Everyday Research Methods

Kids' brains while reading: Screens vs. paper

The study found that children's brains were differently active when they read on screens vs. on paper. Dasha Trofimova/Shutterstock

As a reader, you've certainly had experience reading text on both screens and on paper. You probably have a preference for one over the other. Have you ever wondered whether your brain reacts differently to these different reading formats? 

A journalist from PsyPost has reported on a study that tested this question in young readers, ages 6 to 8. The study used EEG (measured by placing a net of electrodes on the head) while children read on screens and on paper. 

Here's how the journalist described the study's method: 

Each child participated in a reading session where they were exposed to two different conditions: reading from printed paper and reading from a screen. Each condition involved reading age-appropriate expository texts followed by comprehension questions to assess their understanding. The comprehension questions served both to engage the children actively with the texts and to provide a measure of how well they understood the material under each medium.

The researchers were interested in a specific brain wave pattern called the theta/beta ratio. In lots of past research, a high theta/beta ratio has been associated with worse concentration, less attention, and more mind-wandering. The researchers wondered if the theta/beta ratios of each child would differ across the two conditions. 

a) Based on what you've read so far, what are the variables in this experiment?

Name of the variable (Conceptual level)	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or DV?	For IVs: Is it manipulated as independent groups or within-groups?

b) What type of experiment is this? Posttest only? Pretest/posttest? Repeated measures? Concurrent measures? 

Here's the journalist's description of the study's results:

... a higher theta/beta ratio was observed during screen reading compared to paper reading. This ratio is commonly interpreted as a marker of cognitive load, suggesting that children may experience more difficulty in maintaining focused attention and processing information when reading from screens.

Together, the findings provide neurobiological evidence to support the idea that reading from a screen may impose a greater cognitive burden on children, affecting their ability to concentrate and process information as effectively as when reading from paper.

c) Sketch a bar graph of this study's results.

d)  In a design like this, it would be important to employ counterbalancing. Why? Which of the four big validities is counterbalancing addressing? 

e) In this study, children in the study read two different, age-appropriate reading passages, and they also read the texts under two different conditions. How might you arrange counterbalancing for this set of circumstances? 

The empirical article is open access, on PLOSOne. 

 

Suggested answers:

a) Based on what you've read so far, what are the variables in this experiment?

Name of the variable (Conceptual level)	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or DV?	For IVs: Is it manipulated as independent groups or within-groups?
Reading condition	On screen, on paper	manipulated	IV	Within-groups
EEG pattern: Theta/beta ratio	Higher ratio to lower ratio	measured	DV	n/a

b) Repeated measures (it's a within-groups design, and participants' brain activity was measured as/after they read first one type of text and then another type).

c) Theta/beta ratio should be on the y-axis. The two reading conditions are on the x-axis. The bar for "on screen" reading should show a higher theta/beta ratio than the bar for "on paper" reading.

d) Counterbalancing addresses internal validity; specifically, an order effect, which is an internal validity threat that is applicable only to within-groups designs. 

e) Some children would need to read Passage A on screen and Passage B on paper, and other children would need to read Passage B on screen and Passage A on paper. Similarly, some children would need to read on screens first, followed by paper; other children would need to read on paper first, followed by screens.  In the empirical article, it is not very clear how well they counterbalanced these factors. 

 

Really? Wearing high heels can improve your walking

Should we trust a study that tested only 8 people? Photo: sirtravelalot/Shutterstock

Yahoo news picked up a story from the Washington Post , which, in turn, had covered an empirical journal article about high heels and walking (did you follow that?). The Yahoo headline read, "Can wearing high heels actually improve your walking?"  In this post, we'll check out the weaknesses of this study--and think about why these two news sources thought the study would be worthy of coverage. 

Yahoo's story started out like this: 
Can wearing high heels actually be good for you? A new study published in the Journal of Applied Physiology says that wearing high heels can help you improve your walking overall. This could come as a surprise since there are many warnings against wearing high heels constantly.

Then the Yahoo journalist says:

Researchers wanted to find out what changes occur in the human leg when someone starts wearing heels for the first time. Researchers had participants wear custom heels that were, at most, 3 inches for 14 weeks, per The Washington Post.

The participants that consistently wore the heels over the 14 weeks had become efficient walkers, specifically using less energy to walk than before they started wearing the heels,

a) Based on what you've read so far, what is the independent variable in this study? What is the dependent variable?

Now, read a few more details about the study from the Washington Post article (linked here). 

For the new study, published this month in the Journal of Applied Physiology, Beck and his colleagues looked for healthy young men, as well as women, who had never or almost never worn high heels. By recruiting heel virgins, the researchers hoped to be able to track changes inside people’s legs associated with starting to wear the higher footwear.

