What Characterises an Effective Mindset Intervention in Enhancing Students’ Learning? A Systematic Literature Review

Abstract

In recent years, increasing attention has been paid to interventions designed to enhance individuals’ sustainable development in learning by priming a growth mindset. The current study systematically explored the characteristics of message transformation in growth mindset interventions from the perspective of teaching and learning. According to a three-phase literature search (database, prominent researchers, and backtracking references), thirty-eight empirical studies investigating the efficacy of mindset interventions for adolescents of school age constitute the sample for the current literature review. The results indicate that a supportive but not-completely-saturated learning environment paves the way to implementing a mindset intervention. The three pedagogical characteristics that ensure successful interventions are: (1) Mutual interaction among the person, the context, and the theory to generate the message; (2) Iterative processes to ensure the message is delivered; and (3) a persuasive yet stealthy approach to facilitating its internalization. The findings inspire educators to design effective mindset interventions to enhance students’ learning. Theoretical and practical implications, limitations, and future directions are discussed.

1. Introduction

Due to the launch of the Unites Nations ‘Sustainable Development Goals (SDGs)’ in 2015, students’ academic and social performance towards sustainable development at the educational level has increasingly attracted attention. With increasing studies investigating educational processes to improve learning efficacy [1,2,3], mindset theory has proved in both educational and psychological settings as a tool for improving learning patterns through strengthening self-affirmation. Mindset, also known as implicit theory, is the belief people hold about the malleability of their human attributes [4]. Individuals with a growth mindset believe these attributes are malleable with effort, whereas for those with a fixed mindset they are immutable. In academia, for example, an academic mindset has proved to stimulate students’ learning experiences including but not limited to academic perseverance [5], enjoyment, and engagement [6]. Consequently, it has been found to give academic grades a significant boost [7,8,9]. Accordingly, more and more mindset interventions have come to light in the last two decades, exploring effective approaches to cultivating a growth mindset among individuals to enhance achievement [6,10,11].

It is worth noting that the efficacy of mindset thinking has been doubted given the very small even null effects found. For example, a meta-analysis (k = 273, N = 365,915) yielded weak overall effects of mindset and mindset intervention on academic achievement [12], which raised doubts as to whether a growth mindset did help to improve adaptive behavioral patterns. However, the poor results could have been attributable to the lack of scientificness and efficacy in how teachers transferred growth-mindset messages to students [13]. In other words, growth-mindset pedagogy was lacking [14]. Interestingly, benefits from mindset interventions have been broadly supported for marginalized individuals such as those with low socioeconomic status or at risk academically [12]. Furthermore, research involving a nationally representative sample in the US (k = 65, N = 12,490) has shown that mindset intervention increases student enrolment overall for advanced mathematics courses [15].

Mindset interventions differ in terms of design. The accumulative time of the sessions ranges from several minutes [16] to couple of hours [17], and the number of operating sessions varies from one [18,19,20] to dozens [21]. Most sessions include activities such as reading [18,22,23], discussion [24,25], and writing exercises [26,27,28]. Intervention proceeds on-site or online, or even in combination [29]. Brainology, an online program with animation-based interaction to encourage a growth mindset [30], is one example of an intervention [21,31]. Given the availability of such computer-based approaches, mindset interventions have ranged from small-scale field research to scalable investigations [10]. Measures such as complementing the experiment group with a control group that does non-mindset exercises (e.g., coping skills, see [32]; health issues, see [33]) or has no treatment [7,34,35] have been introduced to verify the efficacy of mindset intervention. Other motivational factors, such as sense of purpose [10], social belonging [36], and social norms [23], have been combined with growth-mindset messages to see if intervention efficacy would be additive. However, several empirical investigations discouraged such a prediction [10,36].

In addition to the various empirical investigations in either laboratory or real-world settings, there have been a limited number of theoretical studies focusing on the characteristics of mindset interventions [5,37] or their effectiveness [12]. Farrington et al. [5], for example, emphasize the interaction between individuals and the social-cultural context through recursive processes in which people and context are mutually reinforced or discouraged in a repeating feedback loop. Hence, without removing objective adversities, how individuals interpret adversity could determine whether it hurts the outcome [37]. The implication is that the mindset is a product of interaction between individuals and the surrounding environment rather than a set of predetermined characteristics [5].

Notably, more review and commentary studies have focused on the traits of general social-psychological interventions rather than on mindset intervention [38,39,40]. Some of the theoretical underpinnings in these studies, related to social-psychological attributes, for example, could be applied to the current review, given the theoretical commonalities across different perceptual points that fall within the scale of social-psychological intervention. Cohen et al. [38] and Wilson et al. [39], emphasize the importance of the surrounding environment. It is also acknowledged that underlying theoretical knowledge of its mechanisms facilitates the development of well-designed interventions and their timely revision [40].

In sum, in the educational context of SDGs, most existing reviews of or commentaries on interventions in educational psychology concentrate on social-psychological interventions in general, and very few studies specifically focus on the mindset. Second, most of the studies on mindset-relevant intervention were conducted by the same group of researchers from the United States, and to our knowledge, the latest review with discussions of characteristics of intervention goes back ten years. Third, none of the existing reviews incorporates a scientific and systematic search of the relevant literature to explore the characteristics of growth-message transmission in effective mindset intervention. Fourth, none of the existing research adequately explores person-context-theory interaction in conducting mindset interventions. Thus, there is a clear need for a scientific and systematic review of empirical studies conducted by non-US researchers and with a focus on transformation mechanisms. On the one hand, “broader evidence …… would strengthen the claims from these authors” [5] (p. 31); on the other hand, up-to-date data will facilitate the understanding of the efficacy of mindset intervention. The aim of the current systematic review is to synthesize existing empirical studies published in the last decade, thereby exploring the following research questions:

• RQ1: How effective are mindset interventions in enhancing students’ learning?
• RQ2: What characterises effective mindset interventions in enhancing students’ learning?

