Fakulteten för teknik och samhälle
Datavetenskap
Examensarbete
15 högskolepoäng, grundnivåTeaching Maternal Healthcare and Nutrition in Rural
Ethiopia through a Serious Game
Alexander Hedvall
Emil Svensson
Examen: Kandidatexamen 180 hp Handledare: Jose Font Fernandez Huvudområde: Datavetenskap Examinator: Carl-Magnus Olsson Program: Spelutveckling
Teaching Maternal Healthcare and Nutrition
in Rural Ethiopia through a Serious Game
Alexander Hedvall
- Game Development Program, Computer Science and Media Technology, Malmö University, Sweden.Emil Svensson
- Game Development Program, Computer Science and Media Technology, Malmö University, Sweden.Abstract
Ethiopia has one of the highest maternal and child mortality rates in the world. The main reason for this is the low accessibility of educational content regarding healthcare in the rural regions. In this thesis we investigate the possibility to address this issue through a serious game, something that has proven efficient as a tool for learning in many situations. By designing and creating a game for rural Ethiopians, we aim to interest them in learning about healthcare and nutrition by playing a game that reflects their culture and life. In order to validate our solution we conducted a field study in Ethiopia, where we organized and carried out a usability test with Ethiopian mothers. Our findings suggest that educational games has a potential in developing countries.
Keywords
Computer science; Game development; Serious games; Healthcare; Computer assisted education; Ethiopia.
1. Introduction
In Ethiopia about 4% of the country’s women die during childbirth or during the postnatal period. Furthermore, one in 17 Ethiopian children dies before their first birthday and one in 11 Ethiopian children dies before their fifth birthday(Central Statistical Agency, 2011). This results in Ethiopia having one of the highest maternal and child mortality rates in the world (Hogan et al., 2010). The reason behind this are the rural regions of Ethiopia, where home birth without proper maternal care is still very common, with as little as 10% of women delivering at a healthcare center and/or with professional care (Roro et al., 2014).
Addressing these issues is crucial to be able to meet the United Nations’ sustainable development goals (United Nations, 2016). Particularly SDG3, which aims to ensure healthy lives, and promote well-being for all at all ages, and SDG5, which aims to achieve gender equality, and empower all women and girls.
Reaching out to these women via digital applications is something that might be difficult to achieve, since a large portion of the population in Ethiopia do not possess the level of literacy needed to understand written content(Bekele et al., 2017). Therefore a serious digital game, without text elements could be a viable way to convey the information to an illiterate player.
In 2011 the Central Statistical Agency (CSA), in collaboration with UNICEF and USAID among others, conducted a study in rural Ethiopia (Central Statistical Agency, 2011). Their report (see table 1.1) shows a declining trend in child mortality rates across three successive five-year periods. Even though the data points at a positive trend, Ethiopia still suffers from one of the highest child mortality rates in the world. CSA’s report also examined maternal mortality rates in Ethiopia, among women ages 15-49. The data includes a seven-year period preceding the study (see table 1.2). Categorized by five-year groups, the report shows that the maternal mortality rate increases with age, until reaching its highest rate at ages 30-39.
Table 1.1 Early childhood mortality rates, expressed in deaths per 1000 children Years preceding the survey Neonatal mortality Postneonatal mortality Infant mortality Child Under-5 0-4 37 22 59 31 88 5-9 48 40 88 49 133 10-14 54 47 101 72 166
● Neonatal mortality: The probability of dying within the first month of life.
● Postneonatal mortality: The probability of dying after the first month of life, but before the first birthday (The calculated difference between infant and neonatal mortality).
● Infant mortality: The probability of dying before the first birthday.
● Child mortality: The probability of dying between the first and fifth birthday.
● Under-5 mortality: The probability of dying between birth and fifth birthday.
Table 1.2 Maternal mortality rates
Age Percentage of deaths that are maternal
Exposure Years Maternal mortality rates* 15-19 21.7% 39 187 0.52 20-24 36.1% 41 289 0.94 25-29 28.7% 36 532 1.03 30-34 36.7% 27 449 2.53 35-39 36.7% 19 166 1.53 40-44 18.8% 11 512 1.43 45-49 9.8% 6 899 0.70 15-49 29.6% 182 034 1.14
* Expressed per 1000 years of exposure
1.1. Related Research
Previous research(Bekele, et al., 2016) suggests using an information and communication technology (ICT) solution that puts emphasis on design and development, in order to make health information more accessible in rural Ethiopia. Furthermore (Bekele et al., 2017) identifies the challenges and opportunities of developing technology designs for rural Ethiopia, it also states that rural women generally do not have access to educational content locally, nor do they have the level of literacy to read and understand written health content.
Another challenge identified by Bekele et al. (Bekele et al., 2017) is the cultural challenge of implementing an ICT solution in the region. This is because the majority of the women choose to ignore information regarding antenatal care, postnatal care, diet, and hygiene. It was also observed that there was a strong1 bias towards males, whereas women had a low social value. This resulted in a large portion of women being reluctant to openly talk about family problems in fear of failing their husbands. Furthermore, most women did not own their own mobile phone, or in some cases they had never touched one.
We believe that the information needed to achieve a better understanding of the importance of healthcare and proper nutrition, can be taught through a serious game, designed specifically for use by rural Ethiopian women. According to (Susi et al., 2007), serious games have as many definitions as there are actors involved, but the core aspects remain the same, in that serious games are digital games, whose purposes go beyond just entertainment, because they also teach the user something.
Countering this problem with a serious (digital) game will require a lot of testing and iterations during development of the game, and since our possible users are mostly illiterate, the game has to be designed without text elements.
1.1.1. Examples of Other Serious Games
Figure 1.1 - Get h2o, digital browser game
Get h2o: In Get h2o (Figure 1.1) the user plays as a city manager, managing house building, work, water supply, and pollution in a small city area (Butterfly Works, 2010). For every level the amount of buildings you need to manage is increased. The game aims to teach about the difficulties in having access to clean drinking water in africa. This is done by increasing the time it takes to get water, the further away the water source is from the buildings. Furthermore, the water pollution increases with the increase in buildings. Making clean water harder and harder to come by. Finally, the game cites informational facts about water in africa in between levels.
1The postnatal period is the period beginning immediately after the birth of a child and extending for about six weeks.
