I. Ways of Knowing About the World
Scientific research is often viewed by people as some sort of specialized enterprise for the intellectually gifted. A mysterious and incomprehensible array of techniques and skills carried out hidden away, out of sight, behind laboratory doors. But science is really just one way out of several to understand and gain knowledge about the world. There are several which we'll cover: Authority, Logic, Intuition and Science. And science is easily the most accessible and democratic way to do so. Science is based on multiple repeatable objective independent observations while following specific detailed procedures. The procedures, though, are not secret or guarded. Anyone, who is willing, can be taught to perform them. That is why science can be viewed as accessible and democratic. Unlike some other ways of knowing abut the world, in science, if I follow a series of procedures and observe a result then anyone else should be able to follow the same procedures and observe the same or similar results. In an odd way it is just like cooking and following a recipe: if the procedures and materials are faithfully duplicated then the results should be the same or very similar each time. Not everyone is Emeril Lagasse, but if you follow his recipes you can cook just like him. Screaming "BAM!" at the top of your lungs each step of the way while you "kick it up a notch" as you make pork egg rolls with sweet and sour sauce is optional, though.

A. Authority
Simply put, authority is taking somebody's word for it. Whether you ask first or you are lectured, apparently out of the blue, someone tells you something about the nature of the world: what something means, how to behave, what not to do, what will happen if you run across the street without looking or eat less than thirty minutes before swimming. Basically, we rely on authority all the time. It isn't necessarily a bad thing to do so. It can save a lot of time and unpleasantness that we might otherwise go through if we had to learn everything about the world around us by trial and error. However, authority does have its problems. Those we look towards for authoritative statements often disagree among themselves. Parents may say that it's acceptable for their eighteen year old child to drink wine with a meal but the government says that the legal drinking age is twenty-one. And authorities sometimes change their minds. For instance, government's legal drinking age at one time was eighteen. And sometimes authorities change their minds because they can be very wrong. The government, at one time in the distant past, said slavery was legal; not anymore.

B. Logic
Logic is a powerful tool for understanding the world. However, it has limitations. It requires that the bases for logical reasoning be true. Otherwise, the conclusions drawn can be wrong. For example, here's a flawed logical argument: The major I pick this semester will determine what I do for the rest of my life. I declared history as my major. Therefore, I will become an historian. It may be a nice thought. It may even be comforting. But even if someone never changes their major, the degree they obtain may not correspond to their eventual career. Even if the bases for logical reasoning are true, the reasoning can still be flawed. Here's another flawed conclusion: Parrots eat crackers. Parrots live in trees. Therefore, crackers grow on trees. Logic has limitations. See what I mean?

C. Intuition
Just about everyone has gut feelings, hunches. Often they're correct. And often they're wrong. If we're asked for the basis of those hunches we're typically at a loss to explain them. Sometimes, we're able to give some sort of after-the-fact explanations but usually those intuitive judgments seemingly come out of nowhere. There are two forms of intuition that we'll discuss, common sense and mysticism.

1. Common Sense
Common sense is something we often hold in high regard. We've all known or heard of people that might be highly educated or intellectual but lacked the sort of practical everyday intelligence that is called common sense. And common sense can be a very accurate description of how the world works in a sort of after-the-fact sort of a way. We've all heard common sense sayings like "you can lead a horse to water but you can't make him drink" which can be true in both a literal as well as a figurative sense.

But common sense has its limitations. For instance, it changes in different times and places. In the 1600's in Salem, Massachusetts numerous accusations were made of witchcraft and several people were tried and executed as witches. Much of what happened can be understood through what was considered "common sense" at that place and time. First, it was common sense that witches existed. It was also common sense that, since water was associated with purity and holiness, one could determine if someone was a witch by submerging them under water for an extended period of time. Therefore, if the accused were unholy witches, then the water would not enter their bodies and they wouldn't drown and they could be hanged or burned at the stake. If the person drowned they were clearly innocent and though they were dead, at least it was certain that they were in Heaven with God.

Common sense, relying on "after-the-fact" labeling of the outcome of an event, also is unable to predict what will happen in some situations. Take, for example, a situation in which two people who are dating and one moves away to a different city. What will happen to their relationship? Common sense says "absence makes the heart grow fonder." We can find many examples to verify that little bit of common sense. But there are also many examples that we can find to support the common sense saying, "out of sight out, out of mind." So, which one will it be in any specific instance? To rely on common sense, we have to wait for the situation to resolve itself before we know how to categorize that specific case.

