{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

CJ 300 Research Design I (Winter 2011)

CJ 300 Research Design I (Winter 2011) - Quantitative...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Quantitative Research Design I CJ 300 Dr. Kierkus January 2010 Introduction Quantitative research is about testing theories using quantifiable, empirical data. In other words, a deductive research process. The key to this is effectively translating complex concepts into numbers without losing the meaning of the underlying concepts. If we oversimplify the concepts while quantifying them we have fallen victim to reductionism. Consider the following example: Theory: good parenting protects kids from delinquency. Hypothesis: "The higher the quality of parenting the lower the level of delinquency." You decide to measure the quality of parenting by asking: "Please rate how good your parents are on a scale of 1-10." Two kids answer this question by writing down the number "7". Consider what these numbers mean? Quantitative Research Process 1. 2. 3. Develop a research question. Specify a theory that is relevant to this question (usually after a review of literature). Create a testable hypothesis based on the theory. Define the variables in the hypothesis and specify the relationships between them. 4. 5. 6. 7. Determine which research method is most appropriate for testing the hypothesis. Operationalize the variables. Collect the data. Analyze the data using statistical procedures. Support or refute your hypothesis. 8. 9. 10. Interpret the findings (acknowledging any weaknesses of your design). Report your findings to the scientific community. Replicate your results. Variables Variables are concepts that can take on a variety of different values. They can have as few as two levels, or an infinite number (etc. "gender" vs. "mass"). Four basic types of variables: Independent, dependent, intervening and confounding (control). A hypothesis must have at least an independent and a dependent variable. Complex hypotheses also have intervening and confounding variables. Types of Variables Independent Variable: The cause. An increase in the level of illegal gun ownership (independent) leads to an increase in the level of gun crime. An increase in the level of illegal gun ownership (independent) leads to an increase in the level of gun crime (dependent). Children who watch a lot of violent television (independent) become increasingly impulsive (intervening) and therefore, are more likely to engage in violent delinquency (dependent). Dependent Variable: The effect. A change in the independent variable causes a change in this variable. Intervening Variable: Explains why the independent variable influences the dependent variable. Confounding (Control) Variable: Causes changes in both the independent and the dependent variable, which makes it seem as if the independent variable causes the dependent variable to change, when really it does not. This is called spuriousness. Developing and Testing Hypotheses The logic of the scientific process allows us to: Refute (or falsify) a hypothesis (show that it isn't true). Support a hypothesis (show that it might be true). However, we can't ever conclusively prove that a hypothesis is true. This is because it's always possible that evidence which seems to support a hypothesis is attributable to some other cause (besides the independent variable). However, negative evidence refuting a hypothesis conclusively demonstrates that it's incorrect. "The Simpsons Bear Patrol". No bears on the streets of Springfield does not conclusively prove that the patrol is working. But if a bear eats Lisa, that's conclusive proof that the patrol isn't working! Null Hypothesis Testing Statistically, when we want to test a research hypotheses, we actually test its opposite. This is called the process of null hypothesis testing. For instance: Null hypothesis: "An increase in the level of prison crowding has no effect on the number of prisoner riots". One Tailed Research hypothesis: "An increase in the level of prison crowding leads to an increase in the number of prisoner riots". Two Tailed Research hypothesis: "An increase in the level of prison crowding leads to a change in the number of prison riots." This is because of the logic outlined earlier: we can't ever conclusively determine if the research hypothesis is true, but we can establish that the null hypothesis is false. Double Barreled Hypotheses Although a complex hypothesis can contain multiple variables, it should never be double barreled. That is, we must test the effects of each variable on each other variable individually. For example, this is a double barreled hypothesis: "People who spend more hours studying in smaller groups tend to do better on exams than those who spend fewer hours studying in larger groups". Should be broken down into two hypotheses: "The greater the number of hours students study, the higher their scores on exams." "The smaller the groups that students study in, the higher their scores on exams." Competing Hypotheses Sometimes researchers create studies that have competing hypotheses. In this case, refuting one hypothesis suggests that the other must be true. For instance: "Lowering the drinking age will decrease the number of DUI collisions" vs. "Lowering the drinking age will increase the number of DUI collisions". Both of these are theoretically plausible (the "forbidden fruit" argument vs. the "more alcohol = more danger" argument). Both can't be true. Evidence supporting one is evidence refuting the other. Hypotheses and Causal Diagrams The more complex hypotheses become, the more difficult it becomes to express them in words. "A picture is worth a thousand words". Researchers illustrate complex hypotheses using pictures called causal diagrams. Causal diagrams show how all of the variables are related. The "boxes" in causal diagrams represent variables. Independent variables go on the left, dependent variables on the right, intervening variables in between. Confounding variables are usually placed below the main causal sequence. Arrows with + and signs indicate relationships and the directions of those relationships. Hypotheses and Causal Diagrams The following is an example of hypothesis dealing with violent television viewing and its effect on delinquency, illustrated using a causal diagram. # of violent television shows watched (Independent) + Low level of Social Control (Confounding) + Impulsiveness (Intervening) + + Violent delinquency (Dependent) Summary Quantitative research is a step-by-step process. It involves the testing of hypotheses using empirical data. Hypotheses are made up of 4 different kinds of variables, and the connections between them. Causal diagrams can help us understand complex hypotheses. ...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online