Terminology for Research Methods

Abstract

Summary of a study, found at the beginning of a research article or listing of research presentations.

Bias

Potential unintended or unavoidable effect on study outcomes.

Control Group

In experimental-type design, independent variable is withheld from these participants.

Control

Set of actions that manipulate factors to minimize extraneous variance.

Correlation Analysis

Method of determining relationships among variables.

Convenience Sampling

Enroll available participants as they enter the study until the desired sample size is reached; non-randomized.

Descriptive Statistics

Used to reduce large sets of observations (scores) into more compact and interpretable forms (e.g., mean, media, mode, standard deviation, etc.).

Descriptive Research

Yields knowledge of population parameters and relationships among those parameters.

Effect Size

Magnitude of the differences among dependent variables.

Directional Hypothesis

Direction of the effect of the independent variable on the dependent variable is clearly defined.

Experimental Group

Group in the experimental design that receives the experimental (treatment) condition.

Evidence-based Practice

Model of practice in which decisions are supported by research, experience, and patient preferences.

Inference

Extent to which the samples reflect the population at both pretest and posttest times.

External Validity

Capacity to generalize findings and develop inferences from the sample to the study population.

Internal Validity

Ability of the research design to answer accurately the research question by minimizing the possibility of bias.

Institutional Review Board (IRB)

Institutional review board; board of experts involved in a research process to oversee the ethical conduct of research.

Likert Scale

Type of scale scored on a 5- to 7- point range; indicating participant’s level of positive or negative response to an item.

Levels of Significance

Probability that defines how rare or unlikely the sample data must be before the researcher can reject the null hypothesis (usually .05, .01, or .001)

Measures of Central Tendency

Numerical information regarding the most typical scores in a group (e.g., mean, median, mode).

Mean

Average score calculated by adding scores and dividing the sum by the number of scores

Median

Point in a distribution at which 50% of cases fall above and 50% fall below.

Meta-analysis

Type of review study in which investigators use quantitative procedures to statically combine the results of multiple studies.

Quasi-experimental design

Experiments that have treatments, and outcome measures but do not use random assignment.

Purpose Statement

Statement that articulates the specific purpose or reason for conducting the research study; identifies goals of the study.

Research Design

Plan that specifies and structures the action processes of collecting, analyzing, and reporting data to answer a research question.

Randomization

Selection or assignment of subjects based on chance.

Retrospective Studies

Describe and examine phenomena after the fact or after the phenomena have occurred.

Sample

Subset of the population participating or included in the study.

Sampling Error

Difference between the values obtained from the sample and what actually exists; selecting another sample from the same population MAY produce different outcomes.

Statistic

Number derived from a mathematical procedure as part of the analytical process in experimental-type research.

Standard Deviation

Indicator of the average deviation of scores around the mean.

True Experimental Design

Classic two-group design in which participants are randomly selected and assigned to experimental or control group condition.

Survey Designs

Nonexperimental designs used to measure primarily characteristics of a population, usually through mail, phone, interview.

Type I Error

Rejecting the null hypothesis when it is true.

Triangulation

Use of multiple strategies or methods as a means to strengthen the credibility of findings.

Validity

Extent to which an investigator’s findings are accurate or reflect the underlying purpose of the study.

t-test

Parametric test to determine the extent to which any significant differences between the means of two sample groups can be inferred from the sample to the population.

Measure of Variability

Range, mean, variance, standard deviation.

Type II Error

Failing to reject the null hypothesis when it is false.

Independent Variable

The treatment or intervention conditions or groups that may be manipulated (e.g., treatment vs. control, or different age groups).

Variable

Concept or construct to which a numerical value is assigned. R

Reliability

Shows outcome measure/test is repeatable and consistent.

Variability

Summary measure, such as range or standard deviation, that describes the distribution of observed values.

Random Assignment

Places participants into conditions randomly; no system

Dependent Variable

Effect or outcome that is measured, as a result of differences in the independent variable.

Statistical Significance

Means that the outcome of the study sample is very likely true in the greater population.

Hypothesis

Prediction about the outcome of the study.

Observational Research

Records outcomes related to a specific topic with no manipulation of independent variable.

