Everyday Research Methods

"'Freshman 15' is a myth, study says"

Going to college adds 15 pounds, right? Not according to this report. The story reports that a group of researchers interviewed over 7,000 people aged 17 to 20. The study found that freshmen in the sample gained, on average, 2.5 to 3.5 pounds in the past year. However, the study

also found that young adults who do not go to college gain weight — about a half pound less than college students of the same age. In other words, college attendance has almost no effect on weight gain.

a) What kind of study is this? An experiment, or a quasi-experiment?

b) What are the independent and dependent variables?

c) The story reports no difference in weight gain between college students and non-college students. This is a null effect. Do you think this null effect is attributable to some obscuring factors (such as those listed in Table 10.2)?

Suggested answers

a) This is a quasi-experiment, because the main independent variable (being in college or not) was not manipulated by the experimenters. People can't be randomly assigned to attend college or not. This is an example of a non-equivalent control group design.

b) The study's independent variable was attending college or not, and the dependent variable was weight gain.

c) The potentially obscuring variables that are relevant to this study are situation noise (there are likely many situational differences within the various college and non-college environments) and individual differences. However, the study used so many people--over 7,000--and this most likely counteracted the large degree of within-groups variance. Therefore it might be safe to conclude that there truly is no difference in weight gain between college and non-college young adults.

Attention to detail in Body Dysmorphic patients

Body dysmorphic disorder (BDD) is a mental illness in which people focus critically and negatively on the fine details of their bodies. The disorder is linked to eating disorders and to behaviors such as repeated checking in the mirror or embarassment about one's body.

Science News Daily reports a study that used fMRI techniques to study brain activity in visual processing regions for 14 people with BDD and 14 healthy controls. They showed all the participants a series of photos of houses that were highly detailed, normal, or less detailed.

According to the news story,

The researchers found that the BDD patients had abnormal brain activation patterns when viewing pictures of the less-detailed houses: The regions of their brains that process these visual elements showed less activation than the healthy controls.

This study is an example of a factorial design.

a) What are the independent variables in this study? How many levels are there for each IV?

b) What is the dependent variable in this study?

c) Sketch a graph of the results of the study, according to the description quoted above. You may not know the exact results for each cell, but you can estimate what the pattern looks like.

d) According to the results described (and the graph you drew), do you detect an interaction in these results? How about main effects?

Most Americans report 2 friends

A report suggests that among Americans, the average number of close friends people report having is just two. You can read an msnbc report of the study here.

The study reports that the researchers

...surveyed more than 2,000 adults ages 18 and older from the nationally representative Time-sharing Experiments for the Social Sciences (TESS) program. The Internet surveys were conducted between April 23 and May 5, 2010.

The story also reports that people were

...asked to list the names of people they had discussed "important matters" with over the previous six months. If respondents said "none," they asked whether this was because they didn't have any important matters to discuss or no one with whom to discuss them in the past six months.About 48 percent of participants listed one name, 18 percent listed two, and roughly 29 percent listed more than two names for these close friends.

The average number of confidants listed was 2.03, the story reports.

 

a.) The headline of the study was, "You gotta have friends? Most have just 2 true pals." What kind of claim is made in this headline?

b.) Which two validities do we prioritize when evaluating this kind of claim?

c.) Based on the information from the article quoted above, how well does the study fulfill these two validities?

"Video games help with creativity in boys and girls"

A story in USAToday has reported that "Video games help with creativity in boys and girls."

a)  What kind of claim is being made in the headline?

 

Here is a quote from the article about the study's method:

The children were given creative thinking tests – shown a drawing and asked to expand on it or comment on what it means – and then asked about their use of cellphones, computers, the Internet and video games. Only with video games was there a correlation to creativity, the researchers say, and that was true for boys and girls, and across all races.

b.)  Given this description, what were the variables in this study?  Were they measured or manipulated?

c.) Is a causal claim justified here?  Go through the three rules for establishing causation and draw a conclusion. 

d.) What can you say about the construct validity of this study? 

e.) When the article says that the result was true for boys and girls, describe what this means.  Use the word "moderate" in your answer.