They wound up with five male and three female volunteers.

The scientists then created high-heeled shoes acceptable to all recruits by attaching foam wedges to the soles of flat Chuck Taylor All-Star Low Top sneakers. Each pair was custom fitted so that a participant’s foot would be flexed downward at a 14-degree angle. In practice, this meant the shoes sported about a 2½- to 3-inch wedge heel.

This summary above should alert you to one of the potential problems with this study: its small sample size.

b) Why might it be a problem to have a sample of only eight people in any study? Your answer should talk not about generalization, but about how a small sample might result in results that are unusual and cannot be replicated. 

The study was a pretest-posttest design. The Post article explains that the researchers tested each of the participants on their walking efficiency (as a pretest), and then gave them their high heeled Chuck Taylors to wear for 14 weeks. After the 14 weeks, everybody's walking efficiency was tested again. 

As you know from Chapter 10, in a pretest-posttest design,  participants are tested on the dependent variable (here, walking efficiency) twice, once before and once after exposure to one level of the independent variable. So let's read more about the independent variable. Some people wore high heels and some did not, but....

[The researchers] asked them to go forth into the world and wear their new heels every day for 14 weeks. Not everyone did. Some of the volunteers “were kind of embarrassed” by the makeshift stacked sneakers, Beck said. Others complained of squished toes or other discomforts, quibbles familiar to any flat-shoe habitue who dons heels on occasion.

But more than half of the volunteers, male and female, wore the shoes most of the time. After 14 weeks, everyone returned to the lab and repeated the earlier tests. Those volunteers who had mostly given up early on wearing heels showed no changes in their legs or walking. But those who wore the heels fairly consistently tended to have shorter calf muscles and stiffer Achilles tendons than before.

c) Check that paragraph again, and ask yourself, Did the researchers randomly assign people to the "high heels" and "no high heels" conditions? Why or why not? 

d) Looking back at c), think about how this poses an internal validity problem for the study. Which of the internal validity threats (from Chapter 10 or 11) does it seem to be? 

e) Why do you think this particular study (which had a tiny sample, put only 5 people in one group and 3 in the other, not randomly assigned) was picked up by the Washington Post and by Yahoo news?  Do you think the study should have been covered by the press? 

 

You can find the original empirical journal article here. 

 

Coping with something? Try holding hands with a person you love

The studies had couples hold hands (or not) during stressful events.
Photo: fizkes/Shutterstock

This post helps illustrate how researchers can start with a simple experiment and then move to a factorial design (both designs here are IVxPV designs).

In this story covered by KHOU, (originally in the Washington Post) we learn that when people hold hands during a stressful situation, it can help lower blood pressure and perceived stress. 

In one study, researchers had couples hold hands during a scary movie. That small act helped limit the stress on our nervous system, calming parts of the brain responsible for vigilance and emotional response.

In the study above, the researchers (Graff et al., 2021) had a group of 83 couples watch 10-minute clips from horror movies.  They randomly assigned half of the participants to watch alone, and the other half to hold hands with their spouse while watching. (In the hand-holding condition, they measured the pupils of one person while their comforting spouse wore headphones and was not able to see the movie.) The researchers used eye trackers to measure pupil dilation as a measure of stress during the movie. Pupillometry has shown that when people are under stress, their pupils dilate. Researchers use cameras to measure this pupillary response. 

The researchers found that the pupils of the participants who watched the scary movie alone dilated more (indicating more stress) than those of people who watched while holding hands with their partner. 

So far, this is a simple experiment.

a) Classify its key variables (so far) using the chart below. 

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV, PV, or a DV?	For the IV/PV: Is it within-groups or between-groups?

Now to introduce a second element of the study. These researchers gave each couple a questionnaire called the Social Relationship Index (SRI)  to measure the quality of their relationship. Some of the couples were classified as "supportive" and others as "ambivalent."

b) Return to the table of variables and add this new variable to the chart above. Hint: Rather than being a true IV, this is a PV, or Participant variable. 

c) The addition of this relationship quality variable makes this study into a factorial design. It's a 2x2. Is this a 2x2 independent-groups factorial? A 2x2 within-groups factorial? Or  a 2x2 mixed factorial? Explain your answer.

d) The results found that in the groups holding hands (compared to watching alone) there was lower pupil dilation for all couples,  but the effect was especially strong among supportive couples. Sketch this result. Put the DV on the y-axis and put the two hand-holding conditions on the x-axis. Use different line or bar graph colors for supportive and ambivalent couple types.

e) Estimate the main effects and interactions following this sentence structure:

There [was/was not] a main effect for the hand-holding condition, such that pupil dilation was higher for the  __________ condition than the _______ condition. 

