Behavior Change

From CS260 Fall 2011
Jump to: navigation, search

Bjoern's Slides

Extra Materials

Discussant's Materials


Reading Responses

Valkyrie Savage - 10/12/2011 13:13:19

Main idea:

People can change their behaviour, and to their own advantage, but to help them do that we need to understand them. What motivates them in the first place? How can we leverage that to different ends?


The concepts covered in the UbiFit paper seem to fit with a schema that a critical person would likely come up with for how people would elect to interact with their health. It seems like kind of a shame that people don’t trust their doctors to make good health decisions for them. I can understand that we aren’t generally excited about the idea of being turned into statistics (e.g. with the national fitness goals), but not trusting one’s doctor to make good decisions seems like, well, a pretty serious flaw somewhere. That’s beside the point, I suppose. The discussion of how people prefer to set their goals was well-motivated, and is important to the spirit of things that are coming to particularly the mobile computing world: things like mapmyride and the myriad diet calculators that allow people to track what it is that they are doing to improve their general health. I appreciate seeing HCI studies that focus on the Human part.

The second paper was both drier and more contentious. It hardly seems worth commenting on either the TRA or the TPB: they’ve both been debunked, and anyway the things that they leave out are somewhat damning. How can we ignore that people are not rational actors? What model can count that forgets the slippery slope? It’s useful that these ideas were incorporated into the Integrated Behavioural Model, and I approved of it mostly. It still doesn’t seem to capture spur-of-the-moment decisions. I know they did a study to determine whether or not the model could be reasonably applied to situations that are “in the moment” (sex), but I wasn’t quite convinced. Also, why should they exclude situations where there are some or many environmental constraints that make the behaviour very difficult? That’s the thrill for some people. Sometimes we’re more likely to do something that’s made more difficult, as it gives a greater sense of achievement.

Laura Devendorf - 10/16/2011 9:24:38

Goal Setting Considerations discusses a study of UbiFit, a mobile application aimed at improving the user's fitness. The Theory of Reasoned Action paper gives background and discusses theories that allow researches to predict and influence behaviors.

The UbiFit paper notes that the two important factors in developing persuasive technologies: goal sources and goal time-frames. The paper reports the results of a user study that takes these factors into considerations and asks uses about their preferences for each. The results were interesting and show that everyone wants and needs something slightly different. While everyone responds slightly differently, trends emerged to support that people prefer to have self-set goals or to work with a fitness expert to set goals. Overall, the paper provided a nice set of study results to aid future research in persuasive technology. It gives other developers an idea of what and how to develop similar applications. I would have liked to have seen more results about whether or not the user study participants health was significantly improved.

The Theory of Reason Action Paper explains the constructs of two theories: Theory of Reasoned Action (TRA) and Theory of Planned Behavior (TPB) and how they can be applied to predict behavior. I'm not sure how one would go about successfully applying these theories since there seems to be a large number of factors to take into a account. Choosing the right factors and investigating them could determine the success or failure of your efforts. As a research paper, I see merit in the theories and applications for them.

Steve Rubin - 10/16/2011 14:53:48

The first paper was from the psychology literature, and explained the theory of reasoned action and planned behavior as well as an integrated model of behavior. The second paper was about goal-setting in persuasive technologies. Both were concerned were, at a high level, concerned with persuading people to behave in certain ways.

I am always skeptical of models that try to explain peoples' behaviors. I don't say this to try to be unscientific, but modifying behavior is a complicated process, and I don't think there is some grand unified theory of behavior. However, for the sake of this write-up, I'm going to accept the TRA/TPB model. The whole issue of "scoring" from (-3,3) on belief and evaluation and using their product as a metric in determining behavior is a strange, somewhat illogical system. They don't give much thought to how these measurements come about--we could enumerate endless logically accurate but totally worthless statements. For example: "If I go to the gym, I will be hit with a meteorite." Well, I don't believe this, so I score it -3 belief. Getting hit by a meteorite is bad, so it gets a -3 evaluation. Therefore that's +9 to the behavioral model that would push me toward going to the gym. But this is an obvious logical fallacy: the bad thing and the action in this case are (almost) totally independent, and should not be scored together. I'm sure the authors know this, but the model should make this normalization-for-unrealistic-outcomes clear in the model.

The second paper was frustrating because none of their analysis was quantitative. For issues dealing with models of behavior and persuasion, I can accept that the qualitative responses given by the participants are important: for example, it's necessary to try to understand why a person is is acting the way they're acting. However, the research in this paper would have been a perfect medium to show quantitative results: how much more did people work out? Did they shed pounds/gain strength? Did they use the UbiFit system uniformly throughout the trial period? The paper was focused on exploring the design space of goal-setting, but while it addressed how users prefer to set goals, it did not address how well these goals work. Their goal-setting considerations are also very domain-specific--I'd be interested to see how these apply to, for example, work-related goals.

Amanda Ren - 10/16/2011 15:36:06

The Glanz paper covers the theories that link behavior and action and applies the Integrated Behavioral Model in a cross-cultural application to prove its use for behavior change.