Figure 1.2 - Darfur is Dying, digital browser game
Darfur Is Dying : In this game (Figure 1.2) the user plays as refugees in the Darfur region of Sudan (TAKE ACTION games, 2006). The user must keep the refugee camp functioning by fetching water, and doing certain activities in the camp to help. The game teaches you about the hardships of refugees in africa, specifically it aims to teach more about the genocide in Darfur. The game aims to teach the user through experiencing the game through the viewpoint of refugees, but it also leans on informational messages via text shown to the player throughout the game.
1.1.2. Android Games for Children
Popular children’s games for Android typically feature clear, animated, lively game elements, and little to no focus on the use of text as a central game feature. If one looks at a top list (Hindy, 2016) of this type of game, it is easy to identify the specific design. Mostly the games are colourful, feature user interface (UI) animations and a clear user experience. To enable the game to teach and engage illiterate adults, many of the aspects of children’s games and serious games will be mixed. The specifics of children’s games design will be incorporated in the application developed alongside this document.
1.2. Purpose
The purpose of this project is to apply previously successful game learning mechanics to a serious digital game, focused on teaching hygiene and antenatal and postnatal healthcare. This will be done by examining other previously tested and evaluated mechanics and applying them to a mobile game. The result will be a digital game for Android (Google, 2008), whose aim is to improve the understanding and importance of healthcare and nutrition for the end user. By improving the knowledge of healthcare it aims to reduce maternal and child mortality rates in rural Ethiopia, a country which has one of the highest maternal and child mortality rates in the world (Hogan et al., 2010).
We will also investigate how to adapt the game to fit Ethiopian culture and the high level of illiteracy that rural Ethiopia has(“UNESCO,” 2014). The illiteracy will be addressed by not using text as a central element in the game, and using elaborate artwork, icons and animations instead. Game elements from children’s games will also be incorporated in the game, in order to keep it simple, since the end users are not used to
playing digital games(Bekele et al., 2017). Therefore, another aim is to find out if serious games can be a feasible way to learn, among illiterate users.
1.3. Questions - This thesis will answer the following:
● How can a serious game be used as an educational platform to teach ante-/postnatal healthcare and nutrition among illiterate users?
○ How should a modern game design be adapted to provide a compelling experience for rural Ethiopian users?
○ How can a game design similar to children’s games be used to effectively interest illiterate adult players to enjoy it?
1.4. Target Medium
In this thesis we will work specifically with the region of rural Ethiopia, a region that has a strong tradition of home birth, reliance on traditional birth attendances, and misconceptions about the services provided by the modern healthcare system(Roro et al., 2014). Severe malnutrition among children is another important problem in the region, where the women do not know the importance of feeding her child a balanced and nutritious diet (Christiaensen & Alderman, 2004).
Figure 1.3 - ICT Application
Previous studies have attempted to address this by identifying new ways to educate the population concerning maternal healthcare and nutrition, through the use of information and communication technology(Bekele et al., 2016). This research showed cultural differences, e.g. the application would show a picture of how a bed looks in western society, but the Ethiopian testers did not identify the image as a bed, as their idea of what a bed looks like is different, see bottom right of figure 1.3. But since there has been no previous research done with serious games in rural Ethiopia, the game will be tested by Ethiopian women, in Ethiopia. This is to assure that the quality is good enough that they would use the game and most importantly, understand its message. Since an ordinary serious game usually provides information via some form of text (Debow, Paharia, & Werbach, 2013), and because of the high illiteracy rates (“UNESCO,” 2014), we will have to find ways the game would work, without the use of text.
1.5. Expected results
Our contribution will be a field study which aims to investigate if serious games can be used to improve knowledge of healthcare and nutrition in rural Ethiopia. A successfully designed artifact will hopefully prove games as a new platform for learning in rural Ethiopia, but how it compares in terms of effectivity
and knowledge improvement with different learning platforms is something that is out of the scope of this project, and requires further research.
2. Method
The formal methodology of the research will follow the well established design research methodology by Peffers et al. (Peffers et al., 2007). The methodology breaks down the design process into the following six iterative steps. Figure 2.1 illustrates a visual overview of the iterative steps,where the boxes are project steps, and the arrows shows the flow of steps with the option to reiterate to previous steps. Detailed descriptions of each step are as follows (Peffers et al., 2007):
1. Problem Identification and motivation: Define the specific research problem, and justify the
value of a possible solution. The problem definition will shape the solution, so atomizing the problem may be useful in order to find a possible solution.
2. Define the Objectives for a Solution: Here one should infer the objectives of a solution from the problem definition established in step 1, with knowledge of what is possible and feasible. The objectives can be quantitative or qualitative.
3. Design and Development: During this step the artifact is created. This activity includes determining the artifact’s desired functionality and its architecture and then creating the actual artifact.
4. Demonstration: Demonstrate the use of the artifact to solve one or more instances of the problem. Could be by using it in experimentation, simulation, case study, proof, or other appropriate activity.
5. Evaluation:This activity is used to observe and measure how well the artifact supports a solution to the problem. This involves comparing the objectives set in step 2, with the gathered results from where the artifact was used in step 4.
6. Communication: The final step is communicating the problem and its importance, the artifact, its utility and novelty, the rigor of its design, and its effectiveness to relevant audiences.
2.1. How the Method was used during the Project
2.1.1. Problem Identification and Motivation
Ethiopia has one of the highest maternal and child mortality rates in the world (Hogan et al., 2010). The reason for this is the lack of education and low access to educational content (Bekele et al., 2017), this is also related to the low literacy level of the population (“UNESCO,” 2014). Thus we decided on a solution focused on teaching through descriptive icons and symbols, using an ICT solution that can easily be accessible by the target users, without the requirement of transport and/or teachers.
2.1.2. Define the Objectives for a Solution
Our objective is to find a way to convey the required knowledge in a compelling and interesting way. However since the level of literacy is so low, including text elements in our way of teaching would exclude more than half of our target group. And therefore another objective will be to convey this knowledge in a way that can be understood without the need of reading and/or writing.
2.1.3. Design and Development
Using Unity Engine(Unity Technologies, 2017) we created a digital game for Android devices without any text elements, instead focusing on descriptive icons and graphics that the user can associate with different objects and actions. In the game you take on the role of a pregnant woman, from early pregnancy to successful or failed delivery. The game goals consist of keeping her healthy through visits to medical centers, feeding her a nutritious, balanced diet, and also managing her water supply and happiness.