2. Materials and Methods

2.1. Phase 1 of the Literature Search–Databases

As Figure 1 demonstrates, we first searched for terms combined with mindset (implicit theory of intelligence, incremental theory, entity theory, self-theory, implicit person theory, implicit person scale) and intervention (experimental study, experimental design) in three databases: Web of Science, PsycINFO, ERIC (Education Resources Information Center). We searched in April 2021. The current review focuses on studies published between 2012 and 2021. The first-phase search yielded a total of 1484 titles, of which 1451 remained after de-duplication.

Inclusion and exclusion criteria. To select empirical studies for the current systematic review, we screened the titles and abstracts of all 1451 articles following a seven-item set of inclusion and exclusion criteria (see Table 1).

In line with the criteria, we excluded 1307 of the titles, leaving us with 144 articles. Of these, we excluded five because the full text was not available online and the we received no response to request for it from authors. We screened the full texts of the remaining 139 articles for eligibility, applying the same inclusion and exclusion criteria: 36 met all the inclusion criteria and 88 were deemed ineligible.

2.2. Phase 2 of the Literature Search–Prominent Researchers

To locate the relevant literature more precisely as the evidence base, aside from the database search, we studied the curricula vitae of prominent researchers in terms of conducting social-psychological interventions with a focus on mindset, including Carol Dweck, David S. Yeager, Gregory M. Walton, and Geoffrey L. Cohen. We applied the same inclusion and exclusion criteria to their studies published between 2012 and 2021. During this second round, we added two more articles [36,41] to our literature base.

Table 1. Inclusion criteria in the literature search.

Criterion	Inclusion	Note
Participants	Students	Adolescents of school age were included (e.g., [27]).
Methodology	Quantitative studies	Qualitative investigations were excluded.
Language	English	The author is fluent in English, and international readers can check the selected studies.
Research aim	Effective mindset intervention in enhancing students’ learning	Students’ learning, following the goal of sustainable development, included both academic performances (e.g., academic grades, see [15]) and reported social behaviours (e.g., challenge-seeking behaviour, see [42]). If the intervention efficacy was merely rooted in the change-of-mindset scale before and after the intervention (e.g., [43]), it was excluded. Effectiveness is reliant on significance and effect size. For example, the study of Donohoe et al. [44] was excluded because it discouraged the prediction that mindset intervention boosted students’ learning behaviours. In Burnette et al.’s study [19], mindset intervention did not significantly enhance learning directly, but it did so indirectly and was thus included.
Data overlap	Sample independent of the samples in other included studies	In case of data overlap, our preference was the published study with the larger sample or the more recent work (e.g., we included [45], but excluded [46]), or with more details (e.g., we included [27], but excluded [47]).
Mindset scale	Dweck’s implicit theory of intelligence (or ability) or personality.	Studies measuring mindset in general, namely the implicit theory of intelligence or ability, were included, as were those measuring domain-specific mindset (e.g., implicit theory of computer science ability, see [33]; chemistry intelligence, see [48]), and accurately specifying an implicit theory of personality [22,32]. However, the mindset with other socially relevant attributes (e.g., implicit theory of morality, see [49]; malleability of happiness, see [50]; malleability of creativity, see [51]) was excluded. Other mindsets (e.g., deliberate & implemental) differed from Dweck’s concept, and hence were also excluded [52].
Research design	At least two conditions: mindset group and control group.	Studies containing multiple groups other than mindset and control groups (such as social norms, or a group combining mindset and social norms, see [23]) were included. However, we excluded investigations limited to the control condition and the combined condition (e.g., reading strategies and mindset, see [53]; folklore and mindset, see [54]), those with no independent mindset condition, and those including only growth-mindset intervention without a control group (e.g., growth vs. fixed mindset, see [55,56]). We also excluded studies in which treatments for the control and experiment group were not distinguished (implicit theory of athletic ability vs. implicit theory of personality, see [57]). If the intervention group covered multiple psychological traits rather than mindset (e.g., growth mindset paired with gratitude and value affirmation, see [58]), we excluded the study because it was hard to evaluate the sole effect of a mindset intervention.

Table 1. Inclusion criteria in the literature search.

Criterion	Inclusion	Note
Participants	Students	Adolescents of school age were included (e.g., [27]).
Methodology	Quantitative studies	Qualitative investigations were excluded.
Language	English	The author is fluent in English, and international readers can check the selected studies.
Research aim	Effective mindset intervention in enhancing students’ learning	Students’ learning, following the goal of sustainable development, included both academic performances (e.g., academic grades, see [15]) and reported social behaviours (e.g., challenge-seeking behaviour, see [42]). If the intervention efficacy was merely rooted in the change-of-mindset scale before and after the intervention (e.g., [43]), it was excluded. Effectiveness is reliant on significance and effect size. For example, the study of Donohoe et al. [44] was excluded because it discouraged the prediction that mindset intervention boosted students’ learning behaviours. In Burnette et al.’s study [19], mindset intervention did not significantly enhance learning directly, but it did so indirectly and was thus included.
Data overlap	Sample independent of the samples in other included studies	In case of data overlap, our preference was the published study with the larger sample or the more recent work (e.g., we included [45], but excluded [46]), or with more details (e.g., we included [27], but excluded [47]).
Mindset scale	Dweck’s implicit theory of intelligence (or ability) or personality.	Studies measuring mindset in general, namely the implicit theory of intelligence or ability, were included, as were those measuring domain-specific mindset (e.g., implicit theory of computer science ability, see [33]; chemistry intelligence, see [48]), and accurately specifying an implicit theory of personality [22,32]. However, the mindset with other socially relevant attributes (e.g., implicit theory of morality, see [49]; malleability of happiness, see [50]; malleability of creativity, see [51]) was excluded. Other mindsets (e.g., deliberate & implemental) differed from Dweck’s concept, and hence were also excluded [52].
Research design	At least two conditions: mindset group and control group.	Studies containing multiple groups other than mindset and control groups (such as social norms, or a group combining mindset and social norms, see [23]) were included. However, we excluded investigations limited to the control condition and the combined condition (e.g., reading strategies and mindset, see [53]; folklore and mindset, see [54]), those with no independent mindset condition, and those including only growth-mindset intervention without a control group (e.g., growth vs. fixed mindset, see [55,56]). We also excluded studies in which treatments for the control and experiment group were not distinguished (implicit theory of athletic ability vs. implicit theory of personality, see [57]). If the intervention group covered multiple psychological traits rather than mindset (e.g., growth mindset paired with gratitude and value affirmation, see [58]), we excluded the study because it was hard to evaluate the sole effect of a mindset intervention.