This paper was a very lengthy read. It begins by describing how the Theory of reasoned Action was first proposed to explain behavior through behavioral intentions and a person's attitude toward the behavior. Later, the theory of planned behavior was introduced to include perceived control over behavior. To bring together constructs from both TRA and TPB, the Integrated Behavioral Model was formulated to also include motivation as a big part of behavior. I think this paper is important and relates to today's technologies because it can help us understand why people choose to embrace a certain technology or even how we could use certain technology to invoke behavioral intervention in people. I like how as an example, they showed a cross-cultural application, which shows how these theories can be applied even across other cultures.

The Consolvo explores uses self goal setting with a mobile phone system to help individuals monitor their physical activity.

I really enjoyed reading this paper because it connects two things relevant in today's world. First, inactivity has been a notable trend. Second, people are greatly dependent on their smart phones and other technologies(which could also be a reason for the first). The fact that we can use the second to help improve the first thing is a very good idea. I thought it was interesting how the subjects in the experiment were most motivated when they set the goals themselves, but were very negative when it came to set guidelines, which made sense. I was surprised, however, when they even objected to having the goals be set by a medical expert. I hope this idea gains popularity as the mobile phone provides a good platform for motivating people to perform more physical activity because it is a tool they always carry with them and they can always easily check their progress.

Galen Panger - 10/16/2011 17:58:27

I will be leading discussion.

Viraj Kulkarni - 10/16/2011 20:57:42

TRA, TPB and IBM deals with quantifying the factors that go behind a person performing an action. The Theory of Reasoned Action (TRA) states that behavioural intention (BI) is a result of two factors: the person's attitude towards the behaviour (A) and subjective norm (SN). Attitude towards a particular behaviour quantifies what the person thinks regarding how useful the behaviour is to him/her. Subjective norm deals with what the person thinks about what other people around him (friends, partners, family) feel about him/her doing that behaviour. TRA does not take into account factors outside the person's control. To overcome this, TPB introduces a new factor perceived behavioural control (PBC) which quantifies factors outside the person's control. This model provides a very useful way of analyzing why people do/not do a particular thing. It also shows that people might think something is very important but still won't perform that action. For instance, a person might believe that exercise is very important for health (strong positive attitude towards exercise). But he may not exercise because his partner doesn't deem it necessary (weak SN). Or he may not exercise simply because he doesn't have time for it (negative PCB).

The second paper, 'Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity', discusses the system Ubifit which is a mobile technology developed to pursuade people to stay fit by settings goals for them which they would complete if they exercise. The paper presents findings related to various approaches to setting goals and discusses the considerations in the process of selecting the goals to set. This reminds me of a very nice application I had seen. It was for patients who had high blood pressure. High BP patients must control their emotions and teaching them to do that was the purpose of the application. The heart rate was monitored using a band attached to their fingers and there was a screen showing a feather floating in the air. As soon as the heart rate increased beyond a set threshold, the feather would start drifting towards the ground. If the patient calmed down and controlled his heart rate, the feather would again rise and start floating again. It's a simple application but it was very effective.

Cheng Lu - 10/16/2011 22:21:37

The first paper, “Theory of Reasoned Action, Theory of Planned Behavior and the Integrated Behavior Model”, illustrate some important term in the HCI field. The theory of reasoned action (TRA) derived from previous research that started out as the theory of attitude, which led to the study of attitude and behavior. The theory was born largely out of frustration with traditional attitude–behavior research, much of which found weak correlations between attitude measures and performance of volitional behaviors. The key application of the theory of reasoned action is prediction of behavioral intention, spanning predictions of attitude and predictions of behavior. The subsequent separation of behavioral intention from behavior allows for explanation of limiting factors on attitudinal influence. In psychology, the theory of planned behavior (TPB) is a theory about the link between attitudes and behavior. It is one of the most predictive persuasion theories. It has been applied to studies of the relations among beliefs, attitudes, behavioral intentions and behaviors in various fields such as advertising, public relations, advertising campaigns and healthcare. The Integrated Behavioral Model (IBM), includes constructs from both TRA and TPB and other theories.

The second paper, “Goal-setting Consideration for Persuasive Technologies that Encourage Physical Activity”, shows that Goal-setting has been an effective strategy for changing behavior; therefore employing goal-setting in persuasive technologies could be an effective way to encourage behavior change. The paper developed persuasive technologies to encourage individuals to live healthy lifestyles with a focus on being physically active. As part of the investigations, they have explored individuals’ reactions to goal-setting, specifically goal sources and goal timeframes. The paper presents our findings related to various approaches for implementing goal-setting in a persuasive technology to encourage physical activity.

Hanzhong (Ayden) Ye - 10/16/2011 22:44:33

The papers for Monday discuss the issues in behavior change. The first material introduces several theories which I haven’t heard before: Theory of Reasoned Action (TRA), Theory of Planned Behavior (TPB) and the Integrated Behavioral Model (IBM). TRA and TPB suggest many constructs which help to explain the determinant of behavior. Based on these two theory, several factors such as the person’s attitude, subjective norm as well as perceived control would together contribute to the intention to perform a given behavior, and eventually lead the behavior to be carried out. The article also talks about measures of TRA, TPB constructs and the methodology used to test these theories. Integrated Behavior Model (IBM) shows a more comprehensive model to explain the mechanism of behavior, while considering additional factors such as knowledge and skills required, salience of the behavior, habit factor, etc. The application of IBM to HIV prevention in Zimbabwe is a good example of how it is used to solve real world problems.