2.1.4. Demonstration/Test
Because our project is centered around a specific group of users we decided to verify our solution with a case study conducted in Ethiopia. During this study we are going to have 5-10 Ethiopian women test the game. Prior to a testing session, every tester will fill out a questionnaire (Appendix A). After playing the game, another questionnaire will filled out (Appendix B). After the second questionnaire, a short interview will be held. In order to extract the most useful data out of the interviews, an interview structure centered around key questions must first be established. These key questions should examine the following: interpretation of the visuals and UI presented in the game, interpretations of the game's goal, how satisfactory the game was to play, and what the user learned playing the game. For the specific questions asked during the interview, see appendix E.
Observations will be done by adapting Ismail et al’s method of measuring usability of educational games (Ismail et al., 2011) . Observations will with the aid of two checklists (see Appendix C and D), consisting of tasks the users do in the game, measure the following:
● Effectiveness - the capability of the user to complete a task within the application.
● Efficiency - measuring for the smoothness in completing a task.
Depending on the users ability to complete a task, it will be marked as; successful, partially successful or unsuccessful. This will give us a way to measure the usability and viability of our game design. To ensure documentation of any events during the game sessions, detailed notes about the tester’s interactions, successes and failures, will be taken.
2.1.5. Evaluation
During the evaluation phase of the testing we will look at the data gathered in the demonstration phase, and analyze why certain aspects of the game were or were not understandable for the user, and if the aspects fulfilled their purpose in the game. Using (Ismail et al., 2011), we will evaluate the following:
● Learnability:How easy is it for the player to accomplish tasks in the game when the design is first presented. E.g. how easy is it for the user to understand that she can click on the button with food in order to select a meal and drag it to the woman in the game to reduce her hunger?
● Efficiency: Once the player have gotten to know the design, how quickly can she perform the game’s tasks?
● Memorability:Since our testing period is limited (6 days) we will not have time to test the game for memorability, and this step will be skipped.
● Errors: How many errors do players make, how severe are these errors, and how easily can the
player recover from the errors?
● Satisfaction: How pleasant is it to play the game? Does the player think that games could be a good medium for teaching?
Another subject of evaluation is every sprite of significance in the game. Each sprite’s design and look will be evaluated and discussed with the TEMACC Ethiopia team. This is to ensure that the sprites are clearly identifiable, and properly reflects Ethiopian culture. If it is clear to each team member what the sprite is, and what its purpose is, the sprite will be verified and deemed usable in the game. If it is unclear what the sprite is or what its purpose is, the sprite will be discarded/redesigned, and we will reiterate to the design step, and find a new feasible design of the sprite, or rethink how it is used.
2.1.6. Communication
In the final stage, the results of our testing will be documented and submitted, alongside a discussion for further research options in this field of study.
2.2. Method Discussion
When the problem of was first presented to us, we were informed of the previous methods used in order to reduce the high mortality rates, but a game had never been used, instead the focus had been laid on extensive communication of knowledge via mobile digital devices.
Since no previous research specifically concerning games had been done concerning our target audience, we had very little research to use as framework. So we spent the first month getting to know Ethiopian culture and life. We also learned how a typical day of an Ethiopian woman would be. Because we wanted to solve a real problem, identified by previous research with a digital system, we felt that the design research methodology(Peffers et al., 2007) would be a good fit, and would enable us to work iteratively towards building and evaluating a solution towards the problem.
Furthermore, by adapting the evaluation methods used in Ismail et al’s work (Ismail et al., 2011) we are able to quantify our data into simple understandable metrics that measures the software's effectiveness, efficiency, and satisfaction rate. This enables us to measure the viability of our design towards the user, but it also makes it easy for further research in this area to compare its results towards our own.
3. Solution
The solution we present is a design and implementation of a serious game, specifically for use by women in rural Ethiopia. The game’s purpose is to educate players about the importance of a balanced diet, and the benefits of healthcare during ante- and postnatal periods. All of the UI design has taken inspiration from children’s digital games, where UI design usually incorporates large, clear images and animations to go along with them.
The game itself is an application for Android (see figure 3.1), developed with the Unity 5 game engine. Android was chosen as a platform specifically because of its widespread usage in the world (Goasduff &
Forni, 2017). Android is also more accessible on low end phones compared to other mobile operating systems, and since the cost of the phones/tablets used for testing was limited, this too was a great advantage. We chose to develop the game in Unity 5 since it is the latest version of the engine, and both of us have used it to develop several games previously.
Figure 3.1 - Screenshot of the main game screen during night.
Featuring timeline (1), an upcoming food token refill (2) and healthcare event (3). On the left, the thirst (4) and hunger (5) indicators/menu buttons. The game character in the middle (6), next to the water container (7). In the bottom right corner are the happiness (8) and fook token indicators (9).
3.1. Initial Game Design
Initial game design concepts were adopted by looking at the japanese game Tamagotchi (Bandai, 1996). Tamagotchi is a handheld (Figure 3.2) game device where the player is tasked with keeping a fictional creature happy, healthy and disciplined, through actions such as feeding the creature, playing with it, and maintaining its hygiene. The player’s score is counted as the number of days the Tamagotchi creature has lived. If the player neglects taking care of the creature, it will faint and the game has to be restarted.
Figure 3.2 - A Tamagotchi Device (Sienicki, 2005)
From this concept we designed a game with a similar idea, but some changes were applied to fit the needs to solve the specific problem. Because we needed a platform that was easy to work with and already in wide usage we programmed the game as an application for Android. Furthermore, game mechanics such as discipline and hygiene were outside the scope of this thesis, and as such was overlooked.
The original game mechanic that was kept mainly intact from its designed purpose, was keeping the character nourished through eating. This was further developed to include different food groups, with each food item costing one food token. Food tokens replenish each week depending on the character's happiness. The food tokens are a way to regulate how much food the player has available, and requires the player to plan which food type should be distributed. Alongside this, a thirst system was developed, where the player must keep the character hydrated but also fetch water to maintain the household’s water supply.
Health was kept as a game mechanic, but how the character in the game maintains her health was redesigned. Because the application is focused on maternal healthcare, the health system is based on events that the character needs to attend at a healthcare center. These events range from giving birth to regular pre-natal examinations.
We also added a timeline to the game world, featuring a simulated time lapse (figure 3.3). This was added as a constraint to encourage the player to plan what actions they will do each day, with some actions, like fetching water taking up a certain amount of time during the day. The timeline also functions as a planner for when certain events happen, as well as when the next food token handout happens at the start of each week.
Figure 3.3 - The timeline with an upcoming event (2) and food token (1) refill.