2. Mysticism
Mystical experiences rely on an altered state of consciousness, whether that state is attained by prayer, meditation, ingesting a substance, or perhaps through some sort of "gift." Mysticism often incorporates authority as well, asking us to rely on people with special skills, talents or gifts like psychics or mediums or shamans for guidance. This is not to dismiss the personal effects and significance that mystical experiences may have on some people, which can be beneficial and profound in some cases. However, mysticism has limitations. Communicating the meaning of a mystical experience to another person may be difficult or impossible because of the deeply unique and subjective nature of the experience. It may be not even be possible to find the words to fully and sensibly comprehend the experience ourselves. Also, because consciousness may be altered, perhaps due to drugs, the experience or its meaning may be entirely wrong.

D. Science
Science relies on objective and repeatable observations. There are no special skills, techniques or procedures that are not available to all, though some may require extensive time and practice to use effectively. But anyone can do science. If observations cannot be duplicated by others, they are not accepted. Therefore, only those observations or results that are consistent and repeatable and can be agreed upon by more than one person, even those (or, especially those) who may have been skeptical, can be called scientific. This openness, objectiveness and accessibility defines science. However, this also highlights a limitation of science. It is not a good tool for understanding the subjective, the rare and the unique. Events or occurrences that can only be observed by one point of view or are "one-in-a-million" cannot easily meet the requirement for multiple objective observations. And science also has some implicit assumptions which underlie it as well.

II. Working Assumptions of Science

A. Reality
One implicit assumption of science is that the world around us has some sort of underlying reality apart from our observation of it. To borrow and mangle an old expression, if a tree falls in the woods and no one there to hear it, it still makes a sound. The existence of the universe and ourselves is not a cosmic illusion, nor is it some sort of dream we all share like something out of the movie, "The Matrix." The world doesn't disappear when we fall asleep and magically come back into existence when we wake up. Reality is real.

B. Rationality
Science also assumes that the world is rational, that is, it can be understood through logical reasoning. Even though our reasoning can falter from time to time, the universe is rationally assembled and is potentially comprehensible, even if we cannot understand it presently. Through persistent effort and progress over time, it can be eventually be understood.

C. Regularity
Regularity holds that the rules and laws and properties of the universe are (as well as have been and will continue to be) the same at every point in the universe. The universe, and everything in it, is stable and consistent in that regard.

D. Causality
Another assumption is causality, which means in plain English that things don't just happen. Nothing happens spontaneously without some sort of cause, even though that cause may not be obvious at first. All things or events occur because a previous thing or event caused them.

E. Discoverability
Lastly, science assumes that the universe hold no secrets that cannot be ultimately known and understood. The difference here is that between a puzzle and a mystery. A mystery may not be solvable, but a puzzle, even a difficult one, can eventually be solved and understood.

III. Research Strategies
Scientific research, while it has the basic goal of understanding the world, does employ a variety of approaches and methods in conducting the pursuit of knowledge. Those strategies can be grouped according to how the research enterprise is designed, how the data are collected and the location of the setting in which the research is performed.

A. Research Design
1. Experiment
Experimental studies are the preferred method for conducting research. Conducting experiments involves controlling and manipulating some conditions, called independent variables, and recording the resulting changes in other conditions, dependent variables. Often large numbers of subjects are used in psychological experiments because the variability in individual behavior can make conclusions difficult to draw. Because of the control over variables, experiments are the only method which allow for cause-and-effect determinations to be drawn. While that is a strong point, the weakness of the method is that since conclusions are drawn from aggregate group behavior, very little, if anything can be inferred about what may influence the behavior of single individuals.

2. Correlation
Correlational studies find associations and relationships between factors, or variables. However, there is no control or manipulation of those factors, only the recording and determination of events or conditions which tend to accompany each other. There can be no determination of cause-and-effect relationships with any correlational study, no matter how "common-sense" such cause-and-effect relationships may appear to be.

3. Descriptive
Descriptive studies involve the intensive detailed collection of information about single individuals or specific small groups of individuals. However, while a great deal can be learned about the history, motives and behavior of individuals, there is no control over the conditions or the accuracy of the information collected. For instance, individuals may lie or the observer may have some bias in the information collected or how they interpret or filter it.

