albarracín. “They have negligible effects.”

Unfortunately, many policies and reports are centered around goals like increasing vaccine confidence (an attitude) or curbing misinformation. Policymakers must look at evidence to determine what factors will return the investment, Albarracín says.

Co-author Javier Granados Samayoa, the Vartan Gregorian Postdoctoral Fellow at the Annenberg Public Policy Center, has noticed researchers’ tendency to target knowledge and beliefs when creating behavior change interventions.

“There’s something about it that seems so straightforward — you think x and therefore you do y. But what the literature suggests is that there are a lot of intervening processes that have to line up for people to actually act on those beliefs, so it’s not that easy,” he says.

Targeting Human Behavior

To change behaviors, intervention researchers focus on the two types of determinants of human behavior: individual and social-structural. Individual determinants encompass personal attributes, beliefs, and experiences unique to each person, while social-structural determinants encompass broader societal influences on people, like laws, norms, socioeconomic status, social support, and institutional policies.

The researchers’ review explored meta-analyses of experiments in which specific social-structural determinants or specific individual determinants were tested for their ability to change behavior. For example, a study might test how learning more about vaccination might encourage vaccination (knowledge) or how reductions in health insurance copayment charges might encourage medication adherence (access).

These meta-analyses encompassed eight individual and eight social-structural determinants — part of the original classification made by the authors.

The results from the research are presented in the following three figures, which pertain to a. all behaviors analyzed, b. only health behaviors, and c. only environmental behaviors.

The figures present interventions with individual targets on the left, and interventions with social/structural targets on the right. For each determinant, the figures show whether the effects has been shown to be negligible, small, medium or large.

Individual Determinants 

The analyses researchers conducted showed that what are often assumed to be the most effective individual determinants to target with interventions were not the most effective. Knowledge (like educating people about the pros of vaccination), general attitudes (like implicit bias training), and general skills (like programs designed to encourage people to stop smoking) had negligible effects on behavior. 

What was effective at an individual level was targeting habits (helping people to stop or start a behavior), behavioral attitudes (having people associate certain behaviors as being “good” or “bad”), and behavioral skills (having people learn how to remove obstacles to their behavior).

Determinant	Definition	Example measures	Example interventions
Knowledge	Collection of facts about an object or behavior, which can include information about the properties and consequences of a particular object or event, such as a virus or pollution; knowledge links an object or behavior to an attribute or event with absolute certainty	Measure of literacy: “Contact with a dirty toilet is a common cause of venereal disease or sexually transmitted disease” (participant responds ‘true’ or ‘false)	Health education Didactic instruction about climate change in schools
General Skills	Cognitive or overt routines that enable individuals to carry out various specific behavious; they involve broad capacities such as controlling attention during tasks and being able to inhibit temptations when behaviors require high levels of self-control	Self-report measures of self-control, which include statements about a person’s ability to make a plan or avoid temptations	Behavioral change to programs emphasizing the need to train general skills that might help individuals control undesirable behaviors
General attitudes	Evaluations of objects, persons and events; for example, prejudice is a negative judgement of a group as the attitude object, and an attitude towards cars is a positive or negative evaluation of cars as the attitude objects this type of attitude is often termed ‘attitude towards the target’	Likert-scale measure of attitudes towards enviornmental protections: “Humans are severely abusing the environment” Implicit attitude test concerning alcohol	Mass-media health-promotion campaigns about a behavior  Interventions aimed at weakening associations by instilling goals and threat
Beliefs	Subjective assignments of probability that an object or behavior has a given attribute or outcome	Self-report measure of conspiracy beliefs: “To what extent do you think the virus is part of a biological warfare program?”	Messages that explicitly introduce expectations about a behavior  Growth mindset interventions in academic settings
Emotions	Visceral feelings (for example, happiness or fear) associated with a particular object, person or event; experiencing fear of climate change or disgust about a particular group of individuals are examples of emotion	Likert-scale measure of emotions towards COVID-10: “I feel fearful about COVID-19”	Emotional appeals that sensitize audiences to risks and include discussion of the threat posed by a problem or the audience’s susceptibility to it
Behavioral skills	Routines that enable people to execute a target behavior, often reflected in higher levels of perceived control or efficacy concerning the behavior	Self-report measure of behavioral control and confidence to perform or abstain from a behavior: “If I wanted to, it would be easy for me to exercise for at least twenty minutes, three times a week for the next fortnight”	Practicing and receiving feedback on the behavior and performing homework related to the behavior  Asking individuals to formulate implementation intentions
Behavioral attitudes	Evaluations of a behavior as good or bad; for example, whereas an attitude towards cars is a general attitude, an attitude towards driving a car for transportation is a behavioral attitude; this type of attitude is often referred to as ‘attitude towards the behavior’	Semantic differential measures of attitudes linking recycling to adjectives such as good or bad: “Recycling household waste for me is something…” (participant selects from five-point response scale anchored by adjectives ‘good’ or ‘bad’)	Mode of questioning designed to uncover and reduce attitudinal ambivalence towards a particular behavior
Habits	Behavioral routines that have acquired features of automaticity, meaning that they occur efficiently, without awareness, or continue even without intention and after they are no longer adaptive	Measure of handwashing habit: “Washing my hands would require effort not to do”	Training to stop a behavior when faced with temptations Introducing environmental regularity to promote habit formation Distracting oneself from behavioral cues

 

 

 

Social-Structural Determinants

The researchers also found that what are often assumed to be the most effective social-structural persuasive strategies were not. Legal and administrative sanctions (like requiring people to get vaccinated) and interventions to increase trustworthiness — justice or fairness within an organization or government entity — (like providing channels for voters to voice their concerns) had negligible effects on behavior. 