The second paper discusses the strategy of goal-setting which assists to encourage physical activities. The research develops persuasive technologies to encourage people to live healthy lifestyles with a focus on being physically active. By launching a specific research with the UbiFit system over a period of three months, the authors record and analyze the results separately in term of goal sources and goal timeframes. The authors also discuss the effect of making goal commitment public. This paper helps me to learn the importance of goal-setting and how will goal-setting help to boost efficiency. The goal-setting theory of Locke and Latham is not only of academic significance, but also helpful to know in our daily life. I think for me the most useful piece of words in this paper is its reference to the goal-setting thoery: specific, difficult goals consistently led to higher performance than urging people to do their best--this seems to be a very interesting phenomenon, which has been neglected by people in daily work from time to time.

Alex Chung - 10/16/2011 23:10:39

How can technology be designed to influence behavior change? Theory of Reasoned Action (TRA) was developed to better understand the correlations between attitudes, intentions, and behaviors. According to TRA, behavior is goal-oriented. Yet there are latent factors that influence the behavioral outcome such as beliefs and attitudes, normative beliefs, reasoned action, and perceived control.

What influences are? Two influences on attitudes are belief (b) and evaluation (e) combining in linear regression: A = sum of b * e. If attitude is the major factor dictating the direction of behavior, then it shouldn’t improve your attitude toward weight loss if you strongly believe that you’re overweight. But empirical evidence suggests that people are more motivated to act upon weight loss. Thus it means there are other factors influencing a person’s behavior.

Two influences on subjective norm are normative beliefs (a person’s perception of what sort of behavior is approved or disapproved by society) and motivation to comply with those referents. TRA assumes that the most important direction determinant of behavior is behavioral intention (BI), in which, attitude (A) and subjective norms (SN) combine linearly as a weighted sum to produce intention: BI = Wi * A + Wj * SN.

The intention to perform the behavior from attitude and subjective norm will persuade the person to act but the person must also believe that the outcome of his/her behavior can lead to the desired outcome. These other factors on the behavioral intention are control belief and perceived power. Together, they form the perceived control (PC). For example, a person wanting to lose weight would be motivated by the belief that exercising more will lead to weight loss. That’s the control belief but it only motivates a person to choose exercise as method of weight loss. In order to have a long-term effect, there must be influenced by previous positive experience such as seeing actual weight loss from exercising more. That’s the perceived power. Therefore, behavior intention (BI) = Wi * A + Wj * SN + PC.

Overall, human computer interaction designs focus on human behavior and user’s bias in decision can be used as persuasive influences when it comes to design decision. In general, beliefs influence attitude, which affect behavior. An attitude is basically a collection of beliefs. Since belief is malleable, marketing should focus on popular beliefs. While we can use persuasion to influence decision-marking process, its effect is relative short term. Behavioral change is a systemic change with long-term effect; therefore, it requires stable influences to reinforce behavior change.

The second paper on Goal-Setting Theory demonstrates how the theory of human behavior can drive the design process of a persuasive technology that encourages physical activity.

On the other hand, I wish that UbiFit system could explore the impact of exercising in group setting against personal training. My assumption is that working in group setting would increase the weighted sum of social norm. But it might or might not positively affect the weighted sum of attitude toward exercising. Other social network such as Nike Plus allows individual runners to get together online, how do they compare to the running clubs that meet in person. I would be very interested to find out.

Hong Wu - 10/16/2011 23:21:16

Main Idea:

The two papers are talking about how people’s behavior can be changed.


“Theory of Reasoned Action, Theory of Planned Behavior, and the Integrated Behavioral Model” described the concepts of Theory of Reasoned Action (TRA) , the Theory of Planned Behavior (TPB), and the Integrated Behavioral Model (IBM). The paper also demonstrates how IBM is generated from TRA, TPB and other theories. The paper shows us how behavior and control beliefs determine behavioral interaction. The paper also describes cross cultural application in IBM.

“Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity.” shows that applies goals in technologies is a good way for behavior change. Two things, goal sources and goal timeframes are talked in the paper. From the experiment, the paper shows that most people prefer to set their goal and have a week to accomplish the task.

The purpose of software is to change the way people interact with the world. It is very important to understand the way people think of the world. From “Goal-setting” paper, it come to me that a software will be popular if it can shorten the period that people can set their goal easier, like google calendar. On the other hand, a software will be popular if it can help people to accomplish the task easier, like Microsoft Office.

Ali Sinan Koksal - 10/17/2011 0:01:25

The reading assignments for this week focused on the study of behavior change. The book chapter presents some of the theories that are part of this research field. These theories typically assume that there is a direct causal relation between behavioral intention and behavior itself. Theory of reasoned action considers behavioral beliefs and normative beliefs, while theory of planned behavior augments it with control beliefs. All the beliefs are weighted by the degree to which they are important in a given study. The proposed integrated behavioral model integrates these constructs while also considering other constraints such as knowledge, skills and environmental factors. The ultimate goal is to determine the specific beliefs that should be targeted to induce behavior change.