2.3. Phase 3 of the Literature Search–Backtracking Reference

Finally, we checked the reference lists of the 38 articles from phases 1 and 2 looking for more potentially relevant studies. In line with the above-mentioned inclusion and exclusion criteria, no more were added. Thus, we ended up with 38 articles (see

Table 2. Descriptive information about the selected mindset-intervention studies.

Study	Sample	Country	Domain	Intervention Design (Pretest, Posttest, and Follow-Up Excluded)
				Group	Session	Location	Approach
Broomhead et al. (2012)	High-school students (N = 155)	USA	Academic	1. Positive mindset 2. Control (No treatment)	Single session: 40 min	On-site	Self-persuasive speaking Breathing exercises
Yeager et al. (2013a)	S1 & 2: 7th graders (N = 88) S3: High-school student (N = 76)	USA	Motivational Academic	1. Wise criticism 2. Control 1 (Placebo) 3. Control 2 (Puzzles)	Single sessions: 15–20 min	On-site Online	Wise-criticism-based feedback
Yeager et al. (2013b)	9th and 10th graders (N = 230)	USA	Social Moral Emotional	1. Growth mindset 2. Control 1 (Coping skills) 3. Control 2 (No treatment)	6 sessions (Within 3 weeks): 50 min each	On-site	Lesson-based introduction Group discussion Reflective summarizing
Sriram (2014)	College students (N = 105)	USA	Motivational Academic	1. Growth mindset 2. Control (Study skills)	4 sessions (1 week apart): 15 min each	Online	Video watching Testimonies Reflective summarizing
Paunesku et al. (2015)	High-school student (N = 1594)	USA	Academic	1. Growth mindset 2. Sense of purpose 3. Combined 4. Control (Brain)	2 sessions (2 weeks apart): 45 min each	Online	Text reading Reflective summarizing Letter writing
Andersen et al. (2016)	2nd graders (N = 1587)	Denmark	Academic	1. Growth mindset 2. Control (No treatment)	10 sessions: NA	Online	Video watching Text reading
Hu et al. (2016)	Undergraduates (N = 152)	China	Self-perceptional	1. Growth mindset 2. Emotional management 3. Control (Generic reasoning)	Single session: 3–5 min	On-site	Text reading
Yeager et al. (2016a)	9th graders (N = 3676)	USA & Canada	Academic Self-perceptional Motivational	1. Growth mindset 2. Control (School transition)	2 sessions (1–4 weeks apart): NA	Online	Text reading Testimonies Reflective summarizing Letter writing
Yeager et al. (2016b)	College student (S2: N = 7335; S3: N = 1592)	USA	Academic	1. Growth mindset 2. Social belonging 3. Combination 4. Control (Adjustment)	Single session: 30 min	Online	Text reading Testimonies Reflective writing
Yeager et al. (2016c)	High-school students (N = 265)	USA	Self-perceptional Academic	1. Growth mindset 2. Control (Environment adjusting)	Single session: 25 min	On-site	Text reading Testimonies Letter writing
Schmidt et al. (2017)	7th and 9th graders (N = 726)	USA	Self-perceptional	1. Growth mindset 2. Control (Science)	6 sessions (Each/week): 50 min each	On-site Online	Brainology
Orosz et al. (2017)	10th graders (N = 55)	Hungary	Motivational	1. Growth mindset 2. Control (Bystander effect)	5 sessions (Each/week): 45 min each	On-site	Lesson-based introduction Video watching Group discussion Reflective summarizing Letter writing
Bråten et al. (2017)	Undergraduates (N = 529)	USA & Norway	Self-perceptional	1. Growth mindset 2. Fixed mindset 3. Control (Task requirement)	Single session: <15 min	On-site	Text reading
O’Brien et al. (2017)	6th to 9th graders (N = 196)	UK	Motivational	1. Growth mindset 2. Control (No treatment)	4 sessions (Each per day): NA	On-site	Text reading Lesson-based introduction Reflective summarizing
Burnette et al. (2018)	female 10th graders (N = 222)	USA	Academic Motivational	1. Growth mindset 2. Control (Sexual health)	Single session: 45 min	On-site Online	Lesson-based introduction Testimonies Letter writing
Broda et al. (2018)	Quasi college students (N = 7686)	USA	Academic	1. Growth mindset 2. Social belonging 3. Control (Physical environment)	Single sessions: 20–25 min	Online	Text reading Reflective summarizing Letter writing
DeBacker et al. (2018)	9th and 10th graders (N = 456)	USA	Motivational	1. Growth mindset 2. Control (No treatment)	Single session: 55 min	Online	Lesson-based introduction Reflective summarizing Letter writing
Bettinger et al. (2018)	High-school students (N = 254)	Norway	Motivational	1. Growth mindset 2. Control (Brain)	3 sessions (Two weeks apart): 45 min each	Online	Text reading Reflective summarizing Letter writing
Mills et al. (2018)	Undergraduates (N = 155)	USA	Academic	1. Growth mindset 2. Control (No treatment)	4 sessions (Each/week): About 30 min	On-site	Lesson-based introduction
Bostwick et al. (2018)	College students (N = 278)	USA	Academic	1. Growth mindset 2. Fixed mindset 3. Control (Neutral letter)	Single sessions: NA	On-site	Text reading Reflective summarizing
Schleider et al. (2018)	Adolescents (Aged 12–15, N = 96)	USA	Emotional Self-perceptional	1. Growth mindset 2. Control (Supportive therapy)	Single sessions: 20~30 minutes	Online	Text reading Testimonies Reflective summarizing Letter writing
Fink et al. (2018)	College freshmen (N = 565)	USA	Academic	1. Growth mindset 2. Control (Transition tips)	3 sessions (During a semester): NA	Online	Text reading Reflective summarizing
Rhew et al. (2018)	6th to 8th graders (N = 68)	USA	Motivational	1. Growth mindset 2. Control (Advising support)	40 sessions (5 days/week, 8 weeks, 10.5h in total): 15 min each	Online	Brainology
Calvete et al. (2019)	High-school students (N = 858)	Spain	Social	1. Growth mindset 2. Control (Human brain)	Single session: 50–60 min	On-site	Text reading Reflective summarizing Letter writing
Smith et al. (2019)	Undergraduates (N = 75)	USA	Emotional Academic	1. Growth mindset 2. Control (No treatment)	Single sessions: 75 min	On-site	Lesson-based introduction Testimonies Group discussion Reflective summarizing Letter writing
Goyer et al. (2019)	Middle-school students (N = 669)	USA	Social	1. Growth mindset 2. Combination 3. Control (Neutral activity)	6 sessions (Week 3/9/13/22/25/29): 15~25 min each	On-site Online	Text reading Reflective summarizing Letter writing
Covarrubias et al. (2019)	Undergraduates (N = 468)	USA	Academic	1. Growth mindset 2. Combination 3. Control (Factual information)	Single session: NA	Online	Text reading
Binning et al. (2019)	7th and 9th graders (N = 598)	USA	Academic	1. Persistence mindset 2. Control (Study skills)	Single session: 45 min each Booster session: 45 min each	Online	Video watching Testimonies Reflective summarizing Letter writing
Burnette et al. (2019)	Undergraduates (N = 238)	USA	Self-perceptional Motivational	1. Growth mindset 2. Control (Attention-matched)	3 modules (2 weeks apart): about 15 min each	Online	Video watching Text reading Reflective summarizing Letter writing
Yeager et al. (2019)	9th graders (N = 12,490)	USA	Academic	1. Growth mindset 2. Control (Brain)	2 sessions (20 days apart): 25 min each	Online	Text reading Testimonies Reflective summarizing Letter writing
Fernández-González et al. (2020)	High-school students (N = 123)	Spain	Social	1. Growth mindset 2. Control (Brain)	Single session: 1 h	On-site	Text reading Reflective writing Testimonies Letter writing
Zeeb et al. (2020)	7th graders (N = 59)	Germany	Motivational	1. Growth mindset 2. Control (No treatment)	10 sessions: 45 min each	On-site	Growth-mindset feedback Text reading Group discussion Reflective summarizing Letter writing
Schmidt et al. (2020)	Secondary-school students (N = 707)	USA	Motivational Self-perceptional	1. Growth mindset 2. Control (Science writing)	6 sessions (Each/week): 50 min each	Online	Brainology
Burnette et al. (2020)	College students (N = 491)	USA	Motivational	1. Growth mindset 2. Control (Health issue)	4 sessions (Every 2 weeks): 5–6 min each	Online	Lesson-based introduction
Porter et al. (2020)	8th to 10th graders (N = 354)	South Africa	Academic	1. Growth mindset 2. Control (Animals and nature)	3 core sessions (Every 2 days): 30 min each 2 booster sessions (2 and 23 weeks after 3rd session): NA	Online	Video watching Reflective summarizing Letter writing
Rege et al. (2020)	High-school students (N = 19,719)	USA & Norway	Social Academic	1. Growth mindset 2. Control (Graphic art)	2 sessions (1 to 4 weeks apart): 25 min each	Online	NA
Weisz et al. (2020)	College students (N = 146)	USA	Self-perceptional	1. Growth mindset 2. Social norm 3. Combination 4. Control (Academic difficulty)	3 sessions (Within 10 days): 1h each	On-site	Testimonies Text reading Letter writing Letter reading aloud and reciting
Perkins et al. (2021)	Adolescents (Aged 16–18, N = 80)	UK	Self-perceptional Emotional	1. Growth mindset 2. Control (No treatment)	Single session: 30 min	Online	Video watching Reflective summarizing Letter writing

