Help! Color and Emotion Psychology Experiment

[deleted-47005]

My sixth grade daughter is interested in the effects of color on emotion and behavior. She did a lot of background reading before designing her experiment. This is her design:

She printed our a silohuette of a person. She mounted one copy of it on a pink background, of a particular shade that has been noted in research as having a positive effective on emotion. Then she mounted another copy on a white background. She is intending to interview as many people as possible, without telling them the purpose of her study. She will record each person's age and gender. She will roll dice to determine whether to show the interview subject the picture with the pink background or the white background- they will only be exposed to one of the pictures. Then she will use a timer to make sure that each person looks at the picture for exactly one minute. After one minute she will ask: "What does the person in the picture look like she's feeling?" and "What does this picture remind you of?"

She will then record their responses and will later talley the number of positive words versus negative words used in the responses. The idea is to see whether the color of the background affected the responses to the pictures.

Does this design look adequate for a 12 year old?

What are her dependent and independent variables? Would the color background be the independent variable, and the words used by the interview subjects be the dependent variable?

This is for a non judged science fair, and it is her first science fair-- but I want to make sure that it is a good experience for her. Thanks for your help.

[donnahardy2]

Hi,

This is an excellent science fair project for a 12-year old, especially since this is a topic that she’s interested in. The experiment is very well designed and should yield good data. You are correct; the independent variable is the color of the background, and the dependent variable is the response from the subjects. Your daughter can present results in a bar graph by age and gender groups. I recommend testing as many subjects as possible.

You should check out the science buddies guide for doing a science fair project. Your daughter is off to a very good start because she has done her background reading.

https://www.sciencebuddies.org/science- ... ndex.shtml

Also, since your daughter will be using human test subjects, she will need to have permission slips from her subjects or their parents. Here are the guidelines for doing a project with a psychological survey:

https://www.sciencebuddies.org/science- ... _src.shtml

Donna Hardy

[deleted-47005]

Thanks for the info. We DID do the permission slips and forms you mentioned. She already presented this at her first science fair, and wants to continue the project. She is still very interested in the topic. She is 12 and is wanting to know if she can learn some statistical techniques for determining if the results of her data are statistically significant. She ended up changing the design of the survey, because when she tested it, the respondents all answered with verbs, and she was not able to derive any emotional content from it. So she changed it to be a multiple choice survey. She asked people if, after viewing the picture, they felt, "happy, sad, angry or calm." Then she was going to lump the happy and calm responses into one group and then angry and sad into another. She discovered, however that people who had sad open-ended responses also frequently chose calm. So she realized that "calm" is a state of being, not an emotion. I thought that was a cool insight gained from analyzing her data.

If she changes the survey to offer only two strictly dichotomous choices, such as "negative," and "positive" or "calm" and "not calm," could she use a t-test to determine if her results are statistically significant? Would that student t-test be applicable? Or are there other strategies she could use?

Or, if she discounts the "calm" responses on the current survey, and collects more data could she determine statistical significance?

Thanks for your help.

[MelissaB]

Hi,

A t-test is used to see if there are statistically significant differences between two groups. So, the short answer is yes--if she made a new survey or if she organized the results from the old survey into two different groups, she can use a t-test to see if there is a statistically significant difference between the two groups.

I'm pretty sure there are free online stats pages that will give you t-tests; just google 'free online t-test'. And, yes, Studentized is fine (there are different sorts of t-tests, but that's really not something she needs to know at 12 years old). If you have Excel, you can also get Excel to do a t-test (function is 'ttest(group 1, group 2, tails, type)--tails and type should both be 2).

Good luck!

[deleted-47005]

Thank you so much for this help! We just have one more question:

On the survey after viewing the image she asked them, "Do you feel happy, sad, angry, or calm?" The choices were randomized so the order of the choices wouldn't affect the outcome. She also had a free response section for them to state what they were thinking after viewing the image. She found that the many people who selected "calm" on the multiple choice had very, very sad free responses- for example, "losing my mom to cancer," "the end of a 20 year friendship," etc. That's what led her to conclude that "calm" isn't an emotion- it is a state of being. (I was so amazed that she thought of that after looking at her survey results). She had thought that Happy and Calm / Sad and Angry were two distinct groups.

SO- should she drop the "calm" responses in order to do her t-test? Or should she rewrite the survey to have only two choices, such as "calm" and "not calm" or "positive" and "negative?"

Thank you!

[MelissaB]

Sorry no one has answered your question!

I would go ahead and use the results she already has so she doesn't need to do the project all over again. My inclination would be for her to test positive/negative after removing those who answered 'calm' from the survey. She could do another with calm/not calm, if she wanted. Or, she could lump 'calm' with 'sad' and 'angry' and test the difference between that and 'happy'. I would let her choose based on what she's interested in, or just let her do all the tests and see which one(s) are significantly different (technically you're not supposed to do this in statistics, but for a 6-8 grader, it's not going to matter).