The second paper focuses on the goal-setting aspects of UbiFit, a persuasive technology system that aims to encourage physical activity. Two aspects of goal setting for workouts are investigated upon: i) who should be setting the goals for the workouts, and ii) what should the timeframe for achieving goals.

I believe IBM integrates essential factors that are missing in the previous TRA and TPB. A person's knowledge and skill, as well as the environmental constraints and his habits are expected to influence whether he performs a given behavior.

In the UbiFit paper, the authors draw the conclusion that users desire mainly to set their own goals, but are also keen to consider social groups or fitness experts as goal-setting sources. The most preferred options, e.g. the fixed week timeframe model, are set as default options in the system, however non of the choices that were considered seem to be discarded as the result of the evaluation. I would have expected a more detailed analysis, potentially involving inferential statistics, to evaluate the different design considerations, given that 3 months were spent for running the study.

Yin-Chia Yeh - 10/17/2011 0:26:54

Two papers about how to change people’s behavior change are read. The IBM (integrated behavior model) paper provides a general model of human behavior to guide researchers finding salient factors that affect behavior. It uses HIV prevention in Zimbabwe as example to demonstrate the usage of IBM model. It also reviews two precedent models TRA (theory of reasoned action) and TPB (theory of planned behavior.) The UbiFit paper proposes a cell phone based persuasive system that encourages physical activities. It discusses the effect of different goal source (who sets the goal) and goal timeframes (time frame to achieve the goal.) I feel like the behavior study is pretty like training classifier in machine learning so I will compare this paper with concerns in classification. I think the most critical part of IBM paper is pilot study that elicits important factors that affect behavior. Just as in classification, having really meaningful features solve more than half of the problem, probably more than 80%. If your features are highly correlated to the behavior, you can adopt a simple model to predict the behavior, just as in the IBM paper simple linear combination of features are used. On the other hand, if you have noisy features (or collected non-significant data in your study), you will have to introduce much complicated models to classify the data and the results are still often unsatisfactory. The other thing I notice in this paper is that they translate the response of participants twice to make sure the translation is precise. I think that is really impressive to me how they conduct rigorous research. The UbiFit paper discusses how to set goal to maximize people’s performance. This reminds me an interesting application I heard before. Japanese are well known for saving money. They even invent a special piggy bank to encourage children saving money. The piggy bank has an accompanied video role playing game and to level up your character you need to save more money into the piggy bank! I think that is really a cool idea. On the other hand, this leads me to think that the game must be somewhat attractive but not too attractive, because when the game is too interesting I will want to cheat! I wonder if there are any studies discuss how different strength of incentives affects performance.

Yun Jin - 10/17/2011 0:49:32

The first paper describes the historical development of the Theory of Reasoned Action (TRA), Theory of Planned Behavior (TPB), and Integrated Behavioral Model (IBM) and also describes and explains the main constructs in the TRA, TPB, and IBM. What’s more, it explains the similarity between these theories' key constructs and constructs from other behavioral theories. And it describes measurement of the key constructs of TRA, TPB, and IBM and also explains how and why elicitation should be conducted to identify and select the content for the model construct measures for the health behavior and population studied. In the second paper, it develops persuasive technologies to encourage individuals to live healthy lifestyles with a focus on being physically active. As part of the investigations, they have explored individuals’ reactions to goal setting, specifically goal sources (i.e., who should set the individual’s goal) and goal timeframes (i.e., over what time period should an individual have to achieve the goal). In this paper, it also presents findings related to various approaches for implementing goal-setting in a persuasive technology to encourage physical activity.

Shiry Ginosar - 10/17/2011 2:15:41

Skipping this time (1/3).

Peggy Chi - 10/17/2011 2:25:04

These two papers discussed the persuasive technology and behavior change. Montano and Kasprzyk introduced the models and theories of behavior change. Consolvo et al. presented UbiFit and discussed the importance of goal setting.

It seems applying computer technology is a very effective method of shaping behaviors in large scale. Moreover, different from traditional methodology, using devices such as sensors and mobile phones is more economic. People interact with them most of the time in daily lives. However, I still have doubts about the theories. For example, setting goals helps users be aware of themselves and make progress, but how long the effect could be? Does it apply when there is only one clear goal? What if there are many different goals to care about at the same time? Back to the previous discussion about evaluation, I can hardly see how a system can be evaluated effectively in short term to claim the success. But in general, I believe it's possible to shape behaviors intentionally using technology. It's just the matter of degree of change.

Rohan Nagesh - 10/17/2011 2:27:35

The first excerpt, Chapter 4 from Montano and Kasprzyk, discusses two frameworks in the Theory of Reasoned Action and the Theory of Planned Behavior that enable psychologists to model and predict human behavior. The second paper "Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity" discusses the importance of goal-setting in the art of persuasion and one particular application known as Ubi-fit that has seen success.

The TRA and TPB frameworks are oversimplifications as with all theories and frameworks used to model complex behavior. For instance, TRA models behavioral intention as a function of attitudes and subjective norms. One might also want to take into consideration past history of decision making, any specific extenuating circumstances, but of course these are difficult to quantify and make the modeling hard. I enjoyed reading about the Zimbabwe case study as I believe that was a great application of the these frameworks and the potential modelling benefits.

