Data are the bits of information under study.  Data include measurements, facts, observations, scores, behaviors, etc.  Statistical analysis is used to mathematically understand the data in a study.  A datum is a single unit of a set of data. When referring to data, one must remember that it, grammatically, is the plural form of datum.  Never refer to "data is" or "data was" or "data has."  These should be written and spoken as "data are" or "data were or "data have."

A variable is any quantity or quality whose value varies or can vary from observation to observation.  Examples include age, test scores, IQ, height, weight, attitudes, behaviors, styles, etc.

This is the science (and art) of planning procedures for conducting studies so as to get the most valid findings.  When designing a research study, one draws up a set of instructions for gathering evidence and for interpreting it.

This is research done for the sake of doing research.  Research undertaken with the primary goal of advancing knowledge and the theoretical understanding of the relations among variables.  Hence, it is research that is conducted without a practical end in mind.

This is research that is conducted for some practical purpose -- undertaken with the intention of applying the results to some specific problem.  Applied research often uses the findings of Basic research as its resources. Clinical, Counseling, and School psychologists are often engaged in applied research.

Empirical research is any research in which the data are made up of facts or observations.

Quantitative research is any research that attempts to explore or understand a phenomenon by employing numerical data analytic techniques or methods.

This type of research is any research that produces findings not arrived at by way of statistical procedures or other methods of quantification.  It is an attempt to explore or understand a phenomenon through descriptive narratives, stories, behavior, social movements or interactional relationships.

This is research that describes phenomena as they exist.

This is a type of research that seeks to understand variables by discovering and measuring causal relations among them, and which usually involves variable manipulation and random assignment (i.e., control measures).  Such studies employ either experimental or quasiexperimental designs.

This is a type of descriptive research that seeks to understand relations among variables through measures of association, but not through experimentation.  Hence, though there are variables that may be analyzed, there are no causal relationships between those variables that are under study.

This is any study in which the data collection methods and statistical design are specified before the collection of data.

This is a study in which the statistical analysis is selected after the data have already been collected.  The data in such cases are called archival data.

This is a special type of study in which the data are from an institutional program and whose purpose is often to verify that a program is doing what you think it's doing.  These are typically quantitative in nature.

This is what we use when we use samples to tell us something about the populations from which they were drawn and assessing the precision of our inferences concerning those populations.  That is, "confirming" our hypotheses about populations.  Analogous to criminal law where the rules of evidence are most demanding (i.e., judgment of guilt is rendered based on evidence that suggests guilt  is beyond a reasonable doubt).

Here, we are looking at numerical data to see what they seem to say (i.e., "eyeballing" the data) about a sample from the population.  Attempting to discover potential patterns and relationships among the data, but not to confirm anything.  Analogous to tort or civil law where the rules of evidence are more lax (i.e., judgment of culpability is rendered based only on the preponderance of evidence, much of which may be merely circumstantial).

This is an analysis of analyses -- Distinguished from Primary Analysis (analysis of raw data such as comparing test scores between groups) and Secondary Analysis (re-analysis of raw data originally collected for a primary analysis, but in a new way) such that results of more than one primary or secondary analyses are taken together and analyzed.  It is typically an alternative to narrative literature reviews.  It is the synthesis of a body of research.

This is a question of fact in a subject-matter area from which the research hypotheses or research questions are formed.

This is a question that provides the impetus for experimental or confirmatory research and which is stated in the logical form of the general implication (e.g., if A, then B).

This is an alternative to the research hypothesis, this is a question that provides the impetus for exploratory research and which is, in fact, posed as a question (i.e., How? In what way? Does? etc.).

Once the relevant variables (i.e., Independent; IV; and Dependent; DV; variables) have been identified, the appropriate statistical procedure is selected for testing the statistical hypothesis.  Generally, the type and number of IVs and DVs will dictate the appropriate statistical procedure.