There [was/was not] a main effect for the couple support type, such that pupil dilation was higher for the  __________ couples than the _______ couples.

There was an interaction, such that__________.

 

Now here's another study that measured the same question, with a slightly different design. The journalist described this study as follows: 

It matters whose hand we are holding. 

A very small study on married women found holding a stranger’s hand did lower stress levels. However, that effect was even more pronounced when they held their husband’s hand.

In this "very small study" the researchers (Coan et al., 2006) had 16 heterosexual married couples participate. The wives were the focus of the study. The wives lay in an fMRI scanner, and over several minutes, they were subjected to a series of minor electric shocks. As you might imagine, this is a stressful experience!  During one block of shocks, the held their husbands' hands. During another block, they held the hand of a male stranger they could not see. And during another block, they did not hold anyone's hand. These three blocks were counterbalanced. The researchers measured activity in brain regions that respond to stress. As the journalist reports, holding a stranger’s hand did lower stress levels. However, that effect was even more pronounced when they held their husband’s hand.

f) Classify its key variables (so far) using the chart below. 

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV, PV, or a DV?	For the IV/PV: Is it within-groups or between-groups?

Now for the PV. The journalist wrote, about this study:

For couples with the highest scores on marital quality tests, that stress relief was even more significant. So basically, the better the relationship, the more hand-holding helped.

The Coan team also measured relationship quality, this time using the Dyadic Adjustment Scale. Assume there were two groups for relationship quality: high quality and low quality.

g) Return to the table of variables and add this new variable to the chart above. Hint: Once again, this new variable is a PV, or Participant variable. 

h) Again, the addition of this relationship quality variable made this study into a factorial design. So the design is now a 3x2. Is this a 3x2 independent-groups factorial? A 3x2 within-groups factorial? Or  a 3x2 mixed factorial? Explain your answer.

i) The results of the Coan study found that while holding hands with their husband (compared to holding hands with a stronger or with nobody) women had lower stress responses in the brain. And, the drop was especially strong among wives who reported the best relationships.  Sketch this result. Put the brain activity DV on the y-axis, Put the three hand-holding conditions on the x-axis. Use two different line or bar graph colors for high and low relationship quality.

 

Link to the hand-holding study with scary movies and pupil dilation (Graff et al., 2021) in PLOSOne.

Link to the hand-holding study with shocks and brain scans (Coan et al., 2006) in Psychological Science. 

Bonobos and chimps remember their friends

A journalist from the Washington Post has reported on a study showing that two ape species (chimpanzees and bonobos) appear to recognize old friends and family members.

You might be asking, how can researchers know what apes remember, when they can't speak? In this case, they did it via a concurrent measures design. Researchers at the Edinburgh Zoo worked with 26 apes who lived in ape sanctuaries in Japan, Scotland, and Belgium. They showed each ape two pictures:

Using a straw with fruit juice to keep their heads still and an infrared camera to track their eye movement, the team showed the animals a pair of side-to-side images for three seconds. The apes were not restrained during the experiment

“It’s an elegant, quite simple way to encourage them to keep their heads still,” said Laura Lewis, a postdoctoral fellow at the University of California at Berkeley and lead author of the study.

One image was of an ape they had previously lived with. The other was of a complete stranger.

On average, the bonobos and chimps that saw both images looked at their old groupmates for about a quarter of a second longer than the stranger.

This study illustrates a common ethical approach to animal research on higher apes. The method is typical of research with primates, in that no individual was coerced into participating. Instead, over time, apes in the sanctuaries learn they can sometimes get special treats (in this case, juice) by going to a certain section of their living compound and working with the researchers. Therefore, chimps and bonobos are only in the study if they volunteer.

a) I already mentioned that this study design was concurrent measures.  Use that information to help you identify the study's IV and DV:

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For the IV: Is it manipulated as within-groups or between -groups?

b) Sketch a graph of the study's results, as described in the last paragraph above. In the original article, the authors created a new DV called "looking bias" which was the difference between how long apes looked at former groupmates minus how long they looked at strangers. Try using that on your y-axis. 

c) Later in the story, the journalist drops another level of the independent variable. Read it, and decide, now what are the three levels of the IV? 

And bonobos and chimps looked longer at old groupmates with whom they had groomed and had positive relationships in the past — longer, that is, at friends than foes, though whether the animals are capable of friendship has been up for debate.

“It’s sometimes called the f-word in primatology,” Lewis said. “But it does seem … a lot of us are in agreement now that you could characterize these long-lasting, positive relationships as friendships.”

d) How might you sketch the result that "bonobos and chimps looked longer at old groupmates with whom they had groomed and had positive relationships"? 