The final game mechanic we reviewed was happiness. Happiness was redesigned as a reactive system that is influenced by the other three mechanics (health, hunger and thirst). Failing to maintain your happiness will decrease the amount of food tokens you get the following week.
Unlike Tamagotchi, the game will not progress in the background while the application is closed. This design choice was made because of the limited time available for gathering data, where game sessions spanning several days would be cumbersome to analyze. However, we still designed this part. The final application would send the user a notification, saying that something needs to be addressed in the game, such as the water is close to running out, or a healthcare event is coming up soon.
3.2. Sprite Design
During the design of any sprite to be used in the game, an extensive amount of research was made in order to make sure that the item depicted was something easily recognisable for any Ethiopian. An in-depth example of a sprites design cycle is shown below in 3.2.1.
3.2.1. Example Sprite Design Process
The design cycle of a sprite we would use for the game went through four different phases; The Research Phase, The Rough Design Phase, The Detail Design Phase and The Polish Phase. The four phases of the design of the Food Icon are shown below.
● Research Phase:To be able to create a relevant and relatable sprite, first some research would be made of what dishes are common in Ethiopia. After establishing that Injera (National Dish of Ethiopia, 2017) was one of the most common dishes, some reference pictures were sought up to do a rough design from, one of the pictures used is shown in figure 3.4. The size of the final image is also decided, in order to keep the level of detail needed according to its size on the game screen.
Figure 3.4 - Injera bread with condiments (Rama, n.d.)
● Rough Design Phase: In this phase some rough sketches were created with the digital painting
program, Krita (Stichting Krita Foundation, 2005), each with different designs, with inspiration taken from several of the reference pictures. From these sketches, one was chosen, and developed further. One of the rough sketches is shown in figure 3.5.
Figure 3.5 - Rough sketch of injera sprite
● Detail Design Phase:During this phase a more detailed and refined sketch was made. This one has detailed outlines, and general colours or details are added (Figure 3.6).
Figure 3.6 - Detailed sketch of injera sprite
● Polish Phase: In this phase, final colours and details and shadows are added. During the polish
phase the sprite is constantly be put into perspective in the game in order to make it clearly visible and colourful. The final sprite is shown in figure 3.7.
Figure 3.7 - Injera, finished sprite
3.3. Game Mechanics
The game has five different main game mechanics that will be explained in-depth in the sections below.
3.3.1. Hunger
The hunger mechanic handles everything related to food and nutrition in the game. As time progresses, the game character will start to feel hungry. This is displayed with a progress bar gradually diminishing. (Figure 3.8) .
Figure 3.8 - Hunger progress bar.
If the hunger goes untreated for a long time, the characters happiness will also gradually start to deteriorate. If the hunger bar is completely depleted for an extended period of time, the game will end, as the character faints from hunger. To deal with hunger the character needs to be fed at regular intervals, which can be done by opening the food wheel by clicking the food button. The food wheel contains 4 different food groups, Fish, Meat, Grain, and vegetable (Figure 3.9). Dragging one of the food groups on the player character will feed her, at the expense of a food token.
Figure 3.9 - Food menu (Food Circle).
with the four different types of food, and their corresponding happiness symbols.
Food tokens are a currency for food that replenishes each week with a varying amount, depending on the happiness level of the previous week. This forces the player to manage the allocated food they have for each week to make it last.
To encourage nutritional variation, every food category has a linked satisfaction value; bad, neutral and good. When the player feeds the character with a certain food category the satisfaction gained from eating it decreases by one level, while the other three categories increase by one level. Satisfaction from eating translates into happiness gained or lost depending on the current satisfaction level of the selected category.
3.3.2. Thirst
Thirst manages how thirsty the character is, as well as how much water is stored in the household. The thirst level is displayed in a progress bar that gradually depletes with time. By clicking the icon on the progress bar, two buttons become available, one for drinking water - done by dragging the icon to the character, the other icon allows the player to send the character away to fetch water from a nearby water source, this is done by clicking the corresponding button (Figure 3.10).
Figure 3.10 - Water management menu.
With button used to drag water to the game character (1), and the button used to fetch more water (2).
When drinking water the characters thirst is decreased, and the water progress bar is refilled until full, or the water supply runs out. To indicate the amount of water stored in the household, a jerry can is placed outside the house (Figure 3.11). When the character drinks water from the water supply, the contents of the jerry can container will decrease proportional to how thirsty the character was. If the household supply of water runs out, the player can send the character to fetch more. This consumes a certain amount of time during the day. If the thirst level is left unattended for an extended amount of time, the characters happiness level will also start to decrease gradually until the character is given water. If the characters thirst meter is depleted for an extended amount of time the game will end as the character faints from dehydration.
Figure 3.11 - Household water supply displayed through an in-game container.
3.3.3. Happiness
Happiness is a support mechanic that increases and decreases depending on the characters hunger, and thirst levels, as well as how the player dealt with the various events in the game. Maintaining a high happiness will reward the player with additional food tokens the next week, making it easier to manage. A low happiness level will give a lower food token amount as a punishment, making the next week harder to manage. Care is taken that the weight of this punishment is balanced, so that the player can recover from a bad week, and do better the following week.
3.3.4. Timeline
The timeline manages the cycle of day and night in the game, as well as helping the player keep track of upcoming events, and when a new week begins. Time in the game is artificial compared to real life, with a day in the game only lasting a couple of minutes in real time. Events are displayed on the timeline as icons just below the day of the event, and will remain until that day is over or the player initiates the event. During certain conditions the time will be sped up as to avoid unnecessary waiting. These conditions are: When the character is away fetching water, and during the night. Food and thirst continues to deplete during the night, while only hunger depletes while the character is fetching water.
3.3.5. Healthcare Events
At certain times during the game character’s pregnancy, a small icon symbolizing an event will pop up on the timeline. When the day for these events starts, a window will appear (Figure 3.12), where the player is given two choices. One choice will always be healthcare related, and the other will not be, but both choices still carry a significance to the game.
Figure 3.12 - Event menu.
Showing an ultrasound inspection (1), and a social coffee ceremony (4), as well as their corresponding rewards (2), (5). (3) Shows the drag indicators.
An example of a choice would be that the character now is due for an ultrasound examination, and therefore should go to the healthcare center. This choice will yield the woman a refill of her hunger and thirst, and also give her an amount of food tokens. If the player selects this choice, a silhouette of a pregnant woman (Figure 3.13) will pop up, showing the baby in her stomach. The baby is animated and grows along with the stomach, and will also flash green to indicate its well being.
Figure 3.13 - Silhouette pop up, indicating a healthy baby.