).

3. Result

3.1. Descriptive Findings: How Effective Are Mindset Interventions in Enhancing Students’ Learning?

Effective mindset intervention in the current review connotes positive intervention that significantly enhanced students’ learning. Students’ learning incorporates both objective performances (e.g., GPA, see [9,59] and reported attitudes or attributions (e.g., learning goals and academic stress, see [31]). The reports were from the participants themselves (i.e., self-reported) or others (e.g., parent-reported, see [27]). Table 2 gives demographic information related to the selected 38 articles. In terms of domain, mindset intervention enhanced students’ learning mainly in five domains (k = 39, N = 52): academic (N = 19, 36.54%), motivational (N = 13, 25.00%), self-perceptional (N = 10, 19.23%), social (N = 5, 9.62%), emotional (N = 4, 7.69%) and moral (N = 1, 1.92%). Hence, the academic and motivational domains were the most common indicators of learning that benefited from mindset intervention.

Within the sample, mindset interventions worked particularly effectively among students at secondary school (65.79%, N = 25), followed by undergraduates (31.58%, N = 12) and pupils in primary school (2.63%, N = 1). Moreover, 65.79 per cent (N = 256) were independently conducted in the United States, 13.16 per cent (N = 5) in combination with American samples (N = 3, see [18,26,60]) or in collaboration with American scholars (N = 2, see [9,61]). As few as 21.05 per cent (N = 8) were independently conducted outside of America or included no American scholars, of which 18.42 per cent (N = 7) were conducted in Europe and 2.63 per cent (N = 1) in Asia. Thus, as the cradle of mindset theory, the United States was the most important arena in terms of conducting mindset interventions.