This is a statement, usually in symbolic form, about one or more parameters or other characteristics of a population distribution that requires verification.  Examples: Null Hypothesis (Ho), Alternative Hypothesis (Ha).

• Null Hypothesis

This is the statistical hypothesis whose tenability is actually tested in an experiment.  The process of choosing between Ho and Ha is called hypothesis testing.

This is the decision one makes regarding the statistical hypothesis, usually based on the outcome of the statistical test (e.g., reject the Null Hypotheses,  Fail to reject the Null, etc.)

This is the specific conclusion of outcome of the research study (i.e., the results).

This is a general conclusion derived from the results of the specific study.  This can be thought of as an extrapolated conclusion.

A population consists of an entire set of objects, observations, or scores that are of interest and have something in common. For example, a population might be defined as all depressed adults who seek clinical help or all high school students in America, etc.  Often the population is of such a large size that it is either impossible or impractical to actually observe every singe unit.

A sample is a subset of objects, observations, or scores drawn from a population.  That is, it is a sample from the population.

This is the smallest independently treated unit of study (e.g., treatment group).  Distinguished from observational unit (e.g., subject).

A cause is an event, such as a change in one variable, that produces another event, such as a change in a second variable.  Three conditions are necessary (but not sufficient): 1) X must precede Y; 2) X and Y must covary; and 3) no rival explanations can account as well for the covariance of X and Y.  In experimental studies, the Independent Variable (IV) is the causal variable.

An effect, on the other hand, is an event which necessarily always follows a cause.  In experimental studies, the Dependent Variable's (DV) variability is seen as the effect of the IV's variability.

An experiment is a study, containing at least one IV and at least one DV, designed in such a way that the investigator has full control over all aspects of the study, and manipulates the Independent Variable(s) in order to see how such manipulation affects the variability of the Dependent Variable(s).

A quasiexperment is similar to an experiment.  It is a study, containing at least one IV and at least one DV, but in which the investigator has only partial control over all aspects of the study.  It involves a type of research design for conducting studies in the field or in real-life situations where the researcher may be able to manipulate some IVs but cannot randomly assign subjects to control and experimental groups.  Together, experimental and quasiexperimental designs are often just referred to as experimental designs.

A parameter is a summary index (Mean, Variance, etc.) of a variable from a population of data.  Usually, a parameter is unknown.

A statistic is a summary index of a variable from a sample of population of data.  A statistic is, of course, always known.

This is using statistical results that describe one or more sets of data from a population sample.  That is, Descriptive statistics are summaries of statistics.

Inferential statistics comprie statistical analysis whose purpose is to make inferences of parameters from statistics.  That is, using a sample to draw conclusions about a population.  Inferential statistical analysis is at the heart of most quantitative studies.

This is a puzzle in research that can arise when one wants to study an unknown subject.  (In Meno, Socrates is asked (roughly): How can we seek something if we don't know what it looks like, and, if we already know it, why would we seek it?)  Example: If one does not have the "right" statistical model, one cannot measure the effects of variables, but there is no way to know if one has the right model apart from the measured effects of variables.

Because we cannot study something without affecting it, we cannot know (or must be uncertain about) what it might be like without our interference, when we are not studying it.

This was the demonstration by Kurt Goedel (in 1931) that a formal system, such as logic or mathematics, cannot prove its own basic axioms (self evident propositions).  This has been taken more generally to indicate that we can have little certainty in our claims to knowledge.

This is an error of reasoning committed by coming to conclusions about individuals based only on data about groups.  Example: Concluding that the elderly are more likely to commit crimes because research shows that crime rates are higher in areas with high concentration of elderly persons.  (We'll discuss more of these types of problems later.)

This occurs when there is a possibility that published studies represent a biased sample of all studies that have been completed for a given hypothesis.  The published studies may reflect only those that obtained statistically significant results, and there may be more studies that did not attain significance and were not published.  This is primarily a problem for Meta-analyses (see above).