The second paper is something relevant to my Capstone project as my team develops a glucose monitoring and predictive solution for diabetics. I do believe goal setting is very important to persuasion. For instance, I have stopped using ASAP in team assignments and instead include a specific date and time I want a certain assignment. I agree with the paper that without goal setting, our goals are too open-ended, and that's when we many times end up not accomplishing anything.

Suryaveer Singh Lodha - 10/17/2011 3:10:32

Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity -

The one thing I liked about this paper was the way author included user comments on the experiences of goal setting. It is but obvious that goal setting with feedback is more effective than just plain goal setting. i liked reading about the different viewpoints people had related to setting goals in groups, as opposed to individual goal setting. Also, it was interesting to note that people do not necesarrily approach/consult their medical doctors for directions related to goal setting/ exercise. I share this viewpoint with on of the subject, that if the doctor himself is not in shape, probably he is not the right person to seek for advice! comparison of a rolling 7 day window with a fixed 7 day window was an interesting read and gives insight into how people think about tasks/goals and how critical can setting a deadlines can be for mativation to accomplish goals. I thought the length of the experiment (3 months) was pretty decent for such a study.

Theory of reasoned action, theory of planned behavior and the integrated behavior model - The components of Theory of Reasoned Action (TRA) are three general constructs: behavioral intention (BI), attitude (A), and subjective norm (SN). TRA suggests that a person's behavioral intention depends on the person's attitude about the behavior and subjective norms (BI = A + SN). If a person intends to do a behavior then it is likely that the person will do it. Behavioral intention measures a person's relative strength of intention to perform a behavior. Attitude consists of beliefs about the consequences of performing the behavior multiplied by his or her valuation of these consequences. Subjective norm is seen as a combination of perceived expectations from relevant individuals or groups along with intentions to comply with these expectations. The theory of reasoned action provides a relatively simple basis for identifying where and how to target consumers' behavioral change attempts. Theory of Planned behavior (TPB) is an extension of TRA which involves the addition of one major predictor, perceived behavioral control, to the model. This addition was made to account for times when people have the intention of carrying out a behavior, but the actual behavior is thwarted because they lack confidence or control over behavior. The theory of planned behavior specifies the nature of relationships between beliefs and attitudes. According to these models, people's evaluations of, or attitudes toward behavior are determined by their accessible beliefs about the behavior, where a belief is defined as the subjective probability that the behavior will produce a certain outcome. Specifically, the evaluation of each outcome contributes to the attitude in direct proportion to the person's subjective possibility that the behavior produces the outcome in question. The Integrated Behavioural Model (IBM) - Most important behaviour (in all 3 models) is intention to perform. Behavioural intention is determined by: Attitude toward the behaviour, Perceived norm, Personal agency. To design effective interventions to influence behavioural intentions, one must determine degree to which intention is influenced by attitude, perceived norm and personal agency. TRA, TPB, and IBM provide a basis to evaluate behavior change interventions because they lead to hypotheses about how an intervention targeting a set of beliefs will affect the model component to which those items belong (for example, attitude) and thereby affect intention and behavior. Very different behavioural, normative, efficacy, and control beliefs affects one’s intention to engage in different behaviours; between behaviours, relevant behavioural beliefs are likely to be very different; behavioural beliefs can be different between different populations. In order to analyze these behaviours, one must actually go to the population for full understanding of their behaviours. Interviews of the population are highly essential at this point. This process allows the IBM, TPB and TRA to be highly applicable to all cultures.

Donghyuk Jung - 10/17/2011 4:03:56

  • Theory of Reasoned Action, Theory of Planned Behavior, and the Integrated Behavioral Model

According to the article, “the Theory of Reasoned Action (TRA) and the Theory of Planned Behavior (TPB) focus on theoretical constructs concerned with individual motivational factors ad determinants of the likelihood of performing a specific behavior. TRA and TPB both assume the best predictor of a behavior is behavioral intention, which in turn is determined by attitude toward the behavior and social normative perceptions regarding it.”

TRA can be represented as B ≈ BI = w1AB + 2w2SN (B: behavior, BI: behavioral intention, AB: attitude toward behavior, SN: subjective norm, and w1 and w2 are weight representing the importance of each term). The model has some limitations including a significant risk of confounding between attitudes and norms since attitudes can often be reframed as norms and vice versa. A second limitation is the assumption that when someone forms an intention to act, they will be free to act without limitation. In practice, constraints such as limited ability, time, environmental or organizational limits, and unconscious habits will limit the freedom to act. TPB attempts to resolve this limitation. (From

TPB can be represented as BI = (W1)AB[b + e] + (W2)SN[n + m] + (W3)PBC[c + p] (b: the strength of each belief, e: the evaluation of the outcome or attribute, n: the strength of each normative belief, m: the motivation to comply with the referent, PBC: Perceived Behavioral Control, c: the strength of each control belief, p: the perceived power of the control factor, W: empirically derived weights). TPB is based on cognitive processing and level of behavior change. Compared to affective processing models, TPB overlooks emotion variables such as threat, fear, mood and negative or positive feeling and assessed them in a limited fashion.

IBM emerged from efforts to integrate constructs from TRA/TPB with constructs from other important theories of behavior.

  • Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity

In this paper, the authors showed findings from a three-month field study of the UbiFit System that relate to participants’ reactions to various approaches for implementing goal-setting in a persuasive technology to encourage physical activity. Although UbiFit is developed to encourage individuals to self-monitor their physical activity and incorporate regular and varied activity into everyday life, I think that Groupware or PIM such as outlooks can utilize this concept in order to encourage individuals in doing their hardest. The results that most participants would prefer to set their goal themselves or work with a fitness expert to set a goal are a perfect fit for workplaces. Major features of UbiFit used to encourage participants might be well applied to modern companies.

Sally Ahn - 10/17/2011 4:08:02

Theory of Reasoned Action, Theory of Planned Behavior, and the Integrated Behavioral Model, Chapter 4 from 'Health Behavior and Health Education', Karen Glanz et al. editors, Jossey Bass, 2008. Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity. Sunny Consolvo, Predrag Klasnja, David W. McDonald, James A. Landay. In Proceedings of Persuasive 2009.

The first reading introduces psychological models of influences on behavior: the Theory of Reasoned Action, Theory of Planned Behavior, and Integrated Behavioral Model. I found it interesting to see how something as vague and complex as "behavior" can be reduced into a hierarchy of smaller components. I agree that such models are helpful in analyzing individual factors of a particular behavior for the purpose of influencing that behavior. At the same time, however, I felt that the metrics for these factors (5- or 7-point bipolar scales) were somewhat arbitrary; I would have liked to read more about the reasoning behind the chosen system of metrics.

As for the frameworks themselves, they form a reasonable model in which the components complies with common sense as valid influences of behavior. I don't quite see the benefit of separating the model into TRA, TPB and IBM (perhaps it is just for historical reasons?), since the additional influences that TPB and IBM identifies seem just as important and generalizable for any behavior that one may want to analyze. Finally, since "behavior" is such an abstract concept, I would have liked to have seen more than just one concrete example (Zimbabwe) of the application of these models.

Consolvo et. al.'s paper analyzes goal-setting in a specific context of physical activity. In particular, I find it interesting that most participants opposed the group-set model because they would be upset if they failed to meet their goal because of another group member. However, this response reflects the individual's speculation on a hypothetical scenario, which, during the short time frame of a personal interview, may not be accurate. The responses gathered from real participants provide some insight into designing persuasive technology, but I think the method in which these responses were collected (straightforward interviews) and the small size of the participant pool reduce the reliability of the data.

To illustrate, the response to the group-set model seems reasonable, and I was first inclined to agree. However, after some reflection, I began to disagree. Group-set goals, in this context, seems to me like a contrived "helper" goal whose sole purpose is to boost motivation for the ultimate goal--the individual "my part" of the group's goal. Hence, doing "my part" would be rewarding enough; even if the group failed to meet its contrived goal because of another member, it would not be able to diminish the fact that I had attained my personal--and the ultimate--goal. Of course, the best scenario would be one in which the group and I both attain our goals. However, the impact of the group's final success or failure wouldn't be significant enough for me to reject the positive effect the group goal has already had on my motivation to attain my personal goals. I had to think a bit to reach this conclusion, which makes me wonder if the participants in the study would also have responded differently if they had been given more time to think about the questions. I'm a little skeptical about the conclusions drawn in this paper because they seem to be based largely on responses from interviews.

Apoorva Sachdev - 10/17/2011 7:07:06

Reading for Behavior Change:

The readings for this week were a Chapter from the book, “Health Behavior and Health Education” by Karen Glanz et al that covered the theory of Reasoned Action(TRA), the theory of Planned Behavior (TPB) and the Integrated Behavioral Model(IBM) which is an amalgamation of TRA/TPB frameworks. The second reading was a paper on “Goal Setting Considerations for Persuasive Technologies that Encourage Physical Activity” by Sunny Consolvo that discusses how setting attainable goals is important for instigating behavior change.

The first reading was interesting because it describes the frameworks that can be used to identify the various factors that affect behavior (implicitly and explicitly). The allow researchers to pin-point critical belief targets and hence help them design persuasive methods to change those beliefs. I liked the fact that they described the Zimbabwe example but I was a little uncertain about the truth of their findings. Since the issue was a sensitive subject how does one ensure that the samples being interviewed/surveyed are likely to provide true opinions. It is easier to test people’s awareness about issues through surveys and interviews but accurately identifying factors that influence their behaviors with respect to using birth controls/condoms may not be possible.

The second paper was about an implementation of an App called UbiFit Garden that allows users to set goals of physical activity and also provides visual feedback for the user. The main idea brought about by the paper was that for people to change their behavior one has to provide them unambiguous, attainable goals and also provide incremental rewards(to prevent them from giving up). Although they have a whole section on alternative ways of performing the study they only tested one way of performing the study on the people and instead asked them about how they felt about the alternate ways at the end of the test period. In my opinion, it might have been more effective to break the sample into groups and have each group test some variance of the app to get an accurate opinion about the goal-setting in a system. Interaction with a technique and exposure to it for a while definitely skews the responses.