The other choice will be a social event, such as a coffee ceremony, which is a common Ethiopian tradition where people get together to drink coffee and socialize. This choice will yield the woman happiness as well as having her thirst filled up. It will however come at the expense of the baby’s health, since the woman did not attend the hospital visit. This will be indicated by a red flashing color on the silhouette of the baby. Neglecting maternal healthcare too much and the risk of having an unsuccessful birth increases.
4. Results
In this section the results gathered during our field study in Ethiopia will be presented. The tests were conducted in cooperation with TEMACC Ethiopia (TEMACC, 2017), a research project situated in Addis Ababa, which focuses on technology enabled maternal and child healthcare in Ethiopia. Testing was done during three separate days, at three separate locations, and will be explained in detail below in each subsection below. To ensure the best possible results, tests were conducted with developers and supervisors of TEMACC Ethiopia during day one and two. They could then give their feedback and opinions, so that any changes/additions needed before the final test session with the target users, could be put in place in the application.
4.1. First Test Day
The tests conducted during the first day of testing were situated at the offices of Dr. Rahel Bekele in central Addis Abeba. We invited some members of the development team of TEMACC, as well as the project’s supervisors, Dr. Tesfaye Biru and Dr. Rahel Bekele. The goal of this test phase was to introduce our game the research team, hear their views of the game’s possibilities, and any suggestions for further development. We began with a short introduction of ourselves, the game and its purpose. This was followed by a demonstration, where we played through the full game while they were spectating.
After the game demonstration we held a discussion panel where everyone was encouraged to talk about anything that came to their mind, that in some way could be relevant for the project. During this discussion several feedback points arose, and some ideas for the future development of the game were presented and discussed. Some of the discussed items were as follows:
● Incentive to play the game could be given through real life rewards (e.g. food/clothes). ● Incorporating context relevant content in the game.
● If the player performs poorly, the game can communicate this to a health center, which can help educate the player.
After the discussion, The authors of this text decided on some key concerns we should try to address before the tests on the following day. One problem was that it was difficult for a player to perceive whether it was daytime or nighttime. In order to counter this the background of the game changed during the night, to be a nighttime sky, with a moon and stars (see figure 4.1). The game character also disappeared during the night, to imply that she was sleeping in her hut.
Figure 4.1 - Game during night, with character sleeping in the hut (hidden from sight).
Another issue was that it could be hard to understand that the character was hungry or thirsty for players who are not used to playing digital games. To address this we added thought bubbles (see figure 4.2), which would appear when the game character was very thirsty or hungry, to further indicate her need of nutrition and/or hydration.
4.2. Second Test Day
The second test phase was held at Addis Ababa University, with the non-technical personnel of the TEMACC team. The goal of this phase was to identify potential improvements to the data collection process and to the game itself before conducting the final tests in a real environment. This was also an opportunity to test the changes made after the first testing phase in a more controlled environment. Before testing could begin, all the mobile devices of the TEMACC team were collected for game installation. With six different mobile devices collected, along with two devices (one phone and one tablet) supplied on our arrival for testing, we had a total of eight devices to test the game build on. Out of the eight devices the build was installed on, one could not run the game, but the other seven devices successfully ran the game without any issues. During testing of the game, the TEMACC team were encouraged to ask questions, and write down any feedback that they could think of while playing. After gathering the feedback given during and after the game session, coupled with our own observations, some issues became evidently clear:
● The day/night cycles in the game were too short.
● The character’s disappearance during the night was confusing without proper graphical representation.
● The food token mechanic was confused for a score system.
● The food circle, displaying a choice of varied meals was misinterpreted as a tool to create a single meal using different ingredients.
The TEMACC team deemed our method of collecting data to be good enough for the purpose of our testing, and had no suggestions for further improvement of the data collection method.
Following up on this test phase, some changes to the game was made in regards to the feedback received. Regarding the short days and the character disappearing during the night, it was decided to simply leave the character awake and interactable during the night instead of making her disappear. This worked as a time efficient fix towards the days being too short, while also removing the element of confusion with the character disappearing during night time. Solutions implemented related to food tokens and the food circle was due to time constraints less elegant. The decision was made to leave the food tokens and the food circle system as it were for the final testing session. However, a failsafe system was put in place, that would prevent the player from being able to lose the game through starvation, dehydration, or unhappiness. This was to ensure a smooth testing session, without the player having to restart the game multiple times.
4.3. Third Test Day
As mentioned in the previous sections, a lot of care was taken to ensure the best possible data collection process during our visit to the hospital. The game test process was as follows:
● The tester filled out the first questionnaire.
● The game was tested for a period of 10-15 minutes. ● The tester filled out the second questionnaire.
● The tester was interviewed about the test and the game, this interview was recorded with a microphone, and translated.
The location for the third test day was a hospital, located in one of the poorer neighbourhoods of Addis Ababa (Figure 4.3 and Figure 4.4). Due overcrowding inside the facility, tests had to be conducted in a public hallway. This had no major effect on the testing process, except minor interruptions, and a slight loss of quality on the recordings due to background noise. The testing session yielded a total of five complete data sets, acquired from pregnant mothers.
Figure 4.3 - The street outside the hospital
In total there was three people conducting the tests; the authors of this paper acted observers, each with different tasks. Observer 1 filled out the tables specified in Appendix C and D, regarding the effectiveness and efficiency of the game test session. Observer 2 would write notes describing specifics of each tester’s interactions, failures and successes.
None of the testers were expected to understand or speak english, so arrangements were made in order to be accompanied by a member of the TEMACC research team, who would act as interpreter. The interpreter was given clear instructions about how he should act during the game test. The interpreter’s instructions are specified below.
● No information could be given regarding what we wanted them to learn from playing the game. Instead the testers would be instructed that they would “play a game which tries to teach them something”.
● Before the game test, the interpreter would assist the tester in filling out the pre test questionnaire. ● Before the game test, instructions were given to the tester to ask if anything was unclear. If the tester experienced troubles interacting with the games, she would first be given instructions to keep trying different actions and commands. If the tester still struggles, clear instructions regarding what to do should be given.
● After the game test, the interpreter would assist the tester in filling out the post test questionnaire ● Interviewing and recording the tester, following the interview format (see appendix E). If any tester
did not elaborate their answers on their own, the interpreter was free to explain questions, or to ask for specifics about the game, in order to clarify the answers.