In terms of design, 44.74 per cent (N = 17) of the selected mindset interventions comprised single sessions, 34.21 per cent (N = 13) comprised between two and four sessions, and merely 21.05 per cent (N = 8) had five sessions or more. More than half (52.63%, N = 20) of the interventions were conducted online or on-site (e.g., classroom or lab) and blended locations represented 36.84 (N = 14) and 10.53 (N = 4) per cent, respectively. In terms of intervention approaches, 73.68 per cent (N = 28) adopted hybrid methods, namely combinations of human-interaction-based (e.g., lesson, interactive feedback), reading-based (e.g., text reading), writing-based (e.g., letter writing), computer-assisted (e.g., video watching), and speaking-based (e.g., self-persuasive speaking). Among the hybrid approaches, “writing-based + reading-based” seemed to be the most popular method adopted (50.00%, N = 14, e.g., see [10]). A mere 26.32 pr cent (N = 10) of the interventions used only one approach: Computer-assisted (i.e., Brainology, N = 3, see [17,21,31], human-interaction-based (N = 3, see [33,34,41]), reading-based (N = 3, see [16,18,20] and speaking-based (N = 1, see [35]).

Interestingly, most of the positively effective mindset interventions produced selective benefits. In other words, only some participants were likely to benefit more from the operation. Most beneficiaries were from socially marginalized groups such as Latino/a [59] or Blacks [41], females [20], negatively stereotyped boys [29], and lower-achieving students [15].

3.2. Main Findings: What Characterises an Effective Mindset Intervention in Enhancing Students’ Learning

3.2.1. Precondition: A Supportive Learning Environment

In the terminology of academia, students need to learn, teachers need to teach, and a place is needed to support the teaching and the learning. The key point about mindset intervention is the malleability of how people perceive themselves and the social world, but that any psychological shifting needs a platform. Students benefit not only from transforming their mindset but also from shifting the learning environment in a growth-mindset-oriented direction. Thus, in addition to offering “sufficient materials and human resources” [29] (p. 255), the educational environment should “better embody growth mindset principles and practices” [33] (p. 113). A supportive learning environment with the basic tools required to foster improvement, including but not limited to classrooms or labs, human resources for teaching, and learning materials, are prerequisites to ensure positive outcomes of psychological intervention. Specifically, the learning opportunities afforded by the school environment enable the transformed messages to interact with student motivation to promote adaptive behaviours [42]. Teachers should ensure that they can communicate a growth mindset to their students, for example [25].

From an outside perspective, however, several contextualized disruptions such as community unrest, gang violence, or strikes may explain why mindset intervention does not significantly foster adaptive learning [9]. According to Porter et al., only 47 per cent of students under treatment attended all three of the core intervention sessions, and only 17 per cent attended all five: The low level of compliance may have diluted the intervention effects. In another empirical study, the failure of a Hungary-based mindset intervention to achieve lasting effects was potentially attributed to the lack of growth-mindset pedagogy both in the school climate and in the teaching practice, given that the training of teachers as intervention experimenters only lasted for four hours [24]. The authors concluded that more efforts were needed to reinforce the school’s resources for what was the “performance arena” of mindset intervention.

It is worth noting that intervention might be most effective in an “adequately resourced context with room for improvement” [29] (p. 255). For example, if the school had imparted sufficiently relevant knowledge to the participants before they received the growth-oriented message, they would have been less likely to engage in the interventions [20]. Furthermore, if racial bias is deeply entrenched, intervention might be ineffective or even counterproductive [29].

3.2.2. Message Generating: Interaction Interaction between Person, Theory, and Context

Persons: Participants, stakeholders, experimenters

The persons involved were, first, the targets of the intervention, namely students [15,23]. In terms of the academic setting, the students varied from the lower-secondary [27] and higher-secondary [42] to the college level [28]. Marginal students for whom intervention was particularly effective [15] such as those with learning difficulties [21] and low-SES [29], and ethnic minorities [31], were typically targeted.

Second, “stakeholders” either received suggestions from the participants regarding the intervention or reinforced the intervention messages with their experiences. For example, educationally struggling younger adolescents needed the participating students’ encouragement to help them cope with setbacks [27], regardless of whether they were from the real world [23] or hypothetical [25,28,62], and older peers shared their own experiences with the participants. If the intervention was computer-based, animated characters interacting with the participants to facilitate progress had important roles: Chris and Dahlia helped the participants to cope with challenges in learning, for example, and Dr. Cerebrus introduced brain science in Brainology.

Third, researchers and experimenters conducting the intervention were equally important. They generally had a basic theoretical understanding of intervention procedures and local contexts, and were therefore capable of designing the intervention and making any changes in its ongoing implementation. Some investigations employed locals [21] to help, a practice that could bring the content closer to the participants and compensate for the shortage of specialists. Before the intervention, the experimenters typically received training about the manipulation contents and procedures [32] to complement their existing psychological expertise [21] and context familiarity [32]. In some cases, professional research companies were contracted to collect data and to process the analyses to ensure repeatability and scalability, as well as to strengthen the independence of the intervention results [15].

Theory: Subjective construal

Mindset intervention was rooted in a “solid theoretical background” [24] (p. 8) and contributed to the “subjective creation of meaning” [45] (p. 13). In particular, it enabled participating adolescents facing rejection and aggression from others to “interpret these experiences in a more benign manner, which inhibit retaliatory reactions” [45] (p. 13).

When design-thinking was applied in revising the growth-mindset intervention, it was acknowledged that theoretical guidelines were crucial to increasing the likelihood that it would be more effective for a predefined population [60]. Three theoretical elements guided the revision procedures in that study, namely: (a) How a growth mindset affected the behaviour of individuals; (b) How different cultural values responded to the same growth-mindset messages; and (c) How students could be persuaded to endorse the delivered message. As a result, the revised intervention, which included quotes from celebrities and tailored information to fit the participants’ situations, outperformed a standard intervention with a focus on general malleable messaging and interactive exercises [19,60].

