Chapter 3- Interrogation tools for consumers of research

Question 1

Variable

Answer 1

Something that’s being studied that changes from person to person (trait, condition, etc.). It must have at least 2 levels or values. Variables can be manipulated, but some variables can’t (just have to be measured).

Question 2

Constant

Answer 2

A variable that remains the same (only has one level). You want to control as many variables as possible that could confound your outcome. In a study with only American participants, nationality would be a constant

Question 3

“60% of high school graduates pursue college”. What is the variable and what is the constant for this statement?

Answer 3

Variable- pursuing college. You can have one person who pursues college and one person who doesn’t.
Constant- high school graduates. Every person in the study was a high school graduate.

Question 4

“Men who frequently use social media have less close friends”. What is the variable and the constant for this statement?

Answer 4

Variable- social media use, amount of close friends

Constant- men

Question 5

What types of variables can’t be manipulated?

Answer 5

Some variables are set and can’t be changed, like age (can’t “assign” someone to be older or younger). Other variables would be unethical to manipulate, like the effects of malnutrition on children. We have to measure these types of variables as they occur naturally.

Question 6

What is the difference between a conceptual variable and an operational definition?

Answer 6

An operational definition is how a conceptual variable or construct is defined. A conceptual variable must be defined to be successful. For example, measuring “academic success” in a study would be too subjective, since that can mean many different things to different people. It must be defined using a certain GPA value or other objective measure.

Question 7

What are the 3 types of claims?

Answer 7

1. Frequency claims
2. Association claims
3. Casual claims

Question 8

Frequency claims

Answer 8

Describes a particular rate or degree of a single variable (only involves one measured variable, no manipulation). Example: “1 in 4 working students skip class due to their job”. The variable is skipping class, and the constant is working students.

Question 9

Association claims

Answer 9

State that one level of a variable is likely to be associated with a particular level of another variable. These claims involve at least 2 measured variables. Example- “people with higher incomes spend less time socializing”. The variables are income and socializing, and the constant is higher income.

Question 10

Types of associations (2)

Answer 10

1. Positive association/correlation

2. Negative association/correlation

Question 11

Positive association/correlation

Answer 11

As one variable increases/decreases, the other variable changes in the same direction. The slope of a graph with a positive association goes upward.

Question 12

Negative association/correlation

Answer 12

As one variable increases/decreases, the other changes in the opposite direction. A graph of this association would show a downward slope.

Question 13

How can we make predictions based on associations?

Answer 13

Both positive and negative associations can help us make predictions, but zero associations can’t. We can’t use associations to make causal claims. The stronger the association between the two variables, the more accurate the prediction will be. For example, “people who drink more coffee are happier”. We can’t make causal claims from this, drinking coffee doesn’t necessarily make people happy. We can use this information to predict that a person who drinks coffee might also be happier.

Question 14

Causal claims

Answer 14

Says that one of the variables is responsible for changing the other. Each causal claim has 2 variables.Example- “social media use leads to anxiety”.

Question 15

Examples of verbs used in causal claims

Answer 15

Uses verbs like cause, enhance, affect, decrease, change. A causal claim can also contain more tentative language- could, may, seem, suggest, sometimes, potentially- these verbs are still considered a causal claim. Advice is also a causal claim- it implies that if you do X, then Y will happen (“Boost your salary by hitting the gym”). Causal claims sound more exciting, which is why journalists can sometimes use them incorrectly

Question 16

Examples of verbs used in association claims

Answer 16

Is at higher risk for, may predict, is tied to, link, associate, correlate, predict, tie to, and be at risk for.

Question 17

What are the 4 big validities?

Answer 17

1. Construct
2. External
3. Statistical
4. Internal

Question 18

Construct validity

Answer 18

How well variables in the study are measured or manipulated (operationalized). If you want to know how tall someone is, measuring their height would be useful. However, using their shoe size would be less accurate and would result in low construct validity. The measure of the variable should yield similar scores on repeated testing.

Question 19

External validity

Answer 19

The extent to which the results of a study generalize to some larger population. Using a sample from one private school in NYC would not be very generalizable to all college students.

Question 20

Statistical validity

Answer 20

The extent to which the data support the conclusion. How strong is the association, and what is the statistical significance. The value we get from a single study is not the objective truth- it’s an estimate of that value in some population. Statistical validity improves with multiple estimates (when studies are done more than once).

Question 21

Internal validity

Answer 21

In a relationship between one variable (A) and another (B), the extent to which A, rather than some other variable (C), is responsible for changes in B. Anxiety from social media- maybe people use social media to procrastinate doing other tasks, resulting in anxiety
In this case, social media itself is not actually causing the anxiety.