4.4. Data Presentation
In this section, the data gathered during the third test day of our field study in Ethiopia will be presented. The data collected is divided into five different categories: pre-test questionnaires, observations, game test notes, post test questionnaires, and interviews, and as such will be presented in the different subsections below for easier readability.
4.4.1. Pre Test Questionnaire
The pre-test questionnaire (see appendix A) gathered data concerning the tester’s experience with mobile devices, mobile games, their opinions about healthcare, and their personal opinion on what is important during pregnancy.
Table 4.1 shows a summary of the first four questions of the pre test questionnaire, before the tester played the game. Out of the five testers, two had played a digital game before, and the same two had also done this on a touch controlled device. Furthermore, all five of the testers had attended a medical healthcare facility during pregnancy or delivery, with four having a positive opinion about the care provided.
Table 4.1 Summary of all answers given to questions 1-4 in the pre test questionnaire
Question: Yes No
Have you played any digital games before? 2 3
Have you played any digital games on a touch controlled device before?
2 3
Have you attended medical healthcare during a pregnancy or delivery of a baby?
5 0
Do you have a positive opinion about the medical healthcare offered for pregnant women?
4 1
Table 4.2 shows the average value of answers given by all testers to question five of the pre tester questionnaire, where the tester was asked to rank nine different actions from 1-9 according to what the tester felt was most important when delivering a baby. Note however that due to language constraints between the observers and the interpreter, this table was misunderstood, and every action was instead ranked individually 1-9, thus resulting in multiple actions having the same score.
Table 4.2 Average score of all answers given by the testers on question five of the pre test questionnaire.
Item/Activity Ranking Sleep 6 Physical Labour 7.6 Avoiding Stress 8.2 Staying at Home 7 Varied Diet 8.8 Staying Hydrated 8.4 Drinking Coffee 3 Exercise 8
Visiting Healthcare clinics 8.8
4.4.2. Observations
which measured the smoothness of the user completing a task in the game. This data was then used for a usability evaluation, presented in further detail by Ismail (Ismail et al., 2011). Each task completed by the tester was ranked by three marks: ‘yes’, ‘partial’, and ‘no’. ‘Yes’ indicated that the tester had no trouble completing the tasks on their own and gave full credit (100%). ‘No’ indicated that the tester were unsuccessful in completing a task without help from the researcher, or using multiple attempts to complete it, thus giving 0% score. ‘Partial’ would give 50% score and was given at observer 1’s discretion, to determine if a mistake should give a ‘partial’ score rather than a ‘no’. An example of a ‘partial ‘ score would be that tester successfully manages a task in the game, but with some intervention from the test organizers. Table 4.3 shows 16 different tasks completed by five different testers. Summarized the total amount of tasks completed for the effectiveness evaluation is 80. Out of these 80 tasks, 19 tasks were successful, 25 partially successful, and 36 tasks were unsuccessful. This data was then used for an equation, detailed in Ismail’s work (Ismail et al., 2011), to produce a score, to determine the overall effectiveness of the game that was produced.
Table 4.3 Measured effectiveness in completing tasks within the game
Screen / Evaluation element
Tester 1 Tester 2 Tester 3 Tester 4 Tester 5
Game screen
The tester did not have trouble navigating using the touch screen
No No Yes Partial Partial
The tester understands the meaning of the game icons
Yes No Yes Yes Yes
The tester knows how to interact with the UI buttons
Partial No Partial Yes No
Food management
The tester knows when the character is hungry
Yes No Partial No Yes
The tester knows how to feed the character
Yes Partial Partial Yes Partial
The tester understands the varied food circle concept
No No Partial Yes Yes
The tester understands how the food tokens work
No No No No No
Water Management
The tester understands when the character needs water
The tester knows that the water supply must be replenished
Partial Partial No Yes Partial
Happiness Management
The tester knows when the character is unhappy
No No No Partial No
The tester understands why the character is unhappy
No No No Partial No
Event system
The tester knows what triggers an event
Partial Partial No Partial Partial
The tester understands the concept of events
Partial No No Partial No
The tester understands the consequences of event choices
No No No Partial Yes
Timeline
The passing of time is clear to the tester
Yes Yes No Partial Partial
The Tester understands that food tokens replenish with each week
No No No No No Individual Sums Y: 5 N: 7 P: 4 Score: 43.5% Y: 1 N: 11 P: 4 Score: 18.75% Y: 2 N: 9 P: 5 Score: 28.125% Y: 6 N: 3 P: 7 Score: 59.375% Y: 5 N: 6 P: 5 Score: 46.875%
Testing Group Sums Y: 19
N: 36 P: 25
Average Score: 39.375%
Figure 4.5 - The calculation of average effectiveness score
Effectiveness (%) = (Nr of Yes + (Nr of Partial x 0.5)) / Total Tasks x 100% = (19 + (25 x 0.5)) / 80 x 100%
By using the equation, it was determined that the effectiveness of the game, tested against pregnant women in Ethiopia had a rating of approximately 39 %. The calculation of the effectiveness of the game is shown in figure 4.5.
Table 4.4 shows ten different tasks completed by five different testers. Summarized the total amount of tasks completed for the efficiency evaluation is 50. 20 tasks were successful, six partially successful, and 24 tasks were unsuccessful. This data was then used for an equation, detailed in Ismail’s work (Ismail et al., 2011), to produce a score, to determine the overall efficiency of the game that was produced.
Table 4.4 Measured efficiency in completing tasks within the game
Screen / Evaluation Element
Tester 1 Tester 2 Tester 3 Tester 4 Tester 5
Game elements
The tester manages to keep the character from being critically hungry
Partial No No Yes Yes
The tester manages to keep the character from being critically thirsty
Partial No No Yes Yes
The tester never ran out of food tokens
Yes No No No No
The tester is preventing the water storage from running out
Partial Yes Yes Yes Yes
The tester is giving the character a diverse diet
Yes No Partial Yes Yes
The tester is preventing the character from being unhappy
No No No Yes Yes
Errors and mistakes done by the tester was minimal
Yes Partial No Yes Yes
The tester knows how to recover from errors and mistakes
Yes No Yes Yes Partial
Research interactions
Level of interaction between researcher and the tester was minimal
No No No No No
Guide and help from researcher was minimal
Sums 2 Y: 4 N: 3 P: 3 Score: 55% Y: 1 N: 8 P: 1 Score: 15% Y: 2 N: 7 P: 1 Score: 25% Y: 7 N: 3 P: 0 Score: 70% Y: 6 N: 3 P: 1 Score: 65%
Testing Group Sums Yes: 20
N: 24 P: 6
Average Score: 46%
Figure 4.6 - The calculation of average efficiency score
Efficiency (%) = (Nr of Yes + (Nr of Partial x 0.5)) / Total Tasks x 100% = (19 + (25 x 0.5)) / 80 x 100%
= 39.38%
By using the equation, it was determined that the efficiency of the game, tested against pregnant women in Ethiopia had a rating of approximately 46%. The calculation of the effectiveness of the game is shown in figure 4.6.