You can compare your figure sketches to those in the empirical journal article, which appears in PNAS. https://doi.org/10.1073/pnas.2304903120 

Bonobos and chimps remember their friends

A journalist from the Washington Post has reported on a study showing that two ape species (chimpanzees and bonobos) appear to recognize old friends and family members.

These two chimpanzees, Qafzeh and Edith, are living at the Edinburgh Zoo. See the journalist's story for a clip of Edith's eye tracking performance in the study. Photo: Kate Grounds/Royal Zoological Society of Scotland

You might be asking, how can researchers know what apes remember, when they can't speak? In this case, they did it via a concurrent measures design. Researchers at the Edinburgh Zoo worked with 26 apes who lived in ape sanctuaries in Japan, Scotland, and Belgium. They showed each ape two pictures:

Using a straw with fruit juice to keep their heads still and an infrared camera to track their eye movement, the team showed the animals a pair of side-to-side images for three seconds. The apes were not restrained during the experiment

“It’s an elegant, quite simple way to encourage them to keep their heads still,” said Laura Lewis, a postdoctoral fellow at the University of California at Berkeley and lead author of the study.

One image was of an ape they had previously lived with. The other was of a complete stranger.

On average, the bonobos and chimps that saw both images looked at their old groupmates for about a quarter of a second longer than the stranger.

This study illustrates a common ethical approach to animal research on higher apes. The method is typical of research with primates, in that no individual was coerced into participating. Instead, over time, apes in the sanctuaries learn they can sometimes get special treats (in this case, juice) by going to a certain section of their living compound and working with the researchers. Therefore, chimps and bonobos are only in the study if they volunteer.

a) I already mentioned that this study design was concurrent measures.  Use that information to help you identify the study's IV and DV:

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For the IV: Is it manipulated as within-groups or between -groups?

b) Sketch a graph of the study's results, as described in the last paragraph above. In the original article, the authors created a new DV called "looking bias" which was the difference between how long apes looked at former groupmates minus how long they looked at strangers. Try using that on your y-axis. 

c) Later in the story, the journalist drops another level of the independent variable. Read it, and decide, now what are the three levels of the IV? 

And bonobos and chimps looked longer at old groupmates with whom they had groomed and had positive relationships in the past — longer, that is, at friends than foes, though whether the animals are capable of friendship has been up for debate.

“It’s sometimes called the f-word in primatology,” Lewis said. “But it does seem … a lot of us are in agreement now that you could characterize these long-lasting, positive relationships as friendships.”

d) How might you sketch the result that "bonobos and chimps looked longer at old groupmates with whom they had groomed and had positive relationships"? 

You can compare your figure sketches to those in the empirical journal article, which appears in PNAS. https://doi.org/10.1073/pnas.2304903120 

Open-label placebos

Photo credit: luchschenF/Shutterstock

As you learn in Chapter 11, a placebo is an inert treatment. In a drug study, a placebo would be a saline or sugar pill. In a psychotherapy study, a placebo might be a supportive conversation with no therapeutic structure. A placebo effect is when people benefit from the inert treatment (compared to no treatment). People's beliefs about some expected treatment might be responsible for why placebos can work. 

Placebos were in the news earlier this year because of the antihistamine phenylephrine, sold commercially as Sudafed PE (and under other names, too). This drug has been available over the counter for decades. It's supposed to reduce nasal decongestion. However, while the drug does not cause harm, it also doesn't work any better than placebo (read more about that here).  The FDA, therefore, recently announced that the drug won't be sold anymore for this purpose.

A recent psychological example of placebos concerns the practice of microdosing of psychedelic drugs, such as psilocybin mushrooms. If you do a bit of online research, you'll likely find multiple anecdotes in which people say that taking very small doses of psychedelics has helped them feel more creative, happy, and fun.  But what do randomized controlled studies say? One journalist summarized three major studies on microdosing. Here 's what they found: 

The two largest placebo-controlled trials of microdosing were published last year, and they both suggest that the benefits people experience [from microdosing] are from the placebo effect. In the studies, volunteers used their own drugs to participate and, unknown to them, received either active doses or a placebo packaged in identical capsules. At the end of several weeks, almost everyone’s mood and well-being had improved, regardless of what they had taken.

Dr. Erritzoe, who ran the other study, found that the drug’s efficacy was tied to users’ expectations. If they took a placebo but thought it was a microdose, they felt better, and if they had an active dose but wrongly guessed it was a placebo, they did not.

