BACKGROUND

Chapter 5 discusses the very important issue of causation in social sciences research. A correlation is a relationship between two variables. Height and weight are variables that are correlated. All correlations are either positive or negative. In an adult population, height and weight will be positively correlated. That means as height increases, weight increases. A graph of such a relationship is shown below. Each red dot represents a particular person’s height and weight (called a data point). For example, the lone dot at the top of the chart (closest to the chart title) represents a person who is just over 72 inches and about 215 lbs.

Examining the scatterplot you can see that in general as height increases weight increases. Thus the correlation is positive. This does not have to be true in every case for there to be a positive correlation.

One of the phrases you will hear in any research methods class is that correlation does not imply causation. Repeat that to yourself several times. Correlation does not imply causation. Correlation does not imply causation. While two variables MUST be correlated in order for there to be a POSSIBILITY of causation, two other criteria must be met to establish causation.

There are 3 conditions that are required in order to claim a CAUSAL relationship exists between 2 variables:

Condition 1: Correlation: One variable must be associated or related (i.e. correlated). So in fact, correlation is a necessary, but not sufficient condition for causation.

Condition 2: Temporal order: Change in the independent variable must precede (come before) change in the dependent variable.

Condition 3: Lack of plausible alternative explanations: All other reasonable explanations for the relationship between the two variables have been ruled out.

Here’s an example to illustrate the difference between observing a relationship or correlation between two variables and being able to say one caused the other. Studies have demonstrated a correlation between coffee drinking (A) and likelihood of having a heart attack (B). Data also shows that coffee drinking (A) and cigarette smoking (C) are correlated. Finally, cigarette smoking (C) and likelihood of a heart attack (B) are related.

Therefore, it is possible that what appears at first to be a CAUSAL relationship between coffee drinking (A) and heart attacks (B) is really due to the extraneous variable of cigarette smoking (C). In other words, it might not be that drinking coffee (A) cause a heart attack (B) , but that coffee drinkers (A) also smoke cigarettes (C), and cigarette smoking (C) causes heart attacks (B). The researcher would have to control for cigarette smoking in order to exclude this plausible alternative explanation (cigarettes) and meet Condition 3: All alternative explanations for the correlation have been ruled out. Only then could a conclusion be drawn that coffee causes heart attacks.

Purpose

1. Provide an opportunity to apply the three basic conditions for establishing a causal relationship to scenarios.
2. Continue to illustrate how important knowledge of research methods is to every person as well as criminal justice professionals.
3. Continue to familiarize students with scholarly research articles.

Assignment

Type your answers in the spaces provided below. Once you have completed, be sure to save your work and upload the document into Canvas.

Part I

Watch the 10 minute clip at:

http://www.skepticsfieldguide.net/2008/12/podcast-hunting-humbug-101-tutorial-10.html (Links to an external site.)

on Jenny McCarthy ‘s opinions about the causes of autism. Scroll down to find the video. Be forewarned that the language in this video is inappropriate at times.

1. She makes the classic mistake of mistaking correlation for causation. She observes that two events have been reported by thousands of parents to have occurred one after the other and assumes that means one caused the other. In one sentence explain what pattern of events she connects and how she mistakes the correlation between two variables as evidence of causation.

1. In referring back to the lecture and your notes from Chapter 1 and the discussion of the limitations of personal experience, identify 2 kinds of limitations she makes when talking about the causes of autism. Provide additional examples of each limitation.

1. She states “Parent anecdotal information is science-based information.” Define science and explain how parent anecdotal information is actually the opposite of science.

3a. Read page 59 of your textbook: Making Research Real 3.5. What sanction was imposed on Andrew Wakefield and why?

Part II

Suppose we collected data on the number of fire trucks responding to fires and the amount of damage as a result for every fire in Tampa in the last five years. The scatterplot below illustrates how the data would look if there was a positive correlation between the two variables.

1. Describe in your own words what “there is a positive correlation between the number of fire trucks responding to a fire and the dollar amount of damage as a result of that fire” means.

1. What mistaken conclusion might someone draw if they mistook the observed correlation between the number of fire trucks and the dollar amount of damage for causation?

1. What 3rd variable is likely to be correlated with both variables in the study? Think about this. Why would there be both many fire trucks responding to a fire as well as high dollar amounts of damage from that fire?

1. Given the presence of that 3rd variable that is likely to be correlated with both variables in the study, which of the 3 conditions for claiming a causal relationship is not met and why?

Note: The claim of a causal relationship between the number of fire trucks at each fire and the dollar amount of damages at each fire is absurd. However, many spurious correlations do not seem absurd and some seem compelling. This is why it is so important to be skeptical when cause is implied in news stories. The knowledge you are gaining in this course will help you become an informed consumer of research results. This knowledge will also serve you well if you decide to go into the field of criminal justice.

Part III.

We are going to read a second research article to continue the process of becoming familiar with how studies are conducted and reported on in scholarly journals. The article here: Article -Causation.pdf and is called Hook ‘em horns and heaving drinking: Alcohol use and collegiate sports by Dan J. Neal and Kim Fromme in 2007).

Read the title and the key words (under the Abstract) and speculate (in your mind) what you think the research is about.

1. Read the abstract. Every research study will have an abstract that provides a short summary of the study. It will contain the purpose of the work, a brief explanation of how the study was accomplished, and the results. In your own words what is this study about?

1. Read the Introduction (pages 2681-2683) Notice the style of writing. If there is “no crying in baseball” as Tom Hanks said, there is no commentary in research papers. The researchers are making statements that can be supported by prior research rather than stating opinions or playing to the reader’s emotions. Notice how almost every sentence has a citation. The author’s use a style called in-text citations. What do the researchers say about how their research builds or expands on previous research?

1. The researchers specified four hypotheses this research was designed to test. Which of the four hypotheses do you think is the most interesting and why.

1. Think about the hypothesis you selected above. What is the dependent variable (DV)? The DV is the variable that the research is designed to explain.

1. Read the section labeled 2. Method. This is the section of a research article where the study design, sample, data collection method and measurement of variables are each described. Who was eligible to participate in this study?

1. How was data collected from the participants? (Hint: There are two ways, so specify both).

1. How was “football season” defined?

1. Read section 2.2. Measures. How was the concept of “social involvement” DEFINED? Hint: There is a difference between how something is defined and how it is measured. Defining a concept is the first step.

1. When surveys are used the measurement of a variable is the question(s) used to ask about the concept. You defined the concept of social involvement above. Now tell me what survey question was used to measure “social involvement”.

1. We are not going to read the parts of the article that report the results as our focus is on how the study was conducted. Feel free to read the section labeled 3. Results, but your next question is related to page 2691 of the article and section 4.3 Limitations. All research articles will include the researcher’s assessment of the limitations of their study. Describe two of the limitations that the researchers note might impact the interpretations of the result of this study.