4.4.3. Game Test Notes
During each test session, notes were taken by observer 2. The notes were taken on a mobile phone with keywords, to enable that all of the relevant subjects could be written down. Since the notes were written mainly with keywords, they needed to be transcribed into complete sentences. This was done by observer two straight after each test, this way the observations would still be fresh in memory.
● Tester 1: During the first few minutes of testing, the tester does not seem to understand the character’s happiness, or when and why the character is unhappy. The hunger and thirst also seems to be unclear. After given simple instructions to what indicators to look for, and how to drag and drop the objects, she keeps the character well fed successfully, with diverse and intended food variation. She varies the food intake with 100% success. She does not notice that she is out of water, but when told that the water container is empty, she immediately understands how to get more water. At the end of the session she seems to understand the concept of the game, and plays the game partially successfully.
● Tester 2: She understands when the character wants food and water, but does not understand how
to drag and drop the objects in the menu. When instructed how to drag and drop, she feeds the character the appropriate food types, and successfully varies the food intake about 70% of the time. She does not seem to understand when she has successfully fed/hydrated the character, and keeps feeding the character all the time, and thus runs out of food tokens. After playing a couple of minutes she understands the queues when to feed and hydrate the character, and thus feeds and hydrates at the appropriate time. However, every time she feeds the character, she gives the character one of each food type, instead of just one kind of food, like it was intended. She notices when she runs out of water, and understands how to get more, without any instructions needed.
● Tester 3: She understands when the character is hungry and thirsty, but does not understand how to drag and drop the objects. After instructions how to drag and drop, she properly varies the food 2“Y” signifies amount of fields where yes was written, “N” the amount of fields where no was written, and “P” signifies the number of partial fields.
types with 100% success, without any instructions. However now she drags on all the buttons, even though some are just intended to be clicked. After a couple of tries she figures out she has to click. Now when she has figured out the controls, she plays quite stressfully, and tries to drag food and water all the time, and thus runs out of food tokens. She keeps feeding and hydrating the character even when there is no need. She understands how to get water, but does it so often, that the container never goes lower than 50%.
● Tester 4: She understands drag and drop instantly, and also varies the food properly, with 100%
success rate, without any instructions. She manages to keep the character well fed and hydrated throughout the whole test session. She seems to understand the events and chooses health options every time. Unclear if she does this because her happiness is full all the time, and she thinks she does not need the social events, or if she just did it randomly. She does not understand the food tokens, and keeps feeding the character a lot more than needed, but still with proper variation. She does not notice she is out of water, but after instructed that she is out of water, she instantly understands how to fetch more. She only fetches more water when the container is below 20% capacity.
● Tester 5: The tester does not understand drag and drop, until instructed. But after that she has no problem navigating the UI. She understands that she should vary the food and does so at a 100% success rate. She understands the thought bubbles, and only feeds and hydrates the character when it is needed. During the entire test session she feeds and hydrates the character as intended, and also notices when she runs out of water. She only fetches more water when the container is below 30%.
4.4.4. Post Test Questionnaire
After each game session was over, testers filled out a post test questionnaire (see appendix B), similar to the pre test questionnaire, but with more focus on their experience playing the game. In order to observe any possible changes in opinion, regarding what is important during a pregnancy, the tester was again asked to rank 9 actions from 1-9 (see appendix B: question 5). This questionnaire had the same issue as the pre test questionnaire, where the interpreter asked them to rate each individual item from 1-9 instead of ranking them in a list as it was intended. The sums of all the answers to the questions are displayed in table 4.5. Shown in table 4.6 is the average ranking of how important each item/activity is during a pregnancy, where 1 is least important, and 9 is most important.
Table 4.5 Compilation of the sums of the post test questionnaire answers.
Question: Yes No
Do you feel that the game informed you that a varied diet is important during pregnancy?
5 0
Do you feel that the game informed you that maternal healthare is important during pregnancy
5 0
Do you think the game was hard to understand 0 5
Do you think digital games could be a good medium for teaching?
Table 4.6 Average Ranking of each item among all testers, and the difference from the pre test
questionnaire table.
Item/Activity Ranking Difference*
Sleep 6.4 + 0.4
Physical Labour 6 + 1.4
Avoiding Stress 8,2 Unchanged
Staying At Home 5.8 + 1.2
Varied Diet 8.8 Unchanged
Staying Hydrated 8.4 Unchanged
Drinking Coffee 2.6 + 0.4
Exercise 7.8 + 0.2
Visiting Healthcare clinics 9 + 0.2
*A positive number indicates that the testers ranked the item as more important after playing the game, whereas a negative number would mean the item is less important.
4.4.5. Interviews
The final stage of each individual test session was an interview containing a set of predetermined questions (see appendix E). All questions in the interview were designed to get input from the tester regarding the game they had just tested, focusing on three things: Visual accuracy compared to their real life and their culture, how fun the game was, and what the perceived goal of the game was. In this section each question will presented alongside its corresponding answers in the tables below (see table 4.7- 4.11). All of the interviews were recorded and transcribed later except tester 2. Tester 2 did not wish to be recorded, so the interpreter wrote quick notes about her answers, and as such her answers are not as detailed as the others. The answers given below are direct translations made with the help of two translators. Because of the amount of people around the hospital, there was a lot of background noise in the recordings. So whenever the first translator had problem hearing what the tester said, another translator was brought in, and they would help each other in translating the audio piece. The words in parentheses are clarifications, since the tester’s wording might be incoherent.
Table 4.7 Answers given by the testers to question 1 of the interview.
Question: Did you feel like the game reflected Ethiopian culture? a. If yes: What made you feel this way? b. If no: What did you feel was out of place?
Tester 1 The fact that she is carrying water shows Ethiopian culture, and we also eat vegetables and the fish too. Health centres are also common, and everyone (in Ethiopia) goes to them. If any Ethiopian sees the game, they would be able to see that it takes place in Ethiopia. The way she carried water looked like how we do it in Ethiopia. And also the coffee looked like Ethiopian coffee. The way the character was dressed and how the hut looked also looked Ethiopian.