Much of the work on placebos makes it clear that, while they can't actually kill bacteria, affect neurotransmitters, or shrink tumors, placebos have real effects on people's felt experience. People taking placebos feel less pain; they may feel less depression or anxiety, and their cold symptoms might improve. 

The true efficacy of placebos has led some doctors and researchers to speculate whether they could be used to treat pain in real patients --after all, placebos have no harmful side effects and they seem to actually work. The one issue, though, is ethics. It's not considered ethical to lie to real patients. Outside of a research study, we can't tell people they're getting a treatment when they are not. 

With that in mind, the most fascinating recent research on placebos shows that they can work even if people know they are taking placebos. One of the scientists who does this work summarized the issue in this article. Here's his summary of two studies: 

As it turns out, placebos can work even when patients know they are getting a placebo. In 2010 my colleagues and I published a provocative study showing that patients with irritable bowel syndrome who were treated with what we call open-label placebos — as in, we gave them dummy pills and told them so — reported more symptom relief compared with patients who didn’t receive placebos. (These placebos were given with transparency and informed consent.) In another blow to the concept that concealment is required for placebo effects, my team recently published a study comparing open-label placebos and double-blind placebos in irritable bowel syndrome and found no significant difference between the two. A medical myth was overthrown.

Might open-label placebos work on other symptoms, too? Here's an article about how open-label placebos can help people feel less guilty. While guilt can motivate us to do the right thing, guilt is also an emotion associated with depression and PTSD (at least in the West).  It can be both helpful and unhealthy. 

In the study, 112 healthy volunteers between the ages of 18 and 40 took part. Their guilt was measured at the beginning using questionnaires including the state shame and guilt scale (SSGS). This questionnaire asks people whether they feel remorse or bad about something they’ve done. Next, the participants did an exercise intended to make them feel more guilty. The exercise involved writing a story about a time they had treated someone they loved unfairly.

The participants were then divided into three groups. One group received a “deceptive placebo”: a blue pill they were told was a real drug. Specifically, they were told that the pill contained phytopharmacon, a substance designed to reduce the feeling of guilt by making whoever took it feel calmer.

Another group received an “open-label placebo” – the same blue pill, but this group was told it was a placebo. They were told that placebos benefit many people through mind-body self-healing mechanisms.

The third group did not receive any treatment at all. This was the “control” group.

After getting the treatment, the guilty feelings were measured using the same questionnaires to see whether the deceptive placebo or open-label placebo was more effective than no treatment.

The main outcome reported in the study was that the deceptive placebo and the open-label placebo ... were more effective at reducing guilt than no treatment.

a) Looking at the study above on guilt, what was the IV? What was the DV? What kind of experiment was conducted (Was it posttest only, pretest-posttest, repeated measures, or concurrent measures?)

b) Sketch a graph of the results of the guilt study. Be sure to plot the pre and post test data.
You can compare your graph to Figure 2 of the original empirical article, here. 

c) Here are the instructions for the open-label placebo condition, which I copied from the original empirical article's supplementary information. What do you think of these instructions? Are they manipulating just an open-label placebo, or might they be inducing demand characteristics? How might you change these instructions to reduce the demand? 

[This] is a placebo and therefore does not contain any medicinal active ingredient.” (Study personnel shows the package and removes the pill from the blister). “The pills consists only of sugar (lactose, sucrose, glucose) and stabilisers. However, we know that placebos are very effective. Openly-administered placebos have been used for various clinical disorders and problems and have shown very good efficacy. This has been shown in studies on pain and depression, among others. In the case of depression, feelings of guilt in particular are a major component. Placebos work through expectation and learning processes. The body reacts automatically and symptoms are reduced. An open, positive attitude towards placebos can be helpful, but this is not necessary for a positive effect. We expect the activated guilt feelings to be noticeably reduced again and use it because we want to use its positive emotional effect on the guilt feelings. In doing so, we do not expect any side effects. Here is the pill" (Study participant takes the pill.)

 

It matters how professors talk to students....

Students imagined asking a question in class, and getting one of three responses from the professor. Photo credit: Ground Picture/Shutterstock

This example gives you practice extending a simple experiment (Chapter 10) into a factorial design (Chapter 12).

Think back to the last time you asked a question in your research methods class--or any other class. What was the professor's reaction to your question? Maybe they said, "That's a great question!" or maybe they said, "Didn't we just go over this?" As you probably know, it matters how professors respond--you might feel validated, encouraged, rejected, or ignored, all of these, depending on the professor's tone. 

