Moodnotes is a CBT-based thought and mood journal available on the app store for $4.99. It does a good job introducing the core CBT skill of identifying cognitive distortions and rewriting thoughts. The mood capture doesn’t reflect current theories of emotions, and the “Insights” (i.e., analytics) leave much to be desired.
- You can set multiple daily reminders to track mood
- It incorporates very basic recommendation of activities based on mood—if you say you’re in a positive mood it suggests activities that may increase positive feelings while if you’re in a negative mood it takes you through a CBT-style cognitive distortion identification and thought rewriting activity
- Identification of thought traps provides descriptions of each one.
- Doesn’t use a therapist-like simulation or chatbot, which is probably a good thing. Uses structured data capture that apps are so good at!
- Provides export of mood tracking data.
- The specification of moods allows the user to pick multiple emotion words for one point in time and allows them to also specify percentages associated with each mood word. This doesn’t reflect current theory around emotions and doesn’t support good analytics/machine learning opportunities in the future.
- Chart of moods over time isn’t organized according to time.
- No goals that user is working towards or higher-level guidance to give a sense of achievement. Why would a user keep using this over time without some sense of progress and achievement?
- Privacy of data: “Although Moodnotes collects anonymous, aggregate usage data, absolutely none of your personal information or entries are ever collected.” I consider this a negative as this will seriously limit their ability to build in machine learning supporting better application effectiveness.
Mood tracking in Moodnotes
The basic starting point in Moodnotes is creating an entry where you specify your mood by swiping up or down on a face. As you swipe up (indicating positive mood), colors change from green to yellow. As you swipe down (indicating negative mood), colors change from blue to gray.
Once you’ve chosen a level of positive or negative feeling by swiping, you can either save the entry as is or “add detail.” When adding detail, you specify what you’re doing right now, and then select feelings, from positive or negative lists. You can select multiple emotions (positive AND negative at the same time) and specify the amount of each emotion.
This is too much! When designing a mood tracking capability, designers should consider what sort of data set they’re generating. If you want to be able to predict mood or act on mood values, you are going to be able to do that most easily with just ONE emotion specified at one point in time.
Since the circumplex model of valence and arousal has been well-confirmed in research, this would be a good way to allow people to select their mood at a point in time.
You wouldn’t have to have them select on a circumplex, but could prompt them for how positive/negative they felt (valence), how tired/energetic they felt (arousal), and also prompt them to select a name for their emotion.
This would form the basis for a machine learning capability to classify emotions later based on valence and arousal (or based on other predictors, such as heart rate, blood sugar level, activity, time of day, …).
You can set multiple reminders per day for capturing your mood, which, if you have the motivation, is a good thing. This potentially could allow you to understand your diurnal mood variation.
Activities based on mood
Moodnotes includes a very basic but still useful “recommendation” engine for activities based on mood. If you indicate a positive mood, Moodnotes directs you through an activity that promotes additional positive feelings. If you indicate a negative mood, it takes you through a CBT-style thought rewriting activity that starts with identifying cognitive distortions.
An example of a positive mood activity is specifying some of your skills and positive qualities, then identifying how you might use these strengths to help yourself or others:
If you have a negative mood, you can go through the “check a thought” activity:
The traps include:
- All-or-Nothing Thinking
- Downplaying Positives
- Emotional Reasoning
- Fortune Telling
- Intolerance of Uncertainty
- Mind Reading
- Negative Filtering
- Not Accepting
- “Should” and “Must” Statements
Helpfully, when you select from these traps a description of each is provided:
Then you are prompted to rewrite your thoughts and reassess the intensity of your feelings.
Chart of moods
After you’ve logged moods for a few days, you can look at your mood chart, which bills itself as “Your Mood Over Time.” It’s nothing of the sort!
Data that is collected through time should be plotted over time (with time on the x-axis), not summarized like this in a way that completely obscures any time trends and cyclicality. The goal in looking at moods over time would be to answer questions like:
- Is my overall mood improving, staying the same, or decreasing?
- What sort of intraday mood variation do I see? Is this relatively consistent across days? There is plenty of evidence that most people have mood peaks in the morning and evening, and valleys in the afternoon.
- How does my use of positive mood activities and “check a thought” affect later moods?
These “Insights” give almost no insight at all.
However, the app allows you to export your data, so you could conceivably do your own time-series type visualization and analysis of your mood data.
The designers of this app have made some great choices around mood tracking and introducing the core activity of CBT, checking thoughts for distorted thinking and rewriting them.
Their decision not to use a therapist/chatbot-style interface is a good one, I think, but I wonder if the lack of higher level guidance and ability to sense any progress could be a downfall for engagement over time and overall effectiveness. There’s evidence that self-guided online CBT programs don’t do as well as therapist-led online programs.
But they’ve also made some questionable choices around what data they collect and how it’s displayed for analysis purposes. Capturing multiple emotion words at one time limits analysis and machine learning opportunities down the line. You might think more is better, but in this case, a more theory-based capture of emotion content (capturing valence, arousal, and optionally an emotion name) could lead to better analytics in the future.