 

Toward the end of the article, the journalist reports:  

The researchers have proposed a study to test whether a special touch-screen computer can help kids, ages 3 and 4, with school readiness skills.

f.) How might researchers design a study that could show that a touch-screen computer can actually cause improvement in 3 and 4 year olds' school readiness?

 

Suggested answers:

a.)  This is a causal claim (the verb "help" is a causal verb.)

b.)  The description suggests strongly that both variables were self-reported by the children, therefore both were measured variables.

c.) The study's result suggests there is covariance--video game use apparently associated with creativity scores. 

Given that both variables were measured at the same time, we cannot establish temporal precedence.  It is possible that more creative kids become more likely to play video games, or also that kids who play video games become more creative.  We cannot tell from this design.

It is not clear from the description which, if any, third variables were controlled for (internal validity).  Perhaps intelligence or family income are associated with both video game use and creativity.  You would have to consult the original study to see.  However, because it is not an experiment, this study cannot control for all possible third variables. 

In short, a correlational design like the one implied in the journalist's description cannot establish or support the causal claim in the headline. 

d.) To interrogate construct validity, you ask how well each of the variables was measured.  How good is the drawing interpretation test for measuring creativity?  Is this measure a well-established measure of creativity? (We can consult the original article to be sure). And how accurate is it to ask children to self-report about their use of electronic media? 

e.) This phrase suggests that gender did not moderate the relationship between creativity and video game use.  In other words, the correlation coefficient (between creativity and video games) was the same in both gender groups. 

f.) The best design for this research goal is an experiment, in which the researchers would manipulate the use of a touch-screen computer (for example by randomly assigning some chidlren to use this equipment and other children to use some control equipment) and then measuring school readiness. 

A recent case of research fraud

A social psychologist is in the news this month.  He admitted to fabricating and falsifying data: the data in dozens of published studies by this scientist are suddenly suspect.  

Here's a version of the story in the New York Times. 

a.) Data falsification and fabrication are two slightly different problems.  How are they different?  (see Chapter 4)  Which one is discussed in this story? 

b.) Reflect on why falsifying and fabricating data count as ethical violations.  Who is harmed by these actions?

c.)  The article discusses how the falsification case hurts psychology's public reputation as a science.  What do you think?  Do you think that stories like these will damage psychology's public reputation?

d.)  Data falsification also damages the process of science itself.  Reflect on how falsifying or fabricating data affects the theory-data cycle, the most important one in science. 

e.) After you read this article, think of changes that the field could make to the peer review or writing process that would make it less likely that psychologists would fabricate or falsify data.

Negative parenting and aggressive behaviors

This Fox News article describes the results of an empirical journal article (published in Child Development), presenting the results for a popular audience.  Here is an excerpt from the journalist's description:

The researchers assessed the infants' temperaments and the mothers' parenting styles in the first six months of life (by observing the pair during feeding time) and during the toddler years... The researchers followed up with the mothers and the children's teachers when their kids were in kindergarten (ages 5 and 6).

The results showed that

...children who exhibit aggressive, defiant and explosive behavior by the time they're in kindergarten very often have tumultuous relationships with their parents from early on.

What they didn't find was a correlation between difficult early-life behaviors of the child (if the child was irritable or quick to switch mood in his or her first six months) and later aggressive actions and attitudes.

a.) Is this study correlational or experimental?  How do you know?

b.) There are four variables described in the exerpt above.  What are they?

c.) Following the model in Figure 8.3, sketch the results of this study.  The exact correlations are not given here, but you can estimate which correlation is strongest. 

d.) The headline of the story is "Negative parenting starts aggressive personalities early."  How does the study establish temporal precedence for this causal claim?  Can the study establish internal validity? If so, how?

There's more to the study than revealed here--read the full journalist's coverage or look for the original study in the journal Child Development in your library.

 

Michael Gazzaniga short video

Here is a link to a short video featuring Dr. Michael Gazzaniga discussing his experiences as a scientist.  His work on split brain patients is described in Chapter 12.