Skip to Content
Advanced Search

Scientific Reasoning


Borgman, D. (2001). Understanding Scientific Reasoning.  S. Hamilton, MA: Center for Youth Studies.



(Download Scientific Reasoning overview as a PDF)

In the interest of quality in youth ministry, we propose informing our thinking, strategy, and methodology with principles from the social sciences. Such an intent is not meant to divert us from our primary focus of caring about kids for Christ’s sake. Nor is it meant to discourage us with our lack of scientific expertise. It intends rather to encourage us in a positive professional manner. The beginning of scientific thinking is the research question that can be applied to knowing kids better or doing our work more effectively.



  • Research Question: This starting point in research is called by some researchers a proposition or hypothesis. A research question (e.g., What did the availability of guns have to do with school shootings in the 1990s?) can be converted into an initial statement and refined by restatements.
    • "U.S. gun homicides are higher than those in other developed nation."
    • "I am interested in what (when, why, how)..."
    • "Young people believe or agree that…"
    • "Ten percentage of young people…"
    • "Factors such as…affect…"
  • Objectives of the research questions:
    • To describe (what?)
    • To explain (why?)
    • To influence (how?)
  • Concepts (or Terms): Research Questions contain concepts like "bullying," "drug abuse," "pregnant" or "leisure time." Concepts are shorthand expressions of key phenomenon. These terms are generally variables that may include values. Research seeks to define and then measure variables. We must consider the level of abstraction of a concept. Its usefulness diminishes as it becomes either too broad or too narrow. A concept may be too complex or too simple, unmanageable at one end or worthless at the other.
  • Values. Values express a range of possibilities. The concept "marital status" includes the values "married," "separated," "divorced." The concept "student" might break down into upper-class (juniors and seniors), lower-class (freshmen and sophomores), or different types of students. Values express the range of possibilities within variables.

  • Variables. Variables are concepts that serve as the basic building blocks of a research question. They are measurable concepts with a range of possible values. Research questions are clarified by differentiating, choosing, and considering causes of variables.




George Gray and Neil Guppy (1999:35) suggest 10 steps in the development of a research question:

  • Stating the research question.
  • Identifying the variables in the research question.
  • Defining the variables.
  • Specifying the independent and dependent variables.
  • Thinking causally.
  • Embedding a rationale in causal thinking.
  • Relating the research question to the literature.
  • Introducing other variables (antecedent/intervening).
  • Revising causal connections.

  • Drawing a causal model.


Successful surveys build on two kinds of questions—the research question and the survey questions. The research question is the puzzle, problem, or issue that the research addresses (what do we want to know, e.g. what kind of high school student would likely attend a youth group, club, or camp—or how do various types of kids define "cool"). The survey questions are the questions asked of respondents; the exact questions asked depend on the research question.



Defining the research question is both the starting point of the research process and the backbone of the survey, because it directs the researcher’s attention to exactly what survey questions need asking. Formulating intelligent survey questions requires considerable time and effort developing the research question—and the help of a trained consultant.


You may be interested in the influence of derogatory rap music. This might lead to the following questions that need a consultant’s critique and expertise. (What types of kids listen to Eminem? What influence does Eminem and his music have on these various types of kids? Do more boys (blacks) than girls (whites) listen to Eminem?)


Your research question might lead to the following survey questions:

  • What do you like and/or dislike about Eminem and his music? (and for those who don’t like Eminem or are neutral about him)
  • Why do you think Eminem is so popular?
  • Why do you think friends or classmates of yours listen to him?
  • Why do you object to or not listen to Eminem?
  • What possible influence do you think his music can have on those who like his music? (and for those who like Eminem)
  • When do you tend to listen to Eminem or similar rap music?
  • What do you most like about Eminem’s style and music?
  • What do you say to those who object to Eminem and his lyrics?
  • What kind of influence do you think Eminem has on his fans?


Such open-ended questions are enjoyed by most teen respondents, but they are difficult to tabulate and interpret. Analysis of questions based on a Likert Scale is much easier and measurable. [Statement] "I enjoy listening to Emimem," or "I think Eminem’s lyrics contribute to male disrespect of women" [Check Response] Strongly Disagree, Disagree, No Opinion, Agree, Strongly Agree.