A group of researchers studied how a professor's responses can make students feel, especially in STEM (science, technology, engineering, and math) courses. They used a scenario method in which students imagined asking a question in class. The researchers summarized their research for a general audience on the SPSP news website.  Here's why they were interested in this topic: 

Students' confidence and belonging may influence their interest and motivation in STEM. Confident people believe that they can succeed. Belonging refers to how much people feel they fit in and are accepted. We expected that when students imagined their instructors responding to their questions in a positive way, students would feel more confident and accepted, which would boost their motivation and interest in STEM.

Study 1 (Chapter 10)

The researchers conducted five studies. Let's talk about Study 1: 

College students imagined receiving a response from an instructor to a question they asked.  The response was either positive ("That's a great question, I'm glad you brought that up"), negative ( "I'm not sure why you're asking this question, we went over this already") or neutral ("We're actually out of time today, so please hold your question till next time"). They then reported their feelings of confidence, belonging, interest in joining the research lab, and intentions to recruit others to the lab.

When students imagined receiving a positive rather than a negative or neutral response from their STEM instructor after asking a question, they felt more confidence and belonging in STEM.

a) Classify the variables in this study: (Hint: There's one IV and four DVs)

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For IVs: is it within groups or independent groups?

b) Which kind of experiment is Study 1: Posttest only? Pretest/posttest? Concurrent measures? Repeated measures? How do you know?

c) Sketch a bar graph or line graph of the results they described (select one of the four DVs). 

Study 2 (Chapter 12)

In another study, the researchers compared whether the feedback's effects depended on the type of student. Here's the theoretical reasoning: 

Some students may be more vulnerable to these effects because they doubt their belonging or ability in certain areas. For example, being 'bad at math' is a common stereotype that negatively affects women's experiences in STEM. Women often feel uncertain about their belonging in STEM and perform worse when reminded of negative gender stereotypes about women's math abilities. Based on these ideas, we reasoned that women, compared to men, would react more strongly to imagining receiving certain types of instructor responses to their questions in STEM contexts.

Here's the study's results:

Indeed, compared to men, women reported greater confidence, belonging, and intentions to join a STEM lab after imagining their instructor responding to their question with, "That's a great question, I'm glad you brought that up." Women also reported lower confidence, belonging, and intentions to join a STEM lab after imagining a negative instructor response or even a neutral instructor response of, "We're actually out of time today, so please hold your question till next time."

d) Classify the variables in this study: (Hint: The IV and three DVs are the same as before, but now there is a participant variable (PV)/independent variable :

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV, PV or a DV?	For IVs and PVs: is it within groups or independent groups?

e) Which kind of experiment is Study 2: It's a factorial design, so state it like this ___ X ___ and state whether it's an independent groups factorial, within groups factorial, or mixed factorial design. 

f) Sketch a bar graph or line graph of the results they described (select the confidence DV). I recommend putting the "feedback condition" variable on the X axis and the Gender PV as a line or bar color. 

g) Can you detect an interaction in the data? How about a main effect for type of feedback? 

Do TikTok dance videos influence body perceptions?

Seeing different size models on TikTok videos affected people's body esteem.
Photo: AlessandroBiascioli/Shutterstock

Have you ever watched dance challenge videos on TikTok? If so, you probably noticed the skill and appearance of the dance model.  Some psychologists have started to study how TikTok dance models affect viewers.  A set of three studies asked, can the size of the dancer on TikTok influence the body satisfaction of a viewer?  Journalists at PsyPost wrote up a summary of this empirical work. The research was originally published in the journal, Computers in Human Behavior; may be paywalled. 

The theory behind the study is related to the potential of social media to impact body esteem:

Media, especially its portrayal of the “thin ideal,” where extremely slim bodies are considered attractive, has a significant impact on shaping these negative body perceptions. With the rise of social media platforms like TikTok, researchers are curious about how this new form of media might influence people’s body image.

Here's a description of the first study. As you read this description, try to identify the variables in the study:

...the researchers recruited 262 women aged between 18 and 25 years for an online experiment.  The participants were randomly assigned to watch TikTok videos featuring thin dancers, large dancers, or videos of animals (the control condition). [Both before and after] watching the videos, participants rated their weight satisfaction, overall appearance satisfaction, and body shape satisfaction.

Body satisfaction significantly increased after watching videos featuring large dancers and control videos (videos unrelated to body appearance). However, after watching videos of thin dancers, body satisfaction showed a slight decrease.

a) Use this chart to identify Study 1's variables (Hint: There are one IV and three DVs)

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For IVs: is it within groups or independent groups?

b) Which kind of experiment is Study 1: Posttest only? Pretest/posttest? Concurrent measures? Repeated measures? How do you know?

c) Sketch a bar graph or line graph of the results they described (select one of the three DVs). Don't forget to plot both pretest and postest data. 