Tester 2 It reflects Ethiopian culture, and the coffee made it feel like our culture.
Tester 3 Yes it reflected Ethiopian culture, it’s good that it came to our country and it is trying to help us. It will help us (Ethiopians) for our health and our mind set. The food looks Ethiopian. It also showed how the baby was affected by my actions while it was in the uterus. The coffee also looked Ethiopian, the way the coffee cups was place, and also the jerry can (water container).
Tester 4 Yes it reflects our cultural setting, mainly by the way she fetched water. And also having coffee with friends. It looks like Ethiopian context.
Tester 5 Yes it looks like Ethiopia, because of the food and the area the she (the character) is in, and the house that she lives in too. And by the way she carried the water container, and also the coffee. There is nothing in the game that makes it look like it isn’t Ethiopian.
Table 4.8 Answers given by the testers to question 2 of the interview.
Question: What was fun about the game?
Tester 1 I enjoyed watching her go get water, and going to have coffee. And her going to the clinic, is something that should be done by everyone
Tester 2 I found the game interesting and educative.
Tester 3 To be able to tell the character what to do, and I like that I can make the character do what I want, and also how she fetched water.
Tester 4 Watching her eat and drink.
Tester 5 The fact that the game teaches a pregnant woman how to take care of themselves, so that we accept it and learn that from the game. Also that you could touch the structure and the setting of the game.
Table 4.9 Answers given by the testers to question 3 of the interview.
Question: What wasn’t fun about the game?
Tester 1 The game should show her resting or sleeping, and moving around. Tester 2 Nothing
Tester 4 The game doesn’t show the character washing or washing her hands. And the game shows her standing still in front of her hut, it would be nice to see her moving around and going places. I would like her to go in and out of her hut.
Tester 5 Nothing needs to be improved, it’s good.
Table 4.10 Answers given by the testers to question 4 of the interview.
Question: What do you feel the goal of the game was?
Tester 1 It is supposed to teach pregnant mothers that they should eat a balanced diet and that they should drink water frequently. The game also communicates the message that food might not be available to everyone.
Tester 2 The goal was to be educative.
Tester 3 I was really happy about the game, It teaches you what to eat during the pregnancy and about my situation (pregnancy in Ethiopia).
Tester 4 It showed that there is a pregnant lady, and different kinds of food, and that she should eat different kinds of food and vegetables. It is also clear that the pregnant lady should visit the health centre.
Tester 5 The game teaches how a pregnant mother should take of her health. But if a player is not told of this goal, it might be hard for them to understand. I think a player should be told the goal of the game before they play.
Table 4.11 Answers given by the testers to question 5 of the interview.
Question: What do you think the game tried to teach?
Tester 1 The game was easy to understand, and I had no difficulty understanding what it was teaching.
Tester 2 It taught about eating during pregnancy, and healthcare for the mother.
Tester 3 It taught me to apply what is in the game into my life, if I have the means to do it. Tester 4 [combined with previous question by interviewer, see table 4.10 - Tester 4] Tester 5 It teaches a pregnant woman how to take care of herself.
4.4.6. Individual Tester Profiles
For a detailed overview of each tester profile individually, see appendices F-J. Each appendix contains a compilation of the data gathered from one user during the questionnaires, data logs, observations, and interviews.
5. Discussion
In this section we will evaluate and discuss the results from the third day of testing, since both of the earlier days were covered in the results section.
Looking at the game design we constructed, it feels highly applicable to the area of concern. We had another design for the same problem but with a different audience, where the health extension workers would be 3 the main target group of the game, and thus the objective was to educate them how to best do their work. We disregarded this design since we wanted to counter the root of the problem at its source - lack of education for women. However, the disregarded design could still be viable and applicable for the problem. We would like to extend the game’s current design by adding a garden, where the player can tend to plants, and learn basic facts about agriculture, that can be applied in their own backyard. Another extension could be to let the game progress further, after the pregnancy, into the postnatal period, up to the child is five years old. This is because child mortality is critical until the age of five (CSA, 2011). In this stage of the game the player would then take care of the woman, and her child, including the child’s education. This stage of the game was in our original design, but was removed due to lack of time.
An important finding of this research are those related to using a game as a medium for teaching in a developing country, without the use of text elements. The answers in the final questionnaire and the interview confirmed that the game we developed taught the user something about maternal health and nutrition. What they learned differs slightly between each tester. All of the testers confirmed in the interview, that the game taught them about nutrition and healthcare. Four out five testers said that games could be a good medium for teaching, which also confirms that the game taught them something. Through the use of colourful, clear and animated UI and symbols, testers understood the message of nutrition and healthcare the game tried to relay. This suggests that the implementation of game design inspired from children’s games was feasible. This also indicates that the type of game design presented in this document is appropriate for further research and verification. The game also managed to successfully reflect Ethiopian culture in all ways, as there were no remarks about anything that felt out of place. These findings are especially important as there has been no previous research made within this area.
The discussions held with the TEMACC team also confirm that this type of game is something that is relevant for use as a learning platform. They also confirm the game’s possibilities if it were to be developed further. It was suggested that the game could use real life rewards when a player performs well. A player’s success could be communicated to his or her corresponding health centre, where the player can exchange their achievements for rewards. It was also mentioned that the cultural context of the game was appropriate for residents around and in Addis Ababa, but might not be appropriate for different regions of the country, where for example the Red Cross symbol should be exchanged with the Crescent Moon symbol, in order to avoid any cultural or religious complications.
One of the limitations of this study that should be noted, is the limited amount of data acquired in terms of quantity. This was due to certain circumstances of our field study, such as the fact that we needed support staff in the form of translators as well as transportation. This coupled with the fact that some of our assigned support staff did not have the ability to attend the data gathering session, left us with only one translator with a limited amount of time to work with us. Many sources recommend that a usability test should consist of no more than five people (Sauro, 2010). The recommendation of five testers, is because generally, testing with five users will yield discovery of 85% of the usability issues. For every tester you add, the percentage of new discoveries will stagnate. In figure 5.1 is a display of the stagnation for a problem that would occur for 31% of the testers (the red curve), and a problem that would occur for 10% of the testers (the blue curve). Even though each new issue discovery will cost more time, we had hoped to gather at least 10 different sets of data, since we only had one shot at gathering the data necessary, and the time consumed for testing was not a concern for us.
3 Health Extension Workers are women who travel around rural Ethiopia to visit women, and educate them about