To be effective in thinking causally, you need to develop a rationale explaining how it is that one factor might be related to a second. The debate over smoking nicely illustrates this issue. For a long time, cigarette manufacturers were able to win costly lawsuits because although smokers were more likely to develop lung cancer than nonsmokers, the process by which smoking caused cancer was not understood. There was a correlation between smoking and cancer, but that correlation did not prove causation. To be able to say that smoking caused cancer, we had to know how the influence worked. Medical researchers have now made this link by showing the pathway by which the inhalation of smoke causes cells in the lungs to mutate and become cancerous. Similarly, we can be comfortable saying that aging causes a reduction in physical activity levels because we know that as we age our flexibility lessens, our lung capacity erodes, and our joints become more arthritic.

[Obviously, proving that violent electronic images produce aggressive conduct, that Eminem’s lyrics encourage people to molest women or bash gays, or that pornographic material leads to sexual promiscuity and perversions are much more complex matters.]


Now here is a tricky issue. How do we know that our rationale (in research) is correct? We don’t. That is why it is a research question (as opposed to a research answer). A research question is a hypothesis; it is a hunch about how the world works. We begin by postulating how the variables in the research question are related and go on to conduct a study to assess whether our claim is true. We gather evidence to test our research question. This is tricky only in the sense of understanding how the research process works. We think our rationale makes sense, but the research we undertake is designed to test that claim. (Gray & Guppy, 1999:43-44)


Drawing a Causal Model. Good social science problems are never as simple as X causes Y. The social world is too complex for that. On the other hand, care must also be taken to avoid overcomplicating the research problem or defining problems that border on speculation or unsubstantiated theorizing—that is the other extreme. Generally speaking, however, good theories move beyond the realm of simply stating X and Y.




It is helpful to carry a causal model with us applying it to what we say about youthful behavior and youth ministry itself. We try to remain aware of multi-causality in complex human behavior and social realities as well as of spiritual factors. The spiritual is often missed by social scientists when they negate the supernatural or consider spiritual factors too difficult or impossible to measure.


Take a look at two suggested (and unscientific) causal diagrams (adapted from Gray & Guppy, 1999: 47) in regards to an eating disorder and school shooting. Curved lines between Antecedent Variable 1 and 2 suggest causal flow. Straight lines between other variable suggest correlation. The single-headed arrow pointing from Independent Variable to the Dependent Variable indicates the causal flow of your research question. The single-headed arrows pointing from Antecedent and Intervening Variables to the Dependent Variable represent other causal influences.


Example One

Antecedent Variable 1

Intervening Variable 1

Feeling powerless, lack of control home and out

Advertising’s emphasis on thin female bodies

Independent Variable

Dependent Variable

Feeling unnoticed and un-nurtured

Eating disorder

Antecedent Variable 2

Intervening Variable 2

Low self-esteem


Peer pressure: popular girls, guys, friends dieting


Example Two

Antecedent Variable 1

Intervening Variable 1

Family neglect or abuse

Bullying from cohorts

Independent Variable

Dependent Variable

Unmitigated rage

School shooting

Antecedent Variable 2

Intervening Variable 2

Psychological predisposition

Habitual consumption violent video games/music


Intervening Variable 3



Encouragement from friend(s)


Two things come to mind regarding these examples. They are relevant illustrations of what we are dealing with in youth culture today. Secondly, they illustrate the complexity (going even beyond these diagrams) of youthful behavior. These two examples are so complex they go beyond a possible research question. As they stand, they represent research problems rather than single research questions. Still, they express a way of thinking which can make us more professional and effective in our ministries.


The hypothesis or proposition in Example One might be expressed as follows:


When a girl from a home in which she feels powerless and lacking in control develops a low self-image in a society which emphasizes thinness as essential to beauty and popularity and is surrounded by guys who favor popular girls and friends who are continually experimenting with dieting and unhealthy forms of weight control, her feeling un-noticed and un-nurtured may lead her into dangerous control of her eating and weight.


Similarly we could state Example Two in this way:


When domestic neglect or abuse combines with a particular genetic disposition in a boy, he may develop a conduct disorder making him particularly susceptible to teasing and bullying which can drive him to take out his feelings in violent video games and music until his anger becomes too much to stand and, with encouragement from anti-social friends, he strikes out in a violent, irrational and random manner at his tormentors.