The researchers did Study 2 to replicate Study 1:

Study 2 largely replicated the findings from Study 1. Watching videos featuring large dancers led to increased body satisfaction, while watching videos featuring thin dancers resulted in decreased body satisfaction.

Study 2 also measured the potential mechanism, or mediator, of social comparison (how often someone compares their appearance to those of others.).  Study 2 found: 

A majority of participants in the thin dancer condition engaged in upward comparison (comparing themselves negatively to the dancers), while a majority of participants in the large dancer condition engaged in downward comparison (comparing themselves positively to the dancers).

One potential internal validity threat to both Study 1 and Study 2 is that participants were told the study was about body satisfaction. In addition, the pretest-postest nature of the design may have nudged participants to report changes in line with the study's goals.

Study 3 attempted to correct this issue with a posttest-only design, and by telling participants that the study was actually about how engaging different TikTok videos are. 

d) What is the name for this type of internal validity threat? 

e) Why would a posttest-only design help control for this internal validity threat?

Here's the journalist's summary:

In Study 3, the researchers recruited 375 women aged between 18 and 25 years and randomly assigned them to watch either TikTok videos of thin dancers or TikTok videos of large dancers. Participants were told that the study aimed to investigate the engagement of TikTok videos. After watching the videos, the participants completed the same measures of body satisfaction used in the previous studies.

Even after concealing the true purpose of the study, the results consistently indicated that watching videos of thin dancers led to lower body satisfaction compared to videos of large dancers.

f) Reread the summary of Study 3 and identify its variables on the table below:

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For IVs: is it within groups or independent groups?

g) Graph the results of one of the dependent variables in Study 3. 

h) What do you think of this set of studies? Ask questions about the four aspects of validity.

Specifically, when you address statistical validity, use terms like replication and effect size.
When you address internal validity, think about how the study rules out most internal validity threats.
When you address construct validity, ask a question about the measurement of the dependent variables.
When you address external validity, think about the samples used in these studies.

Here's how the researchers concluded an interview about their studies:

“I would like to add that this study shows that watching TikTok videos is not universally negative and that there are positive and negative effects,” Joiner said. “The findings also raise the possibility that TikTok could change its algorithm and show a wider variety of body types, and this could have a major public health benefits. Body dissatisfaction is major risk factor for eating disorders and eating disorders have the highest mortality rate of all psychiatric diseases due to medical complications or suicide.”

Easing the angst of upsetting news

The study tested two different types of stories to see which might ameliorate the stress of reading about tragic news events. Photo credit: Nomad_Soul/Shutterstock

This post contains descriptions of violence in the news.

Have you ever been saddened or dismayed by a news about a terrible event, such as a natural disaster, mass shooting, or a crime? You're not alone; such stories can make us emotional. One psychology researcher, Kathryn Buchanan, was inspired to study people's reactions to news events by her own experience. She was listening to news about the mass shooting at the Ariana Grande concert in 2017. She felt upset at the news, but she also felt some relief, when she later encountered news about examples of people's kindness in the shooting's aftermath. 

This journalist's account in the Washington Post describes how Dr. Buchanan turned her experience into a research idea:

She began to contemplate whether being exposed to heartwarming content could counteract the known negative impacts of consuming harrowing news stories. Common symptoms include heightened stress, hopelessness, anger, anxiety and depression.

So she started a years-long study in 2017,[...]. [The research team] split 1,800 participants into several groups — each of which were shown news clips or articles about a recent U.K.-based terrorist attack or a similar catastrophic event. Some of the groups were solely exposed to upsetting news, while others were subsequently shown a story about [either] a heroic act or a more lighthearted subject.

Here is a description of some of the results:

The group that was strictly shown unsettling news reported increases in negative emotions and decreases in positive feelings. That group also indicated feeling more pessimism about humanity and society. Meanwhile, participants who were also shown stories of kindness reported feeling less negative emotions, coupled with increases of positive emotion. That group reported a more optimistic view of the world.

Participants used two scales from 1 (meaning very slightly or not at all) to 5 (meaning extremely) to assess the level of negativity and positivity they felt before being exposed to the content, as well as after.

Repeatedly throughout their research, Buchanan and Sandstrom saw that uplifting news can provide an emotional buffer against distressing news. They also found that “there’s something special about kindness in particular,” Buchanan said, noting that while amusing stories (such as swearing parrots or classic jokes) diminished the effects of upsetting news, stories about acts of kindness were even more powerful.

a). See if you can classify the variables in the study. Here's a hint: There were four dependent variables and one independent variable. 