The intervention worked best when the specific needs of the participants’ context were met on the one hand, and when psychological processes were accurately mediated on the other. Furthermore, to validate and sustain the intervention to cope with future challenges, it was designed to target the plasticity of intelligence and personality by focusing on developing strategies such as effort, resilience and self-efficacy among participants experiencing both prosperity and adversity [26]. Thus, it could be productive to incorporate goal-relevant messages into the intervention process. Empirically, although interventions aimed at merely persuading students by means of incremental theory, with no detailed suggestions about how to act, could potentially lead to adaptive behaviours, the adoption of detailed strategies could trigger more positive results [16].

Context: Situational and cultural level

Contextual expertise was an essential element in assessing the efficacy of mindset intervention. For instance, participating students with anxiety and depression [27] might have more interest in engaging and may behave adaptively if the value of specific relevant content were to be emphasized in the transforming message [17]. That was another reason why, in some successful intervention experiments assessing the measures or delivering the mindset message, the original implicit theory of intelligence or ability scale was replaced in domain-specific settings (e.g., the implicit theory of entrepreneurial ability, see [28]; the implicit theory of computer science ability, see [33]; views on chemistry intelligence, see [48]). Additionally, participants doing the active “saying is believing” exercise were encouraged to discuss how their domain-specific mindset influenced their studying for upcoming domain-specific exams (e.g., chemistry, see [48]).

It is worth pointing out that context refers not only to the specific local situation of participants but also to the general culture in a broader sense. Even if the materials and the procedure are the same, an intervention conducted across different cultures (e.g., Norway and the US) could differ significantly in effectiveness (effective in Norway, but not in the US, see [18]). Similarly, Brainology intervention (e.g., [31]) and Yeager et al.’s revised mindset intervention [15,60], which originated and worked well in the US, had no significant impact in Scotland [41] or The Netherlands [63]. This could be attributable to the different underlying cultural features in the populations, namely “national and cultural aspects, and educational setting” [63] (p. 827). The commonly utilized approach to mindset intervention was prevalent in the United States, but when applied in other countries the material and format needed tailoring to fit the local culture and context. As Bettinger et al. [61] note, for instance, research teams employed professional translators, and interviewed several focus groups of participant peers before the intervention, in carefully adapting the original intervention of Yeager et al. [60] to the Norwegian language, culture and context.

Thus, more attention should be paid to the cross-cultural applicability of mindset teaching. More specifically, interventions should be contextually situated [29], problem-specific and customized for a context-specific population [60]. When the manipulation message considers the participants’ backgrounds and is tailored to “align the challenges of the specific contexts” [64] (p. 275), positive outcomes are likely. For instance, to help targeted students who just started a new school, the intervention content changed from the iterative process of gaining feedback from teachers and students used in the pilot test. Additionally, each dose of two mindset interventions served a unique purpose: The first intervention focused on growth-mindset message during a transition time, and the second intervention booster these messages after participants had experienced some setbacks within a new environment [64].

3.2.3. Message Delivering: An Iterative Process

Saying is believing

According to the “saying is believing” theory [65], the more people advocate their positions in their own words, the more they are likely to endorse the message of what they were persuaded to receive. The approach was designed to foster the development of a growth mindset, allowing participants to make the messages more understandable and personalized [9,36]. As a self-persuasive tool, “Saying is believing” was thought to be effective because, by mentally rehearsing the information, communicating it to others, and making it more self-relevant, participants would find it easier to convince themselves [23,60].

After reviewing the intervention contents, participants are generally encouraged to summarize the core conclusions [19,33,59,66], respond to reflective questions [9,67], and describe strategies for applying growth-mindset principles to their own lives [27,45,68]. They may also be encouraged to write letters to peers (e.g., a friend struggling in school, see [61] or other students with similar experiences, see [45,68]) or to younger generations [22,59,69].

Interestingly, although “saying is believing” was the common approach adopted to increase the credibility of the intervention messages, very few mindset-intervention studies that did not adopt it reported somewhat positive academic achievements [70]. One plausible explanation for this is that the intervention was conducted at critical time points of high academic challenge (immediately after the midterm exam, see [70], or immediately before the chemistry exam, see [48]). Hence, it was well-timed to engage students with the growth-mindset concept and to motivate effort when participants have enough opportunities to make the effort worthwhile. Thus, the easy and quick mindset intervention could also have “created a recursive process leading to lasting changes” [70] (p. 189).

Repeating the message to accumulate its effects

Traces of incremental theory were present in a variety of ways in the messages conveyed to participants during the intervention, which enhanced their scientific validity. The intervention typically delivered the messages in at least three ways: Text presentation (e.g., slideshows, see [62]), text reading [10], video presentation (see [64]), testimonials (e.g., from older peers, see [27]), and “saying is believing” writing exercises. Before advocating a position, the participants usually had a basic understanding of the intervention from watching video clips and listening to lectures about the brain of the malleability of the intelligence [66]. After the “saying is believing” exercise they were guided in repeat-practicing the materials, such as rewriting letters with examples from their own lives to encourage other adolescents, then synthesizing the two letters and turning them into speech, audiotaping them and listening to the recordings [23]. After all this, it was reasonable to assume that the growth-mindset message had been effectively delivered.

The accumulative effects of message repeating were also reflected in “the number of direct interventions” and the processes “catalyzed by previous session” [24] (p. 9). Many authors refer to single-session intervention [27,35], but a large number of studies report two or more sessions (see [10,15,36,42,60] for two sessions; [23] for three sessions; [66] for four sessions; [9,32] for five or more sessions). It is possible to deliver the mindset message repeatedly even in single-session intervention, in up to eight steps from a quiz-based introduction to letter writing (four presentations plus three interactive activities in the classroom, see [62]). Alternatively, the message could be refreshed and activated via post-intervention reminders containing intervention materials [9]. The accumulating effects are also reflected in blended approaches. For example, up-to-six-session online workshops were paired with teacher reinforcement in the form of supplementary classroom activities to enhance incremental message delivery [17].