Allie - 10/17/2011 8:17:12

In "Theory of Reasoned Action, Theory of Planned Behavior, and the Integrated Behavioral Model", Montano and Kaspryzk introduce the development of TRA (Theory of Reasoned Action), TPB (Theory of Planned Behavior), and IBM (Integrated Behavioral Model). TRA and TPB are theoretical constructs that can be used to describe individual motivations that determine the likelihood of performing a specific behavior. Both assume the best predictor of behavior is intention, determined mostly by attitude and social norms. TPB further extends in TRA and has an additional component of perceived control over performance of the behavior. Studies have shown that changing TRA or TPB constructs lead to changes in behavior, and can be used to predict and explain a range of health behaviors including smoking/drinking/health services use/exercise/sun protection/etc cetera.

Operationalization of TRA is determined by expectations concerning attributes of the object or action. The paper proposes IBM as an extension of the TRA and TPB. TRA and TPB assume a csasual chain that links certain behavior beliefs, normative beliefs, and cotronl beliefs to behavioral intentions and behavior via attitudes, subjective norms, and perceived control. A particular behavior is most likely to occur if 1) a person has a strong intention to perform it and the knowledge and skill to do so, 2) there is no serious environmental constraint preventing performance, 3) the behavior is salient, and 4) the person has performed the behavior previously.

Although TRA/TPB/IBM are "Western" concepts, the elicitation process is what makes it applicable to other cultures.

IBM has been used to better understanding of the intentional/behavior surrounding condom use and HIV/STD-prevention behaviors. It was used to identify issues in ways TRA/TPB cannot. It provides a theoretical basis to understand behavior/identify beliefs to target.

In the cross-cultural application of IBM in Zimbabwe, the behavioral focus is using condoms all the time with steady partners in the next three months.

In "Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity", Consolvo et al discuss goal-setting as an effective strategy for changing behavior. In their investigations, they explore individuals' reactions to goal-setting, resources, and timeframes. The UbiFit system was used to evaluate 28 people over 3 months during the winter holidays.

The UbiFit is a mobile, persuasive technology developed to encourage people to self-monitor their physical activity an incorporate activity into their daily lives. Since it is a mobile application, it can run whenever/wherever. The findings found that people give the highest level of effort/performance to the highest/most difficult goals. Further, "specific, difficult goals consistently led to higher performance than asking people to do their best".

The findings concluded that the calendar week option was found to be the most popular, with Mondays/Sundays as preferred start days, and reset at the end of the week. Most participants would prefer to set goals themselves.

Jason Toy - 10/17/2011 8:22:09

Theory of Reasoned Action, Theory of Planned Behavior, and the Integrated Behavioral Model

This paper is about three new frameworks for determining individuals' actions based on their beliefs and possibly influencing their future actions. The frameworks provided allow researchers to determine what beliefs are most important to an audience. According to the paper, a person's propensity to doing an action depends on their attitude towards the behavior (Do they think that their action will bring about the intended result, and if this result is positive or negative.), the perceived norm (The social pressures based on friends: how your friend feels about your action and how interested you are in pleasing your friends.), and personal agency (How you feel the environment either does or does not support your cause, and how confident you feel that you can accomplish your goal in the face of obstacles and challenges.) The ideas in the paper are similar to the real life techniques used by brand marketers. Marketers try to target a segment of the population that would be receptive to the attributes of the product they are trying to sell. Both behavior and beliefs, on both life and the products themselves, of the target audience, are important to marketers and they go to great lengths: using surveys, questionnaires, buying information, and focus groups to try and understand consumers. The research done here could be useful in understanding customers further.

The paper does a good job discussing the cultural and behavioral differences that may occur in different audiences. However a problem with the paper is the fact that the only proposed method of obtaining information is through the use of surveys. In the Zimbabwe example, what if there was a social stigma against talking about sex related subjects? How does one deal with such a large scale problem that could reduce the accuracy of your findings? In addition, how do we deal with the problem of segmentation? Given the results of the survey, do we generalize and say all people in Zimbabwe feel the same way, and thus we should apply the findings universally through the country? Maybe we accidently segmented by geographic region, and such findings should apply to only to the northern regions of Zimbabwe. The problem of the case study is that it is so focused on the surveying, that a reader might lose the main ideas of attitudes towards behavior, and start wondering about the techniques used to obtain them. I also think that it is a negative of the paper that the authors' ideas are split across three different frameworks, as it made it more confusing. Each framework goes hand in hand, with the authors acknowledging that TPB adds only the idea of control factors, and IBM being an extension of TRA and TPB.

Goal-Setting Considerations

This paper is about using goal-setting in persuasive technologies to encourage people to change their behavior.

The paper presents a new idea of using goal-setting to persuade users to be more fit in the UbiFit system. This idea is similar to that of the quantified self: using self-tracking through devices like computers or mobile phones to record behavior, process data, and feed it back to the user. This allows us to better understand patterns and adapt behavior more intelligently. An example of this is which allows users to track their finances. Another example is a smartphone application written by U.C. Berkeley students that allow a user to track their methods of transportation (car, bus, bike, foot) and give the user a measure of their transportation sustainability (low-environmental footprint transportation) through CO2 saved among other measures. The use of smartphones, which we can carry with us definitely allows people to learn more about themselves wherever they go.