Literature Review

Critical evaluation of materials already published.

p value

Statistical “score” used to determine whether results are *probably* true.

Confidence Interval

Range of outcome scores that are likely to contain the true population mean.

Research Question(s)

Narrowly focused question that identifies independent and dependent variables.

Population

Overall group represented by the participants in the study.

Alternative Hypothesis

Prediction that there WILL be a difference between control and treatment groups.

Confounding Variable

Additional variable that can influence the dependent variable and the independent variable and lead to invalid results.

Quantitative Research Designs

Data consist of numbers, scores, or frequency counts.

Random Selection

Pulls participants for the sample at random; no system.

Case Study

Published reports of case materials obtained systematically while working with an individual.

Inferential Statistics

Used to make judgements of the probability that an observed difference between groups is likely true rather than a chance occurrence.

Replication

When a research repeats a study protocol with a different sample group from the same population; lends to generalizability of the study.

Statement of the Problem

Sets the scene for a research article by describing the overall issue addressed by the study.

Homogeneous

Alike; of the same kind.

Empirical Research

Tests the effectiveness of a treatment by manipulating the independent variable.

MBD

Multiple baseline design - uses a single-subject research design with multiple participants who are treated individually.

Filtered Evidence

Review article that identifies, appraises, and summarizes all of the research on a specific topic.

Non-parametric Statistics

A branch of statistics that does not assume that sample data come from a normally-distributed population.

n =

Sample size, represents the number of participants in the study.

Qualitative Research Designs

Data consist of words in the form of interviews, diaries, journals, newspapers.

Citation

Information in a publication or poster that attributes specific information to its author.

Single-subject research designs

Each participant treated and results analyzed individually.

Plagiarism

Expressing or using the writing of someone else as your own, whether intentional or not.

Self-plagiarism

When a researcher presents his/her own work as new scholarship when it was previously-published.

Mode

Most frequently occurring score in a distribution of scores.

Heterogeneous

Diverse in character or content.

Research or Treatment Fidelity

Data that demonstrate the research protocol or treatment was administered correctly and consistently.

People First Language

Practice of speaking and writing that syntactically places the “person” noun first, followed by descriptors.

Null Hypothesis

Prediction that there will be no difference between control and treatment groups.

Parametric Statistics

A branch of statistics that assumes that sample data come from a normally-distributed population.

Disseminate

To share information (esp. research results) widely and using a variety of venues.

No quotes!

Rule about including direct quotations in papers assigned by Dr. Hanson’s classes.

Concomitant

Naturally accompanying or associated with, as in: hypokinetic dysarthria is concomitant with Parkinson’s disease.

Stakeholder

Person or person’s with a direct interest in an intervention (e.g., client, family, teachers, etc.).

Internal Validity

The ability of a research design to minimize bias/implement control, and accurately answer research questions.

Participants

Individuals who serve as the test subjects in a research study.

Population

Overall group that is supposed to be represented by participants in the study.

Treatment Integrity

Selection in a research article that shows the independent variable was handled the way it was supposed to be and was consistent.

Phase Shift

Clear difference in data points showing results plotted on a scatter plot in single-subject research.

Compelling

Describes evidence from well-controlled research designs that may show statistical significance and large effect sizes.

Equivocal

Describes evidence from a weak, uncontrolled research design, published expert opinion, and/or small or missing effect size.

Blinding

Used in experimental research to prevent participants and researchers form knowing who is in the control vs. treatment groups.

Methods

Section of a research paper describing the participants, procedures, and technology used in a research study.

Trend

General direction of data points on a scatter plot (e.g., up or down) suggesting effective treatment.

Stable Baseline

A minimum of 3 data points with very little variability used to show no treatment effect in single-subject research.

Trajectory

A paht, progression, or line of development; e.g., upward indicating improvement

Suggestive

Describes evidence from a research design that may have some bias and/or a moderate effect size.

Longitudinal

Data are gathered for the same participants repeatedly over a period of time - years or decades.

Bar graph (bars don’t touch)

Compares discrete, categorical data.

Histogram (bars touch)

Compares, non-discrete, quantitative data

Results Section

Part of abstract where the outcomes of the study are reported.

Control Group

In experimental research, this group does not receive the treatment or undergo something represented by the independent variable.