Framing effect
Read the below two scenarios carefully:
Scenario 1
Imagine that you have a friend over for dinner and you are preparing the meal. You know little about food products, but your friend is quite knowlegeable. One of the ingredants you are cooking with is ground beef. You are trying to learn a little about food from your friend and you two begin discussing meat, ground beef in particular. Your friend happens to mention that the ground beef you are using is “85% lean”.
How lean or fatty do you think this ground beef is relative to other ground beef on the market?
Scenario 2
Imagine that you have a friend over for dinner and you are preparing the meal. You know little about food products, but your friend is quite knowlegeable. One of the ingredants you are cooking with is ground beef. You are trying to learn a little about food from your friend and you two begin discussing meat, ground beef in particular. Your friend happens to mention that the ground beef you are using is “15% fat”.
How lean or fatty do you think this ground beef is relative to other ground beef on the market?
You may have noticed that the two scenarios are identical except one describes the ground beef in terms of how fatty it is and the other describes the ground beef in terms of how lean it is.
Did you have the urge to rate the 85%-learn ground beef as leaner than the 15%-fat ground beef?
45% of the participants that did this framing task in my study did!
This phenomenon you have just witnessed–people responding differently to different descriptions of the same object–is known as framing effect.
Are framing effects reasonable?
Pyshcologists’ opinions
Many psychologists argue that framing effects are mistakes because they violate the rational pricniple of description invaraiance: people’s responses towards the same object should not change simply based on description. Here’s an argument for description invariance:
It is unreasonable for people to be influenced because the two descriptions convey exactly the same information about the ground beef.
If the ground beef is 85% lean, then it is also 15% fat, and vice versa. No matter what wording the speaker uses to describe the ground beef, the objective information is the same.
Because all that matters is the objective information in the description, it is unreasonable for people to be influenced by the way the speaker describes the ground beef.
Others believe that framing effects can be reasonable because of information leakage: speakers do not choose their frames randomly, and different frames may convey different information from the speaker. Here’s an argument for information leakage:
It is reasonable for people to be influenced because when a speaker singles out one characteristic of an object (e.g., fatty or lean), it communicates that the object may stand out on that characteristic.
For example, a speaker may describe the ground beef in terms of how fatty it is if its fat content is higher than other ground beef on the market. The speaker's description of the ground beef's percentage of fat therefore suggests that the speaker thinks it is fattier than other ground beef.
Because the speaker's wording may convey their beliefs, it is reasonable for people to be influenced by the way the speaker describes the ground beef.
Which principle is more rational, description invaraiance or information leakage? Is it a mistake afterall to show framing effects? Some researchers suggest that if a principle is truly rational, people should accept it once they understand it. Therefore, in my study, wea asked participants whether they think it is reasonable to respond differently to different descriptions of the same object, and whether they accept one or the other principle discussed here.
Lay people’s opinions: Our study
Participants read the two scenarios and rated the ground beef. We showed them their ratings for the ground beef in the two scenarios side by side, and asked them to respond to the scale below about how reasonable it was to rate the two ground beefs differerntly based on how it was described.
To manipulate their understanding of the two principles, we showed them both the description invaraiance argument and the information leakage argument as shown above, (half of the participants read the description invaraiance argument first and half read the information leakage argument first), and responded again to how reasonable it was to rate the two ground beefs differerntly based on how it was described.
Finally, to further test participants’acceptance of framing effects, we offered them the chance to update their fat/lean ratings of the ground beef after before and after reading the arguments.
Our results showed that:
• After reading both arguments, participants rated framing effects as slightly more reasonable. In other words, deeper understanding of the two arguments seems to slightly decrease acceptance of description invariance.
• Participants were more likely to provide the same fat/lean ratings for the two frames after reading the two arguments.
Conclusion
34% of our participants initially reported that it is reasonable to respond differently to different frames, and presenting arguments for and against description invaraiance increased the percentage.
