The first step to starting a sociological research involves asking and defining a research question. You are interested in understanding a social phenomenon and you want to find the sociological explanation for your question. Examples of sociological research questions are: Does the US have a better health care system than Canada? Does having tighter gun control decrease violence? Is income inequality increasing or decreasing as a result of globalization? Do women have different career opportunities compared to men? Is there an educational achievement gap between racial groups? Have divorce rates increased or decreased in the past decade, and why? How will longer life expectancy change the lifestyles of retired citizens and their families?

Once you have a research question, you need to be able to define your research question and recognize the variables involved. Taking the question “Does the US have a better health care system then Canada?” example, the variables involved are the two health care systems and the health of the populations. Your unit of analysis is countries—Canada and the US. The next step is to figure out how to measure your variables. Measuring in social science research entails converting abstract concepts into a concrete and measurable concept (76). To conceptualize whether or not a health care system is “better” involves defining what a good health care system results to. This process of operationalization involves figuring out a way to measure your abstract concept. In this case, “better” can be measured by infant mortality rates, the percent of people having access to affordable health care, average life expectancy, and people's level of satisfaction with the system, and doctor-patient ratio. These are indicators of the quality of health of a population. By looking at this data, one can better asses the general level of the population's health. In conceptualizing the health care system, an important aspect to define is the level of government involvement, the role of private corporations and pharmaceutical companies. These points can then be measured by looking at each country's budget allocated to health care, laws and the constitution regarding health care policies

Second step in the process is to review previous research or a literature review. This involves reading what other scholars have written about your topic and their findings. In conducting a literature review one has to take note of the legitimacy and accuracy of the sources. For example, if one is reviewing the literature on US health care and came across an article written by the CEO of a pharmaceutical company, one has to take into account the writer's position and possible personal biases. If your research question is “Does having tighter gun control decrease violence?” reviewing a literature by a prominent supporter of the NRA (National Rifle Association) is going to have certain biases, which the researcher has to take into account.

Third step is the formulate a testable hypothesis. A hypothesis is a statement of the predicted relationship between variables. A variable “is a measurable condition or characteristics that is subject to change under different conditions (Schaefer 30, 2007).” In our research question, “Does having tighter gun control decrease violence?” you are dealing with two variables, gun control and violence. To operationalize violence we can look at crime rates in the country per state. An example of a hypothesis is: having tighter gun controls decrease the amount of violence. You are hypothesizing on a relationship between the two variables. Formulating a hypothesis also involves defining your dependent variable or response variable and independent variable or explanatory variable. In this case, you are testing if the level of violence is dependent on gun control. So, violence is the dependent variable and gun control the independent variable. You are hypothesizing that level of gun control explains violence.

Fourth step in the research process is to design the research and collect the data. In designing the research, you have to decide whether you are going to conduct a qualitative or a quantitative research. Quantitative research deals with more statistical methods and focus is on data and variables that can be reduced to numbers. For example, variables such as income, GDP, number of children, and crime rates can be used for a quantitative analysis. On the other hand, qualitative research focuses on “interpretive descriptions” (Kendall 23, 2006) and data that cannot be easily converted to numbers. In our sample research comparing the health care systems of Canada and the US. Looking at laws and the constitution in deciphering the type of health care system a country has involves a quantitative analysis. Contrastingly, looking at average mortality rates and life expectancy to operationalized the population's health is quantitative. Thus, when choosing a research design, one does not have to be restricted to either a quantitative or qualitative analysis. A combination of the two is possible. The type of design you choose is dictated by the variables you choose and the best way to operationalize them.

The method of collecting data is also influenced by your variables and your research question. This is the art of research design. There are no set ways to designing a research rather, the sociologist needs to use the tools available that best fit their questions and hypothesis. In collecting data, one needs to be aware of the validity and reliability of the data. Validity refers to the degree to which the indicator is accurately measuring a concept. In our research question, “Does having tighter gun control decrease violence?” gun control can be operationalized or measured by looking at gun laws. But what if even though laws are in place, they are not enforced? Are gun laws then still a good measure of gun control? In our health care example, what if we measured the quality of health care based on people's opinion on their doctors only? Is having a good or bad opinion indicative of the quality of health of the population? What if people have high opinions of their doctors, but there are only a few people in the country that have access to health care? Is this a valid measurement in this case? Reliability refers to the extent to which the measurement methods yield consistent results when applied over time and to different subjects. For example, a ruler is a reliable tool for measurement since if you measure the length of the same object over and over again, granted that the object being measure was not manipulated in any way, you will yield consistent results. For example, in our research question on gun control and violence, we operationalized violence by looking at crime rates. However, what if some states only included violent crimes in their data, while others included all types of crimes. Thus, your method of data collection on crime rate is not reliable, since each state has a different way of measuring crime.

Another aspect of data collection is the method of selecting the sample. A sample is a subset of the whole population that is statistically representative of the whole population. If conducting research on gun ownership in the US, it would be expensive and inefficient to ask each person in the US if they had a gun, thus a sample is used. However, you want your sample to mirror your population. For example, in 2016, the US has a population of 300 million and 50.8% are females. If you want a sample that is representative of the US, you would want to make sure that at least half of your sample is female. Gender is just one component to take into account, but a representative sample is crucial to have valid results.