The creators of UbiFit did a good job exploring the various possibilities in regards to goal setting (self-setting, group-oriented, medical expert, and fitness expert) and goal-timelines (weekly, rolling seven-day window, customizable week). However the paper does not discuss the idea of personalization, which is the cause of many of the participants' comments. Participants' negative comments about "experts" setting their goals was because of the lack of personalization and feedback. One participant said that he would never use his doctor's goals because they would just look like the American Heart Association's goals. Another participant said that he/she felt that they should be able to talk to their fitness trainer if the schedule wasn't working out, and have the schedule tailored to their own fit. One user pointed out that in general, her professional career mattered more than fitness and that if necessary, she would slack off on her fitness schedule. Finally, standardized schedules such as the American Heart Association's, were universally disparaged. It might be that it is not the fact that people are setting their own goals, that they knew they could accomplish, as according to the paper, that resulted in success of the experiment, but it is the fact that in this experiment, they could personalize their own schedule easier than if they worked with an expert. The paper acknowledges this point only in a simple sentence in the conclusion that "Most participants would prefer to set their goal themselves or work with a fitness expert to set a goal, in which case they would expect an ongoing relationship with the expert."

Vinson Chuong - 10/17/2011 8:32:35

Montano and Kasprzyk's "Theory of Reasoned Action, Theory of Planned Behavior, and the Integrated Behavioral Model" offer an empirical model describing the factors and beliefs commonly known to influence or determine health-related behavior. Consolvo, Klasnja, McDonald, and Landay's "Goal-Setting Considerations for Persuasive Technologies that Encourage Physical Activity" explores user reaction to setting different types of goals with regard to increasing physical activity.

Drawing from a wide variety of behavioral models, like the Theory of Reasoned Action, the Theory of Planned Behavior, and others, Montano and Kasprzyk presents an Integrated Behavioral Model which attempts to provide as wide a range as possible of possible useful causes of health-related behaviors. Through past empirical studies and consensus in the research community, they assert that knowledge, salience, intention, environmental constraints, and habit are the primary determinants of behavior. Moreover, they assert that attitude, perceived norm, and personal agency are the primary determinants of intention, and that each of those are determined by various types of beliefs. They go on to describe studies in which this model has been useful in identifying the beliefs underpinning behaviors which are targeted for change and discuss how changing these beliefs can change behavior. Hence, this model offers a framework for identifying the target beliefs relevant for affecting behavior change in health interventions.

Consolvo, Klasnja, McDonald, and Landay looks at how goal setting can affect the beliefs of users with regard to increasing physical activity. In particular, they examine the effects of different goal sources, goal timeframes, and group interactions. The effects of each of these factors can be described using the IBM and compared by looking at an instance of the model before the change and an instance after the change.

Here, we see that, although the model is useful for describing the contribution of various beliefs to behavior, it doesn't say much about the interaction of those beliefs with each other or with interventions designed to change them. Do such elements belong in this model?

Another question is, does this model generalize? Can it describe other scenarios? What "Other Factors" would come into play?

Derrick Coetzee - 10/17/2011 8:35:51

Both works concerned motivating behavior change in the particular area of health and exercise. The first was a textbook chapter describing theory of behavior change, and the second a field study of a phone-based support and motivation system.

The first reading was a chapter out of health behavior book by Montaño and Kasprzyk describing well-known theories of behavior (TRA, TPB, IBM) based on meta-analyses of large numbers of studies. TRA assumes direct control over behavior and emphasizes beliefs about outcomes, while TRB emphasizes perceived behavioral control in scenarios where agents have reduced influence over results. IBM emphasizes specific underlying beliefs and their role in influencing behavior. The authors have direct experience in behavior change via longitudinal studies of HIV prevention.

Although these systems provide an effective framework, they depend on being able to "elicitate" specific feelings of efficacy or underlying beliefs, a process that may be stymied by both the subjects' personal lack of awareness about these factors and reluctance to disclose them. Such subjective appraisals should at least be evaluated against objective data (for example, the idea that trust is a factor in condom use could be evaluated by comparing general condom use to condom use with prostitutes).

"Goal Setting Considerations for Persuasive Technologies that Encourage Physical Activity" was a 2009 field study by Consolvo et al regarding how goals were set in their phone-based exercise-encouragement system, and how else they might be set. The system uses a garden metaphor, showing the garden growing over the course of the week as exercise goals are met.

Some of the noted effects in goal-success may be correlational: for example, it's natural that users are more successful at goals they believe they can complete, provided they are good at predicting their own performance. Similarly, a user may be reluctant to make a goal public that they're not confident they can achieve. Correlation is sufficient when performance prediction is the goal, but not necessarily for behavior change.

The use of incentive systems to motivate behavior is one that has been thoroughly explored in the domain of video games, and even physical exercise has borrowed from these models (e.g. Fitocracy), and in fact the flower- and fish-based have many characteristics of games like popular virtual pets, which make me wonder why the paper did not mention this related work. It would be interesting to explore how effective game design relates to success of a motivational system.

The paper extensively quotes speculation of subjects on what goal sources would be most effective, which seems unpersuasive to me since it is both subjective and by unqualified persons. Similarly, although multiple goal timeframes were available, they were self-selected. These are partly limitations of being a field study, but even in this context more choices could have been mandated by the study.