Communication 3163 – Industry Research Methods
Content Analysis Example
Welcome! For this in-class practice, imagine that you are a researcher for an advertising agency. Your company is interested in determining whether having a cute or ugly baby in a commercial will affect people's purchasing decisions. You are designing an experiment in which you will assign participants to one of four experimental conditions: cute baby in the ad, ugly baby in the ad, neutral baby in the ad, or no baby in the ad. Your research assistant gathered 50 pictures of baby things (humans, animals, or inanimate objects) that you could use for this experiment. Now you just need to content analyze them to figure out which condition these stimuli belong to.
Your task today is to develop rules for your content analysis, code examples of babies, and evaluate the reliability of your coding scheme. You will need to work in groups of 2 or 3 (no more or less!) to accomplish this.
Once you are in your groups, decide on one person who will act as the record keeper. This person needs to have access to a computer that can download and edit an Excel file. Next, discuss with your group some rules for coding pictures of baby things. The classifications you will use are: Ugly (0), Neutral (1), and Cute (2). Some example rules might include things like:
- All inanimate objects should be coded as Neutral (1)
- Animals without hair or fur should be coded as Ugly (0)
- Smiling animals and humans should be coded as Cute (2)
Be creative, but also specific! There should not be any ambiguity when you look at a picture about how you should code it according to these rules. Each person should write these down for themselves because you will each be coding pictures on your own.
Next, you will need to select a random sample of the population of 50 images that you will code. Follow THIS LINK and write down each of those ten numbers on a piece of paper. This is your sample.
When you are ready to code each picture in the sample, click the link below with the corresponding number. EACH PERSON NEEDS TO CODE ON THEIR OWN; DO NOT DISCUSS YOUR SELECTIONS WITH YOUR GROUP MEMBERS UNTIL YOU HAVE FINISHED! When you are coding, write down a "0" for ugly babies, a "1" for neutral babies, and a "2" for cute babies. Which number you write should depend on your interpretation of the rules you established as a group.
Record keepers: download THIS FILE if you are in a group of 2, or THIS FILE if you are in a group of 3. Place the names of each group member at the top of each column C through D or E. In column B, enter each of the ten numbered pictures you coded. Then in each named column (C through D or E), enter that person's numeric code (0, 1, 2) for the corresponding picture. The reliability statistic will automatically update as you fill in the information.
As a group, evaluate the statistic. Is this acceptable? If it is acceptable, discuss with your group which rules you found most helpful and which could use some improving. If it is not acceptable, revise your rules to be more clear and concise. Then, get a different set of 10 random numbers and code that new set of pictures. Does this provide a better or worse result?
When you have reached a good stopping point, double check that all of the names are spelled correctly and save the file. You need to email this to your lab instructor to receive attendance credit for today.