In clinical terms, diagnosis is defined as detection of a disease that has already occurred in a patient based on its symptoms. Along the same lines, diagnosis in design is detection of anomalies, fallacies, bugs or holes in theories, concepts, processes, products or service offerings. In this method, the researcher wants to uncover the root causes of the problem. He describes the factors responsible or possible explanations for the symptoms.
Quick details: Diagnosis
Structure: Structured, Semi-structured
Preparation: Information, Documents, Brief
More about Diagnosis in User and Design Research
Diagnosis is a problem solving research design that consists mainly:
- Detection of symptoms
- Evaluating the symptoms
- Diagnosis or possible explanation of the problem
- Solution for the problem and
- Suggestion for the problem solution
When talking about rational thinking, analysis and diagnosis, a name that comes up most frequently is that of Thomas Bayes. Bayes postulated a rule based on probability, which is widely used to draw inferences in the field of science and mathematics.
Bayes theorem is used to make an educated guess and to determine the likelihood of an event occurring given that another has occurred. Bayesian thinking can be applied to predict the probability of the occurrence of an event given another occurred, or determine that a hypothesis holds true given specific evidence, or diagnosis of a bug or a problem given the symptoms.
To explain Bayesian thinking in simple terms let us consider an example. Let us say that a survey was conducted to study the preference of the color ‘teal’ between men and women. Let us also say that the ratio of men and women participants was 1:1. On analysis, the survey results indicated that 60% of women preferred ‘teal’ to any other color whereas only 30% of men preferred ‘teal’ to other colors. One way to interpret the above-mentioned findings is that the probability of a user choosing the color ‘teal’ given that they are female is 60%. This can be represented mathematically as P (User choosing ‘teal’ | female) = 0.6. Knowing that in a demographic with equal number of men and women, women are more likely to pick the color ‘teal’, researchers can take important decisions about color schemes to employ when designing for men or women. Similarly, let’s say that of the 50 visitors to a page with 2 buttons – one orange and one teal, only 5 end up clicking on the orange button. Therefore, the symptom is that the orange button received fewer clicks than the teal one. The evaluation of the symptoms help the researcher uncover the possible explanations for the behavior such as the teal button is placed first in line and the orange second, most page visitors happen to be women and women prefer teal, the shape of the teal button is more appealing or the nomenclature used for the teal button accurately describes the function it performs. Then, choosing one of the stated explanations at a time, the researcher can validate or determine the likelihood of the possible explanations to diagnose the problem, make alterations and reiterate.
Diagnosis can be conducted on large amounts of data collected from different individuals together. It is, therefore, not as time consuming when using online tools; however, a critical thinker or good data analyst can determine the patterns in the collected data to diagnose the real issues or problem areas for resolution.
Advantages of Diagnosis
1. Problem formulation based on symptoms
One of the methods that help researchers trace the cause and effect relationship between product elements or features and user behavior.
2. Critical thinking
Diagnosis requires critical thinking and analysis to formulate the real problems that need to be addressed to deduce a solution.
3. Time effective
Diagnosis is quick and time effective when using online tools and can be used to analyze large chunks of data gathered from large sample groups.
4. Online tools for data analysis
Many online tools can be employed to gather data, carry out analysis on this data to spot patterns or symptoms which can the ultimate diagnosis of the problem much more streamlined for the researchers.
Disadvantages of Diagnosis
1. Experienced researcher
An experienced researcher who is familiar with critical or Bayesian thinking would be more effective at diagnosis than any researcher.
2. Incorrect symptom evaluation
If the symptoms are not evaluated correctly or the researcher fails to ask the right questions, the diagnosis of the problem and its optimum solution may be way off.
Think Design's recommendation
Diagnosis requires the researcher to be aware of the symptoms/ indicators and what they tell about the problem. It is hence, important that the researcher understands the subject/ domain before diagnosing symptoms. In general, Diagnosis is a fundamental research activity that is an essential part of every investigation we do, however the level of formalization of the activity itself varies depending on the assignment. For this reason, diagnosis can be at one extreme, highly structured or at the other, totally unstructured. Choose your structure wisely as it primarily affects the outcome of your effort and the solution at large.
Another important element here is the limitation of this method itself. Since Diagnosis fundamentally involves the understanding of symptoms (and without direct contact with users), it does not provide insights into what users think. For example, you may diagnose your product performance and come to the hypothesis that your users may have clicked something or ignored something based on your understanding of symptoms; however, Diagnosis does not delve to asking users why they did what they did. It is recommended hence, that a primary research method is always combined with Diagnosis in order to avoid conclusions that may be ill-informed.