With the internet, downloading datasets has become easier. The US Census Bureau, the General Social Survey (GSS), Demographic and Health Surveys (DHS), and the United Nations (UN) are just a few institutions that have a wealth of data that is free to the public. Data is usually compatible with the following statistical packages: R, Stata, SAS, and SPSS. Numbers are data are an integral part of statistical research. Sociology and statistics have become inseparable and is a requirement for most undergraduate programs in sociology.

This is an example of descriptive statistics. In this graph, we are shown how the US econmy is divided by sector over time. But we have no idea WHY this is or what is causing this trend in the first place.

If we want to go further in our analysis, we need to start asking the WHY questions and establish correlation and/or causality.

Another example of understanding social behavior through sociology is how we view education in contemporary times. It is only in the latter half of the 20th century where going to a 4-year college has become the norm. In hunting and gathering societies, schools—as formal institutions did not exist. In agricultural societies, people learned their basic skills at home or would do apprenticeships to learn their trade. If you wanted to learn how to be a cobbler, a watchmaker, or a dressmaker, you learned by having a mentor. Moreover, what we now consider as “formal education” through institutions was reserved for the rich, the privileged, and males.

How old are you?

• 18 and under
• 19 to 24
• 25 to 35
• 36 to 56
• 56 and above

What do you think about the new Avengers movie?

Data gathered through observation usually involves qualitative research. Researchers opting observation as a method of data collection are involved in field work and need to physically be with the groups of people they want to study. Researcher can either choose to be an active participant of the group or a mere observer. Participant observation would require the researcher to interact and engage with the subjects, whereas an observer does not. For example, in researching her book, Nickel and Dimed: On (Not) Getting By in America, Barbara Ehrenreich investigates the impact of the 1996 welfare reform on the working poor in the US. She worked with the people she was analyzing and “lived” their lives to better understand them.

Another example is the 2004 documentary by Morgan Spurlock, Supersize Me. Wanting to understand the psychological and physical effects of eating fast food, he embarks on a 30-day journey eating only McDonald's food. There are risks to being a participant researcher. Spurlock, who started off the study as a fairly healthy 32-year old gained 24 ½ lbs, suffered from mood swings, sexual dysfunction, and liver damage. An example of observation is the 2006 documentary by Eric Steel, The Bridge. The film captures people committing suicide by jumping off the Golden Gate Bridge. Although he combines his findings with interviews of the family members who committed suicide, Steel did not try to prevent the people from taking their lives.

Observations can either by reactive or unobtrusive. An example of unobtrusive observation is the one conducted by Humphreys on impersonal sex discussed in Chapter 3, where the subject were not aware of the study being conducted. In contrast, reactive studies occur when the subjects are aware of the presence of the researcher. As the term suggests, these studies may generate a reaction from the subjects that is motivated by the presence of the researcher. Referred to as the Hawthrone effect, there is a temporary change in behavior when people are aware of being observed. The term refers to Hawthrone Works, a factory where researchers were conducting a study on how to improve productivity from 1924 to 1932. As the researchers tried different strategies to increase productivity such as manipulating the lighting, salaries, when to provide breaks, and changing work timings, they realized that productivity increased regardless of what they did. Generally, the subject will react in a way that they think will be agreeable to society and to the researcher when they know they are being observed.

In an experiment, sociologists place subjects in a controlled environment. The setting is manipulated and reactions and behavior are recorded and analyzed. Researchers assign subjects to either the experimental group or control group. For example, if one is testing the effect of a new drug, the experimental group would receive the “treatment” or the drug, and the control group would receive a placebo.

An example of a sociological experiment is one conducted by Rosenthal and Jacobson involving student achievement and teacher expectation. The purpose of the experiment was to test the hypothesis on whether ones' achievement is influenced by the society's expectations. In the experiment, the researchers gave the teachers a list of students who did well on an administered test they gave the students the beginning of the school year. The teachers were told that the The Harvard Test of Inflected Acquisition test given to their students is a measure of IQ, and a student's ability to learn. What the teachers did not know is that Rosenthal and Jacobson randomly picked the names on the list, thus, the list did not exactly reflect the test results. At the end of the year, a test was again administered to the students and on average, the students in the list saw an increase of 12 points in their scores, while the rest of the class had a lower average of increase. Other tests that were graded by the teachers and considered subjective tests such as writing and reading showed similar results. Rosenthal and Jacobson theorized that the Pygmalion effect was responsible for the unequal increase in scores. With higher expectation, teachers give more attention to students they believe are smarter (Rosenthal and Jacobson, 1968).

Another famous sociological experiment is the Stanford Prison Experiment headed by Philip Zimbardo at Stanford University in 1961:Staford Prison Experiment.

In all these types of data collection, there are ethical issues to take into account. Whether or not you are invading someone's privacy or endangering their safety, social scientists have a responsibility to their subjects. The American Sociological Association has a code of ethics that sociologists follow when conducing research. What matters is that the research question guides the type of research and data collection to use and that sociologists understand the pros and cons of their methods of choice.

• Research Question
• Variables
• Literature reviewing
• Hypothesis
• Dependent Variable
• Independent variable
• Data Collection
• Qualitative Research
• Quantitative Research
• Validity
• Reliability
• Descriptive Data
• Causality
• Correlation
• Confounding Variables
• Surveys
• Observation
• Reactive
• Unobstrusive
• Experiment
• Hawthrone Effect
• Pygmalion Effect