Running your own social experiment has never been easier thanks to Photofeeler and a plethora of no-skills-required photo editing apps. Just follow the steps below!

Step 1: Wonder about something

Here are just a few examples of things you might wonder about that you’ll be able to test on Photofeeler:

• How do vertical vs. horizontal stripes affect a person’s general impression?
• Do blue or brown eyes look more trustworthy?
• Do I look more influential with a white or black background?
• Is it true that men prefer women with long hair?
• Do blondes or brunettes look more fun?
• Does black clothing look good on every type of skintone?
• How badly does blue hair hurt a woman’s professional impression?
• Are cargo shorts unattractive?

Whether you want to know how something affects a person’s general impression or a specific trait, you can find out. You’re only limited by your own curiosity.

Step 2: Create a “control photo” and “experimental photo”

Now that you have a question in mind, it’s time to set up an experiment. You start by creating two images. (No Photoshop skills required, as I’ll get to.)

The first, the control photo, does not have the characteristic you want to measure the effect of. The second, the experimental photo, is exactly the same except for having that characteristic.

Here’s an example:

The second photo above was edited to test how makeup changes a woman’s impression.

Here’s another example:

The second photo above was edited to tone down masculine features for a study about the perceived trustworthiness of faces.

As you might have picked up from these examples, the best method is to pick your control photo and then doctor it to create your experimental photo. That’s because two separately-taken pictures are likely to vary slightly in ways like distance, lighting, angle, and facial expression. Using one photo ensures that changes unrelated to your test don’t interfere.

Here are some apps can use to create an experimental photo:

• Facetune on iPhone/Android: tools for beautifying/warping features
• Picmonkey on desktop or mobile: add makeup, spraytan, all kinds of neat tweaks (use free trial for premium features)
• YouCam Makeup on iPhone/Android: makeup editor
• Lunapic on desktop: change the color of something
• Fiverr: cheaply hire someone do more more sophisticated photo edits
• Google the thing you’re trying to do, and you’ll find even more options

Word to the wise: as easy as it may be to dive into editing and crank something out, it is important that your changes look convincing. Sloppy or unattractive editing can throw your results.

Step 3: Choose which traits to test for

Photofeeler has three sections (business, social, and dating) which allow you to test for all of these traits: competent, likable, influential, confident, authentic, fun, smart, trustworthy, attractive.

Determine which traits are most pertinent to your question, or if you want to test impressions in general, consider testing for all of them.

Step 4: Collect a statistically significant sample of opinions

The best experiments have lots of data. That’s because the smaller your sample size is, the less reliable your results will be. If you only ask five people for an opinion, for example, they could all agree just as a fluke. You don’t know for sure until a much larger group of people agrees.

To be sure your experiment on Photofeeler is valid, get as many votes as you can (80 is recommended) and check your confidence intervals for a clear picture of how much voters are agreeing or disagreeing.

[Note: If there is an element of surprise to your test, it might be best to test your photos on different days. Photofeeler does generally put some space between photos by the same user to help voters forget how they voted by the time a user’s second picture rolls around, but testing on different days helps to ensure you get an entirely different voter pool and prevent “giving away” your premise.]

Step 5: Publish your results

Did you get a clear answer to your question? Whether your results said yay or nay or were perfectly neutral, chances are someone else would find it interesting!

When it comes to publishing your results, you’ve got a lot of options. If you don’t already have a blog of your own, you can always post on Medium or pitch your story to a writer of a relevant online publication (such as Thought Catalog, Bustle, Lifehack, Buzzfeed, or The Huffington Post).

Step 6: Repeat

Last of all, always stay curious!