B. Data Collection
1. Self Report
Self report data include questionnaires, surveys and interviews. The strengths are that self reports are relatively quick and easy to collect or administer. However, questions can be poorly written or asked introducing some bias or confusion to the data. People may lie on a questionnaire or in an interview out of embarrassment or an attempt to please the researcher. And interviewers may also insert their own bias in their interpretation of a subject's responses

2. Observation
The collection of observational data is as minimally intrusive as possible, which minimizes any changes in behavior that might be unintentionally elicited by the researcher. However, the researchers may still potentially insert their own bias into their observations.

C. Research Setting
1. Laboratory
Laboratory settings include any sort of uniform and controlled environment, such as a hospital, research facility or even a doctor's or therapist's office. Such control is especially beneficial for experimental studies, but correlational or descriptive studies can be conducted there as well. The weakness of the laboratory environment though is that it is artificial and subjects may not behave as naturally as they normally would.

2. Field
The field is anywhere that is not in a laboratory. It could literally be a field, or a school or a day care center or a sidewalk or a grocery store aisle; anywhere that isn’t a laboratory. The advantage is that subjects can now act naturally, but the disadvantage is that conditions are uncontrolled and unpredictable and may change day by day or even minute by minute and make the collection or interpretation of data difficult or even impossible in some cases.

IV. Correlation vs. Experiment
Experimental and correlational studies are often confused by the general population. The distinction between causation and association is often lost, especially in the popular press where many people hear or read mistaken or misrepresented reports of research findings.

Association vs. Causation
Perhaps the source of confusion that arises from the sorts of conclusions that can and cannot be drawn from correlational and experimental studies is that "causes" and "effects" are associated together but it is a very specific type of relationship called a causal (from the word, "cause") relationship. But what scientists and statisticians mean when they say that correlational studies reveal associations is that the factors can only be said to tend to occur together and nothing further can be concretely stated. However, some correlatioanal associations are often taken as indicating a cause-and-effect relationship. For instance, a correlational study may indicate that there is an association between exercise and long life. Informally, most people assume that exercise causes a longer life, but that may not be the case. All that can be drawn from the study is that people who exercise more live longer. There may be additional hidden factors that may influence BOTH tendencies. For instance, general health may influence exercise and longevity. It may be that the healthier people are the simply more likely to exercise and also their better health leads to a longer lifespan. Or it could be other acquired habits that influence both factors. Heavy smokers and drinkers might be less likely to exercise because of their substance use and also more likely to die earlier. To illustrate things perhaps more clearly, there is a strong correlational association between murder rates and ice cream sales, both reaching their peaks in the summer months. But it would be a mistake to conclude that ice cream consumption triggered homicidal tendencies or that committing murder increased the appetite for frozen dairy treats. It would be wiser to suspect that an increase in heat and humidity might underlie both increases.

The experimental approach, however, can determine causation and causal relationships. In the simplest possible example, experimentation does this by taking a practical approach and selecting a sample of subjects from the general population of all possible participants. The selection of the sample should be conducted randomly from the population to best approximate the random distribution of the general population's traits and conditions. Non-random selection procedures, even such innocent-seeming ones like asking for volunteers, may exclude some possible subjects and unintentionally bias the research results. The factors, or variables, to be considered are then weighed or assigned. The independent variable is the variable thought to be the cause and the dependent variable is thought to be the effect. Subjects will be randomly assigned to groups, one in which the independent variable will be present and one in which it will be absent. Any differences in the dependent variable will be observed and recorded. It is the capacity to directly compare the differences in the values of the dependent variable between groups based on the presence or absence of the independent variable that allows the experimental procedure to determine casual relationships.

V. Theory & Experiment
The Scientific Method
The scientific method relies on practical and theoretical methods to experimentally collect facts, or data, control the manipulation of factors, or variables, and explain the outcomes of our observations as well as to refine those procedures. The goal of the scientific method is to continually refine our understanding of the world. Scientific discoveries and understanding are constantly under revision. Not necessarily because the original findings were incorrect, but because our understanding and knowledge is continually growing.