3.2.4. Message Internalising: Persuasively Yet Stealthily

Mindset intervention aimed at persuading participants to view themselves and others as well as the surrounding environment in a positive light. Its effectiveness relied, to some extent, on the persuasive delivery of the intention, namely to facilitate and support rather than directly to “instill”. “Formulating a persuasive message to another person often strengthens one’s persuasion” [25] (p. 6). As “a persuasive technique” [28] (p. 5), growth-mindset messaging is meant to bolster change in an individual’s perception, especially among participants who may not have had sufficient mastery experience. For instance, in addition to summarizing what the mindset-intervention messages gave them, adolescents were asked to complete self-persuasive writing exercises in which they described a time they felt rejected, withdrawn, or disappointed by others, and gave advice to another student with similar experiences [45,68].

Stealthy delivery, in turn, is intended to enhance the validity of transformed messages. Thus, participants were not informed in detail about the intervention process, including the fact that they were the target of an experiment. All 38 articles in the current systematic review adopted a blindly randomized design, meaning that the students were randomly assigned to the experimental or control group. Moreover, double-blind randomness was applied in that allocation was concealed to both participants and investigators [17,33,45,48,68,70]. Specifically, the intervention materials were incorporated into the students’ regular-track curriculum during normal class time (e.g., science classes, see [17]; entrepreneurship classes, see [28]; mathematics classes, see [34]; computer-science classes, see [33]). Hence, the students believed that they were engaging in regular classroom activities [64].

Another way in which the mindset intervention was sometimes implemented was as a course requirement, such that teachers embedded relevant ideas into ordinary courses to boost a growth mindset and motivation among the students. For example, the incremental theory has been integrated into chemistry homework assignments immediately before students have their incoming unit check [48], and given-mindset strategies have been highlighted for outdoor-expedition courses [26].

Figure 2 demonstrated the characteristics of effective mindset interventions in enhancing students’ learning.

4. Discussion

In the context of SDGs at the educational level for students, our work aimed to answer a set of questions concerning (1) the effectiveness of mindset interventions in enhancing students’ learning and (2) the characteristics of effective mindset interventions. In terms of effectiveness, mindset interventions benefitted students in the academic domain of sustainable development most, in particular, the majority of them having been conducted in the United States among students at secondary schools. A single-session design with hybrid approaches, especially a combination of reading-based and writing-based methods, was commonly adopted. First, in terms of characteristics, a supportive learning environment with sufficient resources but not-completely-saturated psychological support (i.e., a psychological need for a growth mindset among the targeted groups) is not only a prerequisite, but also reflects the need to implement a mindset intervention. Second, an effective intervention is reliant on three procedures: (1) Mutual interaction involving the person, the context, and the theory to generate the message; (2) Iterative processes to deliver the message; (3) Persuasive yet stealthy facilitation of its internalization.

4.1. Theoretical Foundations Underlying the Current Review

According to the person-environment fit theory from the developmental perspective, people flourish in environmental settings that are compatible with their developmental growth, including their skills, interests and values, for example [71]. “Subjective meanings do not work in a vacuum but within complex systems” [72] (p. 620), and the social circle’s diversity “enhances individual absorptive capacity” [73] (p. 11). This makes it easier to understand why a supportive learning environment as a precondition guarantees the success of an intervention [38,39].

As a social activity, learning is an internal-development process in which individuals interact with others in the environment [74]. Accordingly, the mindset is constructed in an interaction process, reflecting our statement of interaction needed to generate the growth message. This emphasis on the person and the context also reflects person-environment-fit theory [71], concerning the matching of environments to the psychological needs of individuals. Meanwhile, we followed theoretical guidelines to guarantee that the delivered message precisely targeted the participants in the given context [37,40]. Bodies of psychological research support the notion of a mindset [5], including but not limited to attribution theory [75,76], goal theory [77,78], self-determination theory [79], and expectancy-value theory [80]. Accordingly, rather than via objective numbers or attributes [32,81], one reason why a social-psychological intervention such as mindset intervention was impactful in guiding individuals’ behaviour was that it precisely targeted the participants’ subjective construal [40]. The accessibility of the mental concept may affect how people think about themselves and others, and how they interpret the social experience [82].

Rather than directly conveying ideas to participants who passively receive them, encouraging them actively to engage in an operation may be more effective in helping them to internalise the information [83]. Additionally, stealthy delivery enhances the validity of transformed messages [40,84]: Participants are less likely to reject ideas if “they do not feel controlling……they minimize resistance and reactance to the message” [40] (p. 284). This quiet and careful approach echoes the targeting of participants’ subjective construal of intervention [40], in that the forces that drive it are hard to see [85], and the negative stereotyping or social marginalization of participants is avoided [40]. This could also explain why a seemingly small intervention may produce a large effect [38].

In synthesizing predecessors’ findings in the social-psychological domain [40,86,87], Orosz et al. [24] summarized the core elements required to conduct an effective mindset intervention, which verifies the precision of our currently proposed characteristics to some degree. First, the “psychologically precise theory and tool” (p. 9) is consistent with our targeting of psychologically subjective construal (theory). Second, “targets a specific group” and “appropriate context” (p. 9) match another two of our message-generating components: Person (Particularly the subject of the intervention) and context. Third, an iterative process is what we emphasize in our systematic review. Fourth, “not using direct persuasive appeal” and “not help but give an opportunity” (p. 9) refer to our notion of persuasive yet stealthy message delivery.

4.2. Reflecting on the Mindset Intervention and Rethinking the Meaning of Mindset

Regarding the mindset intervention, first, in accordance with our notion of stealthy delivery, one reason for the null difference between the experiment and the control group might be that the participants were aware of their participation in the test. Sometimes, the intervention visit included several surveys, hence they recognized that they were being tested. Consequently, the students may have been demonstrating resistance to the intervention. In addition, group allocation might lead to cross-contamination in that growth-mindset messages “can be difficult to confine to students in the intervention condition” [33] (p. 113), particularly if intervention-group students shared the “active” ingredients with their control-group peers. Peer cultures and classroom contexts proved to construct the meaning system of mindset among adolescents [88].

