Module 5: Ch 11 More on Experiments

regression threat

Extreme scores return towards average.

Measurement Error Correct label: validated scales Correct label: increased number of measurements Individual Differences Correct label: increased number of participants Correct label: within-groups design Situation Noise Correct label: experimental control

For each experimental impediment, match the solution that would best prevent within-group variability from obscuring group differences.

false

Increases in within-group variability can lead to the illusion of covariance. Null effects are very uncommon in experiments.

attrition threat

More participants decline to continue from one group or condition than from another.

true

Not enough between-group differences can result in a null effect. Null effects can occur in any experiment.

noise; measurement error; individual differences; situational noise

Null effects can be caused by excessive amounts of unsystematic variance, or -. This within-group variance can be caused by a dependent variable with poor reliability or in which insufficient data is collected. This is referred to as -.. It can also be caused by - if extreme scores within the participants obscure changes brought on by the manipulation. Finally, outside factors in the experimental setting can influence participants, otherwise referred to as -.

maturation threat

Participant behavior changes spontaneously over time.

testing threat

Participant's future performance is changed because of interaction with a measure.

history threat

Participants' responses are altered by an event outside the control or interest of the study.

false

To prevent a history threat, the comparison group should be followed at a different time or in a different location than the treatment group. A study typically only has one threat to internal validity.

true

A regression threat can lead to a significant finding, resulting in a Type I error. If participants withdraw from a study in an unsystematic way, there is likely no attrition threat.

ceiling effect

Antonio is studying how different pamphlets promoting local charities may increase volunteer behavior in high school students. He creates one pamphlet that shows students having fun and being social while volunteering, and another that discusses the benefits volunteering has for college and future careers. Antonio has participants read one pamphlet or the other and then fill out a measure about volunteering intentions for the future. Antonio's measure asks how likely participants are to volunteer in the future from 1 ("I will not volunteer at all") to 5 ("I will likely volunteer in the future"). When looking at his results, Antonio finds that both groups scored very high on his measure and there is no difference between the two groups. What problem is most likely causing a null effect in the study?

testing threat

Candice has created an intervention to see if children using their imaginations increases their spatial reasoning abilities. She randomly recruits kindergarteners from city schools and pretests them with a spatial reasoning task. After a week of her intervention training she measures the children's reasoning again using the same survey. All the children scored better on the assessment the second time compared to the first. What threat to internal validity may be present?

Dependent Variable Correct label: victim blaming helping intentions Independent Variable Correct label: environment Control Variable Correct label: gender of victim events

Doris is interested in whether there is a difference in victim-blaming and victim-helping attitudes when it comes to bullying based on environment (in person versus online). Doris creates a fictional scenario in which someone is bullied in a classroom or on a social networking site. In each case, the events and gender of the victim are not changed. Doris then measures how much people blame the bullied person and how willing they are to step in and help the person. Match each aspect of the experiment to the variable it best represents.

null effect; weak manipulation; insensitive measure, ceiling effect; floor effect

If the researcher concludes that there is no difference in the dependent variable, they are concluding a(n) null effect. This can either be the truth (the variables are not related) or could be due to design flaws in the experiment. A(n) weak manipulation can occur when the change in the independent variable is not strong enough to affect the dependent variable. The dependent variable might not be responsive enough to detect change from the independent variable; it could be a(n) insensitive measure. Scores from the dependent variable can also spontaneously cluster near the top of possible scores, known as the ceiling effect, or bottom, known as the floor effect, which can make covariance undetectable.

control

In all studies, experimenters keep certain factors constant in order to be sure the independent variable is what is affecting the dependent variable. Though these factors are not true variables, they are often referred to as what type of variables?

selection-attrition threat

Jermaine is curious if the way students take notes affects their academic ability. Jermaine recruits two groups of students from classes on campus: one group of high achievers and one group of low achievers. He then asks both groups to take notes in the exact same way to see how it will affect their scores. During the study, several students in the low-achieving group drop out from the study, as they find the note-taking task difficult. At the end of the study, the results for each group are similar. What threat to internal validity may be present?

Comparison Group Correct label: maturation threat Correct label: history threat Correct label: regression threat Posttest-Only Design Correct label: instrumentation threat Correct label: testing threat Remove All Scores from the Participant Correct label: attrition threat

Match each threat to the appropriate prevention method.

weak manipulation

Nikki created an experiment to investigate if viewing pictures on different types of social networking sites would cause people to feel more or less social. She had participants look at pictures on a web page that had either a Facebook logo or a Google+ logo, but everything else about the website was the same. Nikki measured sociability on a scale of 1 to 10 that asked several questions, such as "How interested are you in meeting other people right now?" and "How interested are you in talking to your friends right now?" Nikki found that both groups had sociability averages around 6, and there was no difference between the groups. What problem is most likely causing a null effect in the study?

observer bias; masked design; double-blind study

The influence of a researcher's own experiences upon how she views outcomes in an experiment is called -. This can be prevented by making sure that researchers are unaware of which participants are in which conditions, called a -. To further reduce this issue, the conditions can sometimes be unknown to both the experimenter and the participant, called a -.

instrumentation threat

The way the researcher measures a variable changes over the course of an experiment.

experiment; independent; condition; dependent

When researchers manipulate a variable in a study, they are carrying out a(n) -. The manipulated variable is often called the - variable. A manipulated variable always has more than one level or -. Researchers measure the - variable to determine the effect of the manipulated variable.