In the beginning of a scientific inquiry, a researcher becomes curious about a phenomenon or condition in the world. In an effort to understand that phenomenon an explanation is advanced. That explanation is a theory. The theory, or explanation, may or may not reflect the truth about the phenomenon. But if that explanation is true, logical reasoning will lead to a prediction about the outcome should certain conditions be changed. That predictive statement is a hypothesis. Experiments are procedures which test the hypothetical predictions that arise from a theoretical explanation. The next step is to design the actual procedures and methods to test the hypothesis though more abstract reasoning. Once the procedures and methods have been prepared the experiment is conducted and the resulting data are compared to the hypothetical predictions. If the results verify the predictions, the cycle is complete and the theory has been supported. If the results do not correspond to the predictions, which happens more often than not, then the theory (or the experimental procedures or both) need revision. The cycle begins again and continues with constant refinement until eventually the results match the predictions of the hypothesis. One important thing should be noted. Even if a theory is verified by an experiment, it remains a theory, because a theory is an explanation and an explanation is an explanation, whether it is an accurate one or an inaccurate one, whether it has been tested experimentally or not.

VI. Statistical Methods
You may ask how can we be certain that the results of our research are reliable or true? How can we know if our theoretical explanations and our hypotheses reflect the real nature of the world? Especially when we’re often dealing with groups of subjects and their individual responses can vary, sometimes greatly. Scientists do this with statistics, a branch of mathematics that relies on the analysis of probabilities to gauge the reliability differences in groups of data.

A. The Correlation Coefficient
Correlational studies rely on the calculation of a number called a correlation coefficient to determine how closely two values are related. It is a number between -1.0 and 1.0 which measures the degree to which the values observed for two variables would fall on an imaginary straight line, with one variable's values plotted against a y-axis and the other variable's values plotted against a x-axis on a grid. If there is perfect linear relationship between the two variables where as both values increase in a positive direction along the grid (both values increasing), we have a correlation coefficient of 1.0. If there is a perfect linear relationship with one variable's values decreasing as the other variable's values increase, we have a correlation coefficient of -1; a perfect negative correlation. A correlation coefficient of zero means that there is no linear relationship between the two variables.

B. Populations, Samples & Distributions
A population is the sum total of all people, animals, or things that we may study and from which we wish to collect data and draw conclusions. However, it is typically not a practical matter or even a possibility to include an entire population of interest in a research study. Therefore, a sample that is meant to be representative of the population is often studied instead. For example, we may want to study intelligence in children, but we will be unable to include all the children in the world, so we may pick 100 children selected at random. Statistically, we would expect that the distribution of values for a random sample will reflect that of the entire population. Such distributions are typically expected to cluster around a central value.

1. Variability of the Distribution
While the observed values may have a central value, there will be greater and lesser values varying around it. The "spread" around that central value can be arithmetically calculated (the exact formula is not of concern to us here) and that number is called the variance and tells us how scattered the values of our observations are; the larger the number value of the variance, the greater the scatter.

a. Standard Deviation
The standard deviation is the square root of the variance and as such also a number that gives us an indication of the scatter of our measured values. If we go up and down one standard deviation from the central value, two-thirds of all our observations should fall within that range. The larger the standard deviation, the broader the scatter. The smaller standard deviation, the tighter the concentration of values around the central value.

2. Centrality of the Distribution
The values we observe center around one value. However, there are several ways to describe the most central value. Under ideal conditions the three descriptions of the central value will be the same. However, that may not always be the case. There can be a need for more than one way to describe that central value because our samples may not always smoothly represent the variability within the population, especially when we have a small sample.

a. Mean
The mean is simply the arithmetic mean or average. It is calculated by adding up the values of all our measurements and then dividing the grand sum by the total number of subjects.

b. Median
The median is determined by counting the total number of subjects and ordering the measured values from lowest to highest. The value which is greater than half the observations and less than the other half is the median value.

c. Mode
The mode is the most frequently occurring value within the distribution. However, sometimes there can be more than one mode. Distributions with multiple modes can sometimes (but not always) be a clue that our sample may not be a true representation of the population and may contain identifiable subpopulations.

C. Comparing Distributions.
When we compare the distributions that our data form, there are specific statistical tests that we can use to determine if our experimental manipulation of independent variables has reliable changed the values of our dependent variable. Some of the most commonly used are t-tests and ANOVAs (short for analysis of variance). T-tests are used when there is only one comparison made between two distributions. This can be the distributions of two groups or even two distributions of data from one group collected at two points in time; like before a drug is administered and then afterwards. Then more than one comparison is being made, for instance a drug study with a placebo group and two or more doses of a drug, then an ANOVA is used. One thing to keep in mind is that the power to reliably determine differences between distributions depends on the number of observations (the size of the groups). For instance, a small difference in the arithmetic mean between two groups may not be statistically significant when there are only 10 subjects per group. But the same sized difference in the mean may be significantly different if there are 50 subjects (or even more) per group.