This is what we did:
- Decided on which of the factors hypothesized to be related to weight change over three years among middle age New Zealand women to start our analysis.
- These are things like speed of eating, intuitive eating, psychological flexibility, etc.
- Categorized women by baseline BMI into underweight, normal weight, overweight, and obese and then looked at the relationships with factors mentioned above.
- Categorized women by weight change into -5+%, -3 to -5%, -3 to + 3%, 3-5%, and 5+% of baseline weight and looked at the relationships with factors mentioned above.
I got a kick out of our reactions when the p value was and wasn't what we hoped for. Why all the fuss about a p value? Well, I want to run an intervention, but I need to analyze the weight gain with these factors so that I can say, "See! This and this are statistically significantly related to weight gain over three years (because the p value is less than 0.05 - meaning it's less likely the results happened by chance) and so these are what we can target in a weight gain prevention intervention!"
Isn't research fun?! I'm half tempted to print out our results and tape them to the fridge in my flat just to show off our great work. Is that too nerdy of me?