Second, the message repetition in the iterative process was not dull, and it was done as flexibly and with as much enjoyment as possible. In line with the core salience of messages, the delivery style and specific content varied. For example, a professional graphic designer animated the content of the growth-mindset intervention, and all sections were narrated, to make the intervention enjoyable so that students would find it engaging [64]. Paradoxically, long-term interventions with multiple sessions did not necessarily reinforce the iterative processes, because excessive information allows “room for diverse interpretations” [24] (p. 9). Although it is still not clear how many sessions are required or how long it takes for the intervention to yield optimal efficacy, to some degree, in the case of social-psychological intervention, less is more.

Regarding the meaning of mindset, there may be a need to rethink it. In the case of social psychology, interventions aimed at changing people’s self- and social perceptions, attribution of performance, and sense of social connectedness [39], for example. Critically, instead of the malleability of personal characteristics, the tenets of the meaning system of mindset, conceptualizing such characteristics in academic or social terms distinguish different categories of mindset [88]. Specifically, fixed theorists emphasize academic performance while growth theorists on social skills and mixed ones on interpersonal behaviours within the academic context. The current review prefers the social behavioral patterns within given cultural contexts. Rather than the view of the malleability of self-intelligence or personality [4] the mindset was also the construal of social perception. Even though the intervention did not remove the objective adversities in the individuals’ situating environment, mind shifts throughout the process would enable them to change their interpretation of adversity, and then to decide whether they were affected by them [37]. Hence, “it can be as important to change people’s construal—their interpretations of the social world and their place in it—as it is to change the objective environment” [39] (p. 1252).

4.3. Implications

On the theoretical level, the current research is the first systematic review to explore the manipulation mechanism of empirical investigations into mindset interventions. The characteristics we discuss are rooted in message generation, delivery and internalization. Our review enhances understanding of how mindset intervention works and how it could work more effectively. Based on a supportive learning environment, the design is reliant on both theoretical and contextual expertise, iterative processes and persuasive but stealthy delivery. The growth-transmission mechanism mentioned above applies not only to mindset intervention but also to a wider range of social-psychological interventions. Thus, the present study enriches and extends the existing literature on how these interventions work and how they could work more effectively.

On the practical level, we propose a relatively scientific design for mindset intervention. Specifically, effective intervention typically involves, first, exposing individuals to theoretical and scientific information (by means of article reading, video watching, slideshows, for example) about the malleability of intelligence and personality; Second, using persuasive testimonials and case studies from celebrities or neighboring models; Third, encouraging self-persuasive writing (generally linked to personal experiences) addressed to oneself or someone else (typically struggling peers). Growth-mindset information is also incorporated into the participants’ routine activities in which they are randomly assigned to different conditions and engage in the intervention as naturally and actively as possible. Well-designed interventions that precisely target the participants’ psychological construal, and in which context is situated, the pattern of growth-mindset delivery is well-timed and flexibly altered, could be beneficial to students in school, children in families, and trainees in companies, among others, affecting a wide range of adaptive behaviours.

4.4. Limitations and Future Directions

This review is limited in several respects. First, although we used a three-phase search method including databases, prominent researchers and reference backtracking, our sample consists mainly of published journal articles and book chapters. We might therefore have left aside other literature such as dissertations, conference presentations, and even unpublished articles. Second, our seven-item inclusion criteria excluded potentially valuable studies, such as mindset interventions targeted at other socially relevant attributes (e.g., the implicit theory of morality, see [49], for more details see the “Mindset scale” in Table 1), and empirical investigation involving fixed- and growth-mindset conditions but no control group (e.g., [55], for more details see “Research design” in Table 1). The omitted articles might have been informative and could have enriched and extended, even challenged our findings. Third, we failed to discuss the efficacy of the dichotomous scales used in empirical mindset intervention from a scientific perspective, such as in meta-analyses.

Accordingly, to find the best balance between high efficacy and low cost in mindset intervention, the following questions should be asked. What mindset scale (general vs. domain-specific) might be appropriate to yield higher efficacy? How long (long-term vs. short-term) should the intervention take? What kind of sessions (on-site vs. online, single vs. multiple, sole mindset vs. combined with other traits) should be given? What is the best way of generalizing the benefits of intervention to a wider range of participants instead of minorities? What is the best way of achieving lasting rather than a temporary effect immediately after an intervention? Finding answers to these questions would require further empirical investigation, with a specific focus on the comparative perspective. Moreover, given that context plays a crucial role in the implementation of interventions, and those existing empirical investigations of mindset intervention referred mainly to western countries, more evidence from other contexts, such as Asia, is required to strengthen and enrich these claims. A much deeper and stronger foundation would yield better instruments and benefit a broader population from the perspective of cross-cultural applicability.

5. Conclusions

While increasingly empirical interventions of mindset have proved to enhance individuals’ behavioral patterns, none of the existing research has systematically investigated the formation and transmission of growth messages in these interventions. The current study is, to the best of our knowledge, the first systematic review to explore the characteristics of mindset intervention from the perspective of message-transforming mechanisms. Through synthesizing and compiling empirical research published between 2012 and 2021, the current review consists of thirty-eight studies focusing on mindset interventions. Due to the origination and cost-effective rules of mindset theory, most mindset interventions are conducted in the United States and are effective particularly for secondary-school students. A single session with the “saying is believing” exercise is commonly adopted. To maximize the efficacy of growth-message transmission from instructors to recipients, a supportive learning environment provided the “arena” of mindset intervention, then reciprocal interaction and iterative processes with persuasive and stealthy way facilitate the generation, delivery, and internalization of growth messages. The findings not only theoretically enriched the mindset theory from the perspective of intervention mechanism, but also practically inspire scholars and teachers to design effective instruments of mindset intervention. However, the current review failed to discuss the specific efficacy of the dichotomous scale from a statistically scientific perspective including the number, length, location, and approach of the intervention sessions. Thus, future research might focus on the balance of efficacy and cost to figure out a well-designed paradigm of growth mindset intervention. Meanwhile, the cross-cultural applicability of mindset intervention is yet explored to benefit a broader population.