I also wanted to make a name for myself in the competitive field of alphabet factoring in psychology. G factor. P factor. I factor. D factor. Aaaaall the factors. My recent list is already outdated. So here I’ll introduce the S factor, for satisfaction. Look how satisfied this little student frog is!
Seriously though, there is a substantively interesting question: we asked nearly 2000 students, in our large-scale WARN-D mental health survey, not only about context questions researchers are commonly interested in: Do you work? What is your living situation? How are your grades, finances, sleep? We also asked them about satisfaction with these things, because we thought that would be much more important for mental health research. Let’s take a quick look at the data.
1. Satisfaction questions
We didn’t really find questionnaires on this, so we made our own, 14, satisfaction questions. Specifically, about satisfaction with their …
- romantic relationship (and if they don’t have any, their satisfaction with not having any)
- sex life
- living situation
- studies (they are all students)
- grades
- finances
- independence (“Young adulthood is a time where people might move out from their parents, gain financial independence, etc. How satisfied are you overall with your level of independence?”)
- sleep
- leisure activities
- weight
- looks
- family relations
- friend relations
- and their work (this is the only conditional item — they only get it if they indicate they work, which around 38% of students do — so we will skip the item from below).
This was not one specific questionnaire, but questions throughout the 75 minute survey on risk and resilience for depression. Some are about the “current” (how satisfied are you with your work), others more specific — e.g. for leisure, we asked for the last 2 weeks specifically. See our codebooks for details (protocol paper, supplementary, stage 1).
2. Let’s look at the data
I’ll only look at a n=700 subset of our data here. Distributions look kind of ok — overall students are pretty satisfied (1 is “very dissatisfied”, 7 is “very satisfied”). The peak at 3 for sleep and weight dissatisfaction stands out a little.
Correlations between items are all positive (Pearson, Spearman gives virtually same results).
Here the heatmap of the all correlations.
Most correlations are between r=0.1 and r=0.3. The two strongest correlations are between romantic relationship and sex life (r=0.58) — makes sense — and weight and looks satifsaction (r=0.63). This is interesting and appears to imply that a lot of the looks is determined by the weight. Scatter plot between these 2 reveals nothing surprising.
3. Introducing: the S factor for satisfaction!
So I fit a one factor CFA model to the 13 indicators and — ta-ta — it ain’t looking bad folks. Turns out I may also be ready for the alphabet factor literature: CFI 0.91, TLI 0.89, RMSEA 0.05.
I mean, sure, the fit was absolutely abysmal when I first tried my ‘confirmatory’ factor model, which is entirely exploratory. But then I confirmed, after briefly peaking the modification indices table, that I would really like to add two residual correlations, based on theoretical considerations of course. And so based on my unifying domain of satisfaction, it was very easy to get from abysymal fit that we don’t report1 to something that would be publishable. And this took literally half a minute — I’ll get you a non significant chi-square in n=1500 if you give me an hour!
4. Conclusion and ways forward
Getting back to the data I find actually interesting, I’d appreciate any comments and thoughts on what to do with it. Our next step will be to check if the information commonly collected (are you in a relationship, what is your living situation, do you work) relates less to mental health outcomes than our satisfaction questions — and I bet that’s the case. If you have come across similar questions or questionnaires, or know of work on the topic, please let me know!
Looks like a great start! Something I’ve been curious about for a while is the relationship between values & life satisfaction – does holding certain values predict greater life satisfaction (holding constant other factors like some of the ones you’ve explored in this data set)? Or perhaps it’s about the degree of alignment between one’s values and one’s satisfaction with relevant domains of life?
Not an expert on the topic, but there is definitely some recent work on the relation of personal values and life satisfaction, e.g. https://journals.sagepub.com/doi/full/10.1177/02654075231173157.