Norms and forms to monitor and incentivize behavior had some effects, albeit small. What was most effective was focusing on targeting access (like providing flu vaccinations at work) or social support (facilitating groups of people who help one another to meet their physical activity goals).

Determinant	Definition	Sample measures	Sample interventions
Legal and administrative sanctions	Legal and administrative instruments to prescribe, ban or sanction a behavior	State and country records of laws coded through a policy review	Banning smoking in public establishments Mandating vaccination Mandating sick pay Taxing pollution
Trustworthiness	Justice or fairness within an organization or government entity, which leads constituents to follow recommendations	Self-report measure of procedural justice: “How fair were the procedures used to handle the problem?”	Providing channels for Latinx voters to voice their concerns Community-oriented policing that fosters non-enforcement interactions
Injunctive norms	Perceptions of the degree to which others support a person’s behavior	Self-report measure of injunctive norms: “People who are important to me think I should use condoms”	Messages that communicate that others approve of condom use Posting signs stating that taking the stairs is a good way to get some exercise
Monitors and reminders	Physical or digital instrument to track behavioral performance and remind users of the need to execute a behavior	Self-reported use of pill boxes, diaries, and planners	Clinical reminder system for promoting preventive care Digital watches and phone apps that promote physical activity
Descriptive norms	Frequency of a behavior is a particular population	Self-reported perception of what others do: “Most residents would vaccinate their child against COVID-19”	Comparative feedback such as a chart tracking one’s energy consumption in relation too one’s neighbors Using role models to promote a target behavior Posting signs stating that most people use the stairs
Material incentives	Providing financial or non-financial rewards in exchange for a behavior	Introduction of state lottery for vaccinated residents as a rewards for vaccinations	Paying people $24 to receive the COVID-19 vaccine
Social support	Informational, instrumental or financial help to facilitate a particular behavior	Self-reported lists of individuals who can perform instrumental, informational and emotional support functions	Leveraging family or ad-hoc groups to assist individuals to meet their physical activity goals  Groups of Latina mothers led by ‘promotoras‘ who support and accompany each other during health-promoting activities
Access	Material or logistic resources to facilitate the performance of a behavior	Census demographics and self-report of health insurance Self-reported health insurance	Reducing co-payments for medication Providing health insurance Providing basic income

 

Future Use

Granados Samayoa says that knowing which behavior change interventions work at which levels will be especially crucial in the face of growing health and environmental challenges. 

“When faced with massive problems like climate change, policy makers and other leaders have this desire to do something to change people’s behavior for the better,” says Samayoa. “Our study provides valuable insights. Our research can inform future interventions and create programs that are actually effective, not just what people assume is effective. 

Albarracín is glad policymakers will have this resource now. 

“Before this study, analyses of behavior change efforts were limited to one domain, whether that was environmental science or public health. By looking at research across domains, we now have a clearer picture of how to encourage behavior change and make a difference in people’s lives,” she says.

“Our research provides a map for what might be effective even for behaviors nobody has studied. Just like masking because a critical behavior during the pandemic but we had no research on masking, a broad empirical study of intervention efficacy can guide future efforts for an array of behaviors we have not directly studied but that need to be promoted during a crisis.”

“Determinants of Behaviour and Their Efficacy As Targets of Behavioural Change Interventions” was published in Nature Reviews Psychology and authored by Dolores Albarracín, Bita Fayaz-Farkhad, and Javier Granados Samayoa. The research was funded by National Institutes of Health (NIH) grants R01MH132415, R01 AI147487, DP1 DA048570, R01 MH114847, and NSF 2031972 to Dolores Albarracín, and by the Annenberg Foundation Endowment to the Division of Communication Science at the Annenberg Public Policy Center.

About the Social Action Lab

The Social Action Lab is a group of experts and trainees in psychology, communication, and economics who seek to understand the fundamentals of social behavior and apply this knowledge to the solution of social and health problems. The lab is led by Dolores Albarracín, a Penn Integrates Knowledge Professor and Director of the Division of Communication Science in Annenberg Public Policy Center. She holds appointments in the Annenberg School for Communication, the Department of Psychology in the School of Arts & Sciences, the School of Nursing, and the Wharton School. The other two authors are Bita Fayaz-Farkhad, who is Assistant Research Professor in the Annenberg School for Communication, and Javier Granados Samayoa, the Vartan Gregorian Postdoctoral Fellow of the Annenberg Public Policy Center. 

---

---