Question 22

“80% of college students have been depressed over the past year”. Which of the validities are we concerned with for this statement?

Answer 22

Construct validity- how was depression measured in the study? Measuring someone’s depression now is not a good representation of how they have felt over the past year.
External- think about who the participants were and how they were recruited. Using just students from one university doesn’t represent all American college students.
Statistical- how well the data support the claim. How accurate is this percentage? Margin of error takes into account sampling error

Question 23

“Romantic partners who express gratitude are three times more likely to stay together”. Which of the validities are we concerned with here?

Answer 23

Construct and external, statistical- strength and significance

Question 24

Type 1 error

Answer 24

False positive- deciding that there’s a relationship when there actually isn’t.

Question 25

Type 2 error

Answer 25

False negative. Deciding that there isn’t a relationship when there actually is.

Question 26

What validities are we concerned with for causal claims?

Answer 26

Construct, external, statistical validity. We must also interrogate the causal claim using different criteria.

Question 27

3 criteria for causation

Answer 27

1. Covariance
2. Temporal precedence
3. Internal validity

Question 28

Covariance

Answer 28

If two variables are related (correlated). Necessary to begin to establish a causal relationship.

Question 29

Temporal precedence

Answer 29

Cause must be shown to have occurred before this study using a randomly assigned experiment.

Question 30

Independent variable

Answer 30

Manipulated variable, used to create experimental and control groups in an experiment. This is the “cause” variable.

Question 31

Random assignment

Answer 31

Ensures that each participant has the same opportunity to be assigned to any given group.

Question 32

Internal validity

Answer 32

The study’s method ensures that there are no plausible alternative explanations for the change in B, A is the only thing that has changed.

Question 33

Dependent variable

Answer 33

The variable that is measured and affected by the change in the independent variable in an experiment. This is the “effect” variable.

Question 34

Which of the 4 big validities is most important?

Answer 34

It depends on what kind of claim the researcher is making and the researcher’s priorities. When studying if a review class impacts performance on a test, you want demographics to be the same. External validity might not be high. However, internal validity could suffer if you attempt to prioritize external validity (generalization)

Question 35

Measured variable

Answer 35

One whose levels are observed and recorded. This includes gender and hair color. Variables like depression and stress are measured using questions to represent the various levels.

Question 36

Manipulated variable

Answer 36

A variable a researcher controls, usually by assigning study participants to the different levels of that variable. They are assigned to groups doing different activities or different medication dosages

Question 37

Claim

Answer 37

An argument someone is trying to make. Researchers make claims about theories based on data, journalists make claims when they report on studies they read in empirical journals

Question 38

Correlate

Answer 38

Variables that are associated are said to correlate- when one variable changes, the other variable tends to change as well.

Question 39

Zero association

Answer 39

No association between the variables (zero correlation). A scatter plot for this association would have no slope, a line drawn through the points would be almost horizontal.

Question 40

Generalizability

Answer 40

How well participants represent the intended population

Question 41

Confidence interval

Answer 41

The confidence interval is a range designed to include the true population value a high proportion of the time- demonstrates the precision of the estimate. Ex- the 39.2% point estimate had a 37.0-41.4 confidence interval. This means that the number of teens who text while driving is as low as 37% or as high as 41.4%, although the interval could miss the true value.

Question 42

Which validities do we use to interrogate frequency claims (3)?

Answer 42

1. Construct
2. External
3. Statistical

Question 43

Which validities do we use to interrogate association claims? (3)

Answer 43

1. Construct
2. External
3. Statistical

Question 44

Construct validity of association claims

Answer 44

Since there are 2 variables, you must assess the construct validity of each variable. Both variables must be measured well to be able to trust the study’s conclusion. “Study links coffee consumption to depression in women”- how did the researchers measure coffee consumption, and how did they measure depression?

Question 45

What factor makes a confidence interval more precise?

Answer 45

Larger sample sizes make a confidence interval narrower and more precise

Question 46

Experiment

Answer 46

Experiments can support causal claims- in an experiment, one variable is manipulated and the other is measured. Uses random assignment

Question 47

Random assignment

Answer 47

Using a random method, like rolling a die, to assign participants to each group- prevents other factors (different levels of maturity, etc.) from influencing the results.

Question 48

How does a study’s method establish temporal precedence and internal validity?

Answer 48

Manipulating the independent (causal) variable ensures it comes first, since the other variable is measuring the manipulated activity- this establishes temporal precedence. When researchers manipulate a variable, they have the potential to control for alternative explanations- they can ensure internal validity.