Can Your Survey Design Affect Your Matching Methods?

So Megan has run her social media survey over a Saturday and a Sunday, has collected several thousand responses, and you've convinced her that she should use matching methods to find out if her correlation is actually a causal effect. What might be survey design issues that could hurt her attempts at using matching methods to generate a convincing causal effect of going to a restaurant on exercise?

Possible answers

A) If she made some of her questions optional, she may have a lot of missing values in her data

B) Because it's on social media, some responses will be tainted by peer pressure

C) If she doesn't define what she means by exercise or eating, then people might cheat with their answers

D) A and C

All of the above

Exercise

Can Your Survey Design Affect Your Matching Methods?

Instructions

Possible answers