It is important when we conduct a study that the results mean something to other people even if they weren’t involved in the study. As such, we look for validity, reliability, and generalizability to help us determine if the results of the study are applicable to the larger social world.
There are different ways to study social phenomena. If you wanted to study poverty, for instance, you could do a quantitative analysis by picking a neighborhood, getting the census data, and seeing how much money the average household makes. Then you could compare that to the federal poverty line to determine how many people are in poverty. On the other hand, you might not get a complete picture just by looking at the numbers. Some families have high incomes (maybe $100,000/year or more), but if you asked them, they might tell you that they don’t have enough money to get by. If you just look at the numbers, you might exclude these people from your study, yet you might be able to learn something interesting about social life by talking to these people. As a result, many studies include both quantitative and qualitative methods in order to produce more thorough data.
For example, you read somewhere that college graduates are likely to have higher incomes than non-college graduates, so you hypothesize that graduation from college increases salary. You collect some data and analyze it to determine whether your theory is correct.
In this case, you notice that one of your friends is making more money than one of your other friends, even though they have similar jobs. You have no idea why this could be, but you are interested in figuring it out. You think of all the differences between these two people. They are both females, they are from the same state, they like the same music, they work in the same area – but then you remember that one went to college and the other did not. You look back at their work history to see if there was always a big difference in the amount of income they made. You then see that they were making the same salaries while in high school, but after the first friend graduated from college, she got a huge raise. You can conclude that there is a correlation (or a connection) between college graduation and salary!
For example, in Chapter 1, you read about famous college dropouts like Woody Allen and Bill Gates. We might ask ourselves if people become successful because they go to college or if they would have been successful whether or not they went to college. In order to study success, we would want to determine if college caused them to be successful or if college was simply a coincidence and did not cause success. The cases that we mentioned (Allen and Gates) lead us to believe that successful people might be successful regardless of whether they finish college, but we would have to conduct a more thorough study to make a determination. We could do a quantitative study (maybe by looking at SAT scores before college, and salaries later in life), or a qualitative study (possibly by talking to individuals who are successful to find out how college did or didn’t influence their success). Image: http://commons.wikimedia.org/wiki/File:Domino_Cascade.JPG
Causation is a stronger assertion than correlation. Let’s say you have noticed that people who have fender-benders (small car accidents) on their way to work are in a bad mood. But you wonder, did the car accident put them in a bad mood, or were they already in a bad mood, which caused them to have an accident? You see that there is a correlation –- bad moods and car accidents are related to each other (there is an association), but which causes the other? You would want to know which came first, the car accident or the bad mood. If you find that many people who have car accidents were actually already in a bad mood, you might prove causation: that being in a bad mood actually causes car accidents!
In our previous example, car accidents would be the dependent variable. We’re trying to explain whether mood changes the outcome (a car accident or no car accident). Therefore, mood is an independent variable. We want to see if mood has a causal impact on the dependent variable.
A null hypothesis states that there is no relationship between the variables. If we are studying the impact of mood on car accidents, the null hypothesis is that mood does not affect car accidents (there is no effect of mood). The alternative hypothesis is that, as we thought, mood does affect car accidents.
Examples of research methods h ighschool movies - preschool in three cultures - hans rosling
Examples ofResearch Methods
What Makes “Good” Research? Good research should be valid, reliable, and generalizable: Validity: does the study measure what it is intended to measure? Reliability: if you conduct the study again, will you get the same results? Generalizability: will the findings of this study apply to some other population or group of people?2
Research Methods Research methods are standard rules that social scientists follow when trying to establish a causal relationship between social elements. Quantitative methods seek to Qualitative methods attempt to collect obtain information about the information about the social social world that is in, or can be world that cannot be readily converted to, numeric form. converted to numeric form.
Approaches to Research A deductive approach to research: 1)starts with a theory. 2)develops a hypothesis. 3)makes empirical observations. 4)analyzes the data collected through observation to confirm, reject, or modify the original theory.4
Deductive reasoning is more narrow and is generally used to test or confirm hypotheses.
Approaches to Research An inductive approach to research: 1) starts with empirical observation. 2) then works to form a theory. 3) determines if a correlation exists by noticing if a change is observed in two things simultaneously.6
Inductive reasoning is more open-ended and exploratory, especially during the early stages. Sometimes called a “bottom up” approach.
In research it is often a combination.. He noticed that Protestant countries consistently had higher suicide rates than Catholic ones.His theoretical interpretations in turn led His initial observations led him to inductivelyhim to deductively create more create a theory of religion, socialhypotheses and collect more integration, anomie, and suicide.observations.
Causality vs. Correlation Causality is the idea that a change in one factor results in a corresponding change in another factor.9
Spurious CorrelationDoes it meet all three criteria?
Causality vs. Correlation Sociologists conduct research to try to prove causation. To prove causation, correlation and time order are established and alternative explanations are ruled out. 1.Correlation 2. Time order 3. Alternate Explanations.12
Variables – What Are We Studying? A dependent variable is the outcome that a researcher is trying to explain. An independent variable is a measured factor that the researcher believes has a causal impact on the dependent variable. Example: a person’s income (dependent variable) may vary according to age, gender and social class (independent variables).13
The Hypothesis (if…then) A hypothesis is a proposed relationship between two variables, represented by either the null hypothesis or an alternative hypothesis. Null Hypothesis (sometime called no-difference) •Hyperactivity is unrelated to eating sugar. The null hypothesis is good for experimentation because its simple to disprove. If you disprove a null hypothesis, that is evidence for a relationship between the variables you are examining.14
Marijuana and serious mental illness (SMI) research Prevalence of Past Year SMI among Adults Aged 18 or Prevalence of Past Year SMI among Lifetime Marijuana Users Aged 18Older, by Gender and Age Group: 2002 and 2003 or Older, by Age at First Marijuana Use: 2002 and 2003