Pj. Keegan Knittle
Increasing motivation for physical activity: A meta-analysis identifying effective behavior change techniques
Keegan Knittle1, Johanna Nurmi, Rik Crutzen, Stephan Dombrowski, Marguerite Beattie, Nelli Hankonen
1University of Helsinki
Background: Motivation is a proximal determinant of behaviour, and increasing motivation is central to most health behaviour change interventions. This systematic review and meta-analysis sought to identify features of physical activity interventions associated with favourable changes in three prominent motivational constructs: intention, stage of change and autonomous motivation. Methods : A systematic literature search identified 89 intervention studies (k = 200; N = 19,212) which assessed changes in these motivational constructs for physical activity. Intervention descriptions were coded for potential moderators, including behavior change techniques (BCTs), modes of delivery and theory use.
Results: Random effects comparative subgroup analyses identified 18 BCTs and 10 modes of delivery independently associated with changes in at least one motivational outcome (effect sizes ranged from d = 0.12 to d = 0.74). Interventions delivered face-to-face or in gym settings, or which included the BCTs ‘behavioural goal setting’, ‘self-monitoring (behaviour)’ or ‘behavioural practice/rehearsal’, or which combined self-monitoring (behaviour) with any other BCT derived from control theory, were all associated with beneficial changes in multiple motivational constructs (effect sizes ranged from d = 0.12 to d = 0.46). Meta-regression analyses indicated that increases in intention and stage of change, but not autonomous motivation, were significantly related to increases in physical activity.
Discussion: The intervention characteristics associated with changes in motivation seemed to form clusters related to behavioural experience and self-regulation, which have previously been linked to changes in physical activity behaviour. These BCTs and modes of delivery merit further systematic study, and can be used as a foundation for improving interventions targeting increases in motivation for physical activity.
Shifting the mindset from actions to interactions: Network models for studying physical activity determinants in the Let’s Move It trial
Matti Heino1, Keegan Knittle, Nelli Hankonen
1University of Helsinki
Introduction: During the early 2010s, a crisis of confidence in research findings was sparked by failed replications of studies including but, not limited to, psychology, medicine, economics and ecology. A methodological reform has followed, but has yet not been taken to its logical conclusion. Health researchers have an exceptional connection to the complexity of the real world, but sometimes do not take this into account when extrapolating from low-dimensional models to high-dimensional reality. The current work aims to demonstrate a how we can bring models closer to what we believe the reality to entail.
Methods: Psychological network analysis was used to model and visualise interactions between behaviour change technique use, motivation and physical activity. The sample consisted of 1165 older adolescents in the Let’s Move It trial, which evaluated an intervention to increase physical activity and decrease sedentary behavior.
Results: Network analysis turned out to be a fitting conceptual framework for examining the relationships in the light of what is known of the processes under scrutiny. Network structure predicted later activity.
Discussion: With complex multicausal processes, conventional use of standard regression models allow for unsatisfactory inferences. Thinking in terms of mutually interacting, coupled processes makes clear the need for new approaches to studying health behaviours. Networks provide an intuitive framework for this. As with all modeling approaches, reason must be practiced instead of ritualistically applying the analysis.
Precious n-of-1 trial: within person variation in motivation and self-efficacy explain changes in daily physical activity
Johanna Nurmi1, Keegan Knittle, Felix Naughton, Stephen Sutton, Todor Ginchev, Fida Khattak & Ari Haukkala
1University of Helsinki
Background: An important and under-researched method of increasing our understanding of the psychological determinants of motivation and physical activity is through daily, individual measurements. The Precious smartphone app was designed to collect both self-reported and objective data while delivering repeated, randomised interventions with elements from motivational interviewing and biofeedback (Firstbeat) to explore how interventions affect psychological determinants of physical activity and activity itself.
Methods: Fifteen healthy adults answered daily questions on the Precious app on motivation, self-efficacy, perceived barriers, and happiness during a six-week factorial n-of-1 trial (440 observations). The app also collected data on the intervention engagement, and activity bracelets collected objective physical activity data. The association between psychological determinants and intervention effects on daily steps were analysed using multilevel modelling (level 1: daily observations, level 2: participants).
Results: Individuals had highly different motivational and activity profiles , justifying multilevel modelling. Self-efficacy and motivation were the strongest predictors of daily activity within-individuals. Pain, illness, and perceived barriers were only weakly associated with activity. Happiness was associated with afternoon motivation but not physical activity. Neither motivation nor daily steps were higher on intervention days with digital motivational interviewing components or Firstbeat biofeedback than on non-intervention days.
Discussion: Motivational constructs predicted daily steps over and above pain, illness, and perceived barriers. Findings on the association of motivational constructs and physical activity, measured daily within individuals, have value in confirming the predictive nature of these constructs, central to many behavioural theories. Increasing motivation with digital content remains a challenge.
Implicit process interventions in dietary behaviors: A systematic review and meta-analysis
Matthias Aulbach1, Ari Haukkala, Keegan Knittle
1University of Helsinki
Background: Dual-process models integrate deliberative and impulsive mental systems, and predict dietary behaviors better than deliberative processes alone. Computerized interventions developed to directly alter impulsive behavioral antecedents include the Go/No-Go, Stop-Signal and Approach-Avoidance tasks. The current meta-analysis examines the effects of these tasks on dietary behaviors, explores sample- and task-related characteristics as potential moderators of effectiveness, and examines implicit bias change as a proposed mediator.
Methods: Nineteen randomized controlled trials testing one of these tasks (38 comparisons) were included in a random-effects meta-analysis
Results: Interventions had small cumulative effects on eating-related behavioral outcomes (g = -0.18, CI95 = [-0.32; -0.05], p=.008) and stimulus evaluations (g= -0.30, CI95 = [-0.50; -0.09], p= .004). Task type was the only significant moderator of these effects, with Go/No-Go tasks producing larger effects than Stop-Signal or Approach-Avoidance tasks. Effects of interventions on implicit biases were related to effects on eating behavior (B= 0.46, CI95= [0.11; 0.82], p= .01).
Discussion: Future research should focus on Go/No-Go tasks for altering dietary behavior via the impulsive system, and should explore effects of these interventions over longer periods of time with repeated exposures, especially in real-world as opposed to laboratory settings.