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For IVs: is it within groups or independent groups?

b) Which kind of experiment is Study 1: Posttest only? Pretest/posttest? Concurrent measures? Repeated measures? How do you know?

c) Look at the description of the results, above, for level of positivity only. Sketch a graph of the results they described. Here's a hint: it's important to find a way to include the pretest and posttest scores on positivity. 

d) Think about the four big validities and apply them to this study. Specifically, ask one question to help you interrogate external, internal, construct, and statistical validity.

I appreciated that, in the news story, the journalist mentioned other studies that also measured people's reactions to negative news in the media. Journalists do not often talk about recent studies in the context of an entire literature, but this one did! You can read examples of conceptual replications of this study in the journalist's story (scroll toward the end).  

The empirical study can be read for free here at PLOS One. 

 

"Want to raise kind, generous kids? Take them to an art museum"

Did the study manipulate exposure to art museums?
Photo: AChubykin/Deposit Photos

On my first read of this headline, I was skeptical. "Probably a correlational study behind this one," I thought. As it turns out we are actually looking at an experiment. However, there's another problem with the journalist's coverage of this study. Read on.

The journalist writes for Inc.com, which  summarizes research for the business world. The headline was "Want to raise kind, generous kids? Take them to an art museum" The original empirical article appeared in Psychological Science. Let's work through it. The introduction starts out:

Everyone wants to raise successful kids, but a big part of being successful is being a kind and generous person. So how can parents give their kids the best while also doing their best to ensure they raise decent, thoughtful little people? 

They introduce the idea that art might make kids more generous: 

You may be wondering what possible connection a day viewing Van Gogh or Picasso could have to raising an empathetic kid. Certainly, museums make children more worldly and cultured, but how do they make them kinder? The link, according to the new study, recently published in Psychological Science, is awe. 

a) Before you read more, make a prediction. Given this introduction, what independent variable do you think the researchers probably manipulated? What variable do you think was measured?

Now let's read the details. There were two studies. In the first one, 

..the research team asked 159 volunteers aged 8 to 13 to watch short movie clips. Some of these clips were neutral, others cheerful, and others awe-inspiring. . 

In the second study, about 350 kids participated. The researchers used one of the same three clips for the awe variable (neutral, joy, and awe), but in this case, the generosity variable was operationalized differently:

The researchers then asked the kids [in Study 1] to determine how many items in a list of foods should be donated to a food drive for needy families. Alternately [in Study 2], the kids were given the option of donating their reward for participating in the study--a ticket to a local art museum--to a refugee family. 

(By the way, the original article reports that "The awe-eliciting clip was from the film Song of the Sea, in which the main character turns into a seal and transforms the city while flying above it.")

Back to the Inc.com story for the results: 

"Children who watched the awe-inspiring video chose to count 50 percent more items for the food drive than children who watched the joy-inspiring clip and more than twice as many items as children who watched the neutral clip. [in Study 2], children in the awe-inspiring condition were also 2 to 3 times more likely to donate their museum tickets than children in the joyful or neutral conditions," reports the Association for Psychological Science blog. 

b) Take a moment to classify the two variables in Study 1:

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For IVs: is it within groups or independent groups?

c) Which kind of experiment is Study 1: Posttest only? Pretest/posttest? Concurrent measures? Repeated measures? 

d) And now for Study 2:

Variable name	What are the variable's possible levels?	Is this manipulated or measured?	Is this an IV or a DV?	For IVs: is it within groups or independent groups?

e) Which kind of experiment is Study 2: Posttest only? Pretest/posttest? Concurrent measures? Repeated measures? 

Look back at the description of the study. Did either Study 1 or Study 2 manipulate "taking kids to an art museum"? Nope. In fact, Study 1 took place online (kids watched video clips on their own computers). And Study 2 took place in a science museum. Neither of the studies manipulated art museum exposure. Art museum tickets were used in Study 2, but only as a way of operationalizing generosity.

That factual error didn't stop the journalist from asserting:

Parents ... can always go to their local art museum to give their kids a guaranteed dose of awe. 

Guaranteed dose of awe? That conclusion does not really follow from this study!  So here are a couple more questions. 

f) What could be an alternative headline for the Inc.com story--one that is more accurate to what the study actually manipulated?

g) The journalist appears to have generalized from one setting (watching video clips) to another (going to an art museum). Which of the four big validities is the journalist treading on here? 

h) Here's a construct validity question: The two studies operationalized the dependent variable, "generosity" in two different ways. Were each of them a good way to measure this variable? Why or why not? What are the strengths of operationalizing the variable more than one way?

i) Now for a question related to internal validity: The main independent variable was the content of the film clip the children saw. To ensure internal validity, what should the researchers hold constant about the film clips used in the three conditions?