The challenge of social science is to take such undefined terms and rambling (unwieldy and complex) statements—and create from them several focused and measurable research questions.


The point here is to encourage general causal thinking and to recognize that our thinking needs scientific discipline. Recognizing scientific process helps us become more discerning in processing the news and media—as well as in the doing of our ministry.


Using a Causal Model in Ministry


Why does a young person come to a youth group, club, or camp? Can you see in these questions the interaction of variables—home life, personal disposition, friends, what is considered cool, relationship with a leader, spiritual interest? Could you place these variables in a diagram such as those above?


We might also consider the interaction of human/cultural and divine influences:

  • The work of the Holy Spirit.
  • A young person’s interest in goodness and quest for truth which comes from the image of God in everyone.
  • The resistance of everyone’s sinful nature.
  • The additional interference of the Evil One.


I believe it is helpful to consider why a person accepts Christ at camp, or loses a personal commitment in college life, in terms of spiritual dynamics and human causal factors. Putting the human and cultural together with spiritual factors makes for mature and spiritual understanding. We may not reach scientific certainty, but we are being "wise as serpents…" (Matthew 10:16) and learning from "the children of this age" (Luke 16:8b). Knowledge and understanding are valued by Scripture and by Jesus Christ.




An internationally famous American evangelist was quoted just after the terrible Columbine shootings, "Guns don’t kill; people do." I was shocked and saddened. That is as bad as saying, "Guns are the cause of school shootings." Simplistic X à Y thinking and arguments will undermine our ministries.


If we want to use research to test causal ideas, we must choose our sample and construct our survey questions very carefully.


We must avoid spuriousness and misspecification. In X à Y equations, X may seem to cause Y because they are found together or happen at the same time. The real cause of Y may, however, be Z, an overlooked variable. That is called spuriousness. Or Z may actually be a cause of both X and Y, and this is termed misspecification.


If our aim is a scientific report, the respondents of our survey should be representative and therefore randomly selected from the whole group we want to measure.


Scientists also caution us about scope conditions which express limitations of a causal relationship. Water boils at 214 degrees Fahrenheit…at sea level, and certain things may be true of young people in suburbia. It is important for us to understand why any causal claim is limited.


The validity of a survey depends on whether it accurately measures what it intends to describe or explain. Some final exams are not valid measurements of what the syllabus said were the goals of the course. Such exam scores are not in themselves valid measures for a final grade. Or, if a survey’s results vary in its administration, if there is random measurement error, we say that it lacks reliability. Questions of a survey ought to be asked in a similar, objective way every time the survey is administered.




Scientific research can be either quantitative or qualitative research. Simply stated, quantitative research measures amounts and percentages. (Note biblical references to census taking. The words state and statistics come from a common root. God asks Moses to number the people of Israel in Numbers 1:2-3. David’s and Caesar Augustus’ censuses are further examples.) Quantitative research determines the amount or relative amounts of components in a mixture or group asking, "how many" or "what percentage." Qualitative research, on the other hand, generally takes place by interview and tries to get behind the "what" with further "what, when, why, how" questions. While both these types of research can be scientific, they may, when proper control of questions, sampling, and interpretation are lacking, become unscientific.


Our surveys and interviews in youth ministry are usually unscientific as compared to those of trained social scientists working for government, universities, or advertising agencies. But that does not mean our surveys of, and interviews with, young people not helpful. This is especially true when we are careful about our goals and methods, when we allow them to be informed by scientific principles.




Much more needs to be considered in measuring the context of our work, the subjects of our ministry, and the effectiveness of our strategy and methods. We have only introduced the issue of measurements. We have not adequately covered the designing of questions and organization of the questionnaire along with choosing the type of survey and selecting a sample of our target population. Nor have we touched on the processing of data and how our results are to be reported and disseminated so others can profit. These pages are meant to be an introduction to a style of thinking about young people and our ministry. They assume that being attractive to young people as Jesus was, that being dependent on and filled with the Holy Spirit, and that being professional are not opposed but complementary.

Dean Borgman cCYS