Glossary of common research terms

Porter,S (2003) Tidys Physiotherapy 13th edition Butterworth Heinemann Publishers Elsevier Science

Abstract a succinct summary of a research paper

Action research where the researcher uses some form of intervention in a situation and
evaluates the impact of the intervention
Bias Deviation of the results from their true values. Examples include:
Spectrum Bias The sample population chosen is not representative of the
population.
E.g. an appropriate spectrum of students were not included in the study of their
exam results
Measurement Bias the act of being studied or measured can affect the outcome.
Confounding Bias Occurs when two factors are closely associated and the
effects of one confuses or distorts the effects of the other factor. The distorting
factor is called a confounding variable.
Recall Bias The recall of events may differ in cases and controls. Questions may
be asked more times and more intensively in cases compared to controls.
Referral Bias Health care referrers may attract individuals with specific
disorders or exposures.
Volunteer Bias Volunteers may exhibit outcomes that may differ from nonvolunteers (e.g. volunteers tend to be healthier).
Withdrawal Bias Patients who withdraw from studies may differ systematically
from those who remain.
Central Tendency The middle of a distribution
Chance: Random variation. Statistical methods are used to estimate the probability that
chance alone has accounted for the difference in outcome.
Coding a qualitative research term, it refers to the task of ascribing codes to concepts and
themes that occur in an interview transcript
Cohort Study A longitudinal study that begins with the gathering of two groups of patients
(the cohorts), one which received the exposure of interest, and one which did not, and then
following this group over time (prospective) to measure the development of different
outcomes.
Confidence Interval often expressed as 95% confidence interval. The confidence interval
(CI) quantifies uncertainty. Derived from the sample mean and the standard error.
Control group: The study patients that have not received the experimental manoeuvre or test
Correlation The link between two variables, always represented by a figure between minus 1
and plus 1

correlation
correlation
of +1
of -10

Positive correlation
As on e rises , so does the other e.g.
length of shoes versus length of your
foot

Negative correlation, as one goes up ,

the other goes down eg the older an
adult gets, the less tall they become

No correlation
e.g. the relationship between the
colour of your socks and what
you had for your breakfast

Cross-sectional study: Survey of an entire population for the presence or absence of a

disease and/or other variable in every member (or a representative sample) and the potential
risk factors at a particular point in time or time interval.
Crossover Study Design: The administration of two or more experimental therapies one after
the other
Data(singular = datum) the pieces of information, facts and figures that have resulted from
the study
Data can be one of four types:

 Nominal (naming data e.g. male or female, yes or no)

Ordinal slightly more detailed, puts data into an order
e.g. strongly agree. Agree. neither agree nor disagree. Disagree . disagree strongly
Interval as above but assumes equal intervals between the categories
Ratio as above but also possesses a zero point
Dependant variable the variable which alters as a result of manipulation of the independent
variable
Double-blind An experimental method in which both the patients and the research staff do
not know which patients are receiving treatment and which are receiving placebo.
Epistemology the theory of knowledge
Ethnography is the description or study of a culture
Exclusion criteria the criteria by which members of the population of the sample will be
excluded, for example in a study of heart rate, people with a history of cardiac problems
would be excluded
Independent variable the variable which is actually changed (manipulated) so that its effects
on the dependant variable can be seen
Field notes the records or reflections kept by the researcher which act to enrich the data
obtained or put them into some sort of context, commonly used in qualitative research
Focus group the term for a group of people who have been gathered together in order to gain
some insight into their ideas and attitudes towards a particular subject
Generalisability given the validity of the results, the similarity of the study population to the
population in general
Grounded theory a research approach which attempts to generate theory form the data
Hawthorne effect When subjects systematically alter their behaviour when they are being
observed.
Hypothesis (pleural hypotheses) a proposition that may either by supported or rejected
Inferential Statistics: Determines how likely a given result occurred by chance alone. Since
we can rarely study an entire population, we study a sample of the population and by
inference apply that result to the entire population.
Intraobserver variability Variability between the same observer on repeated occasions , for
example ,does a senior clinician give the same reading of knee flexion when goniometry is
performed on more than one occasion?
Interobserver variability The variability between different observers. For example, do two
or more senior clinicians give the same reading from the same knee joint?
Intention-to-treat Analysis In a randomised control trial patients can be randomly assigned
to different treatments After randomisation, patients who have been assigned conservative
therapy may decide to have surgery instead , conversely, patients assigned to the surgical
treatment may decide not to undergo surgery. In an intent-to-treat analysis patients would be
analysed for mortality according to the groups for which they were originally assigned.
Investigator Bias Occurs when the interviewer is aware (not blinded) of the outcome
variable. An unblinded interviewer may be more vigorous in searching for the exposure of
interest.
Multivariate Analysis: An analysis where the effects of many variables are considered.
Mean the average number in a set of values.
Median: When values are arranged in order of magnitude, the median is the middle value for
odd numbers of values and the average of the two middle values in the case of an even
number of values.
Mode: the mode is the value that occurs most often.
Non parametric test a statistical test which is less sensitive than a parametric test but which
can be used on nominal and ordinal data

Normal distribution data which is spread out in a bell shaped curve eg height of people or
the size of peoples feet, a few have small feet, a few have large feet but most are in the centre
see below
Number
of
people

Size of peoples feet

Size of peoples feet

Null Hypothesis: The proposal that no difference between groups. If the null hypothesis is
true then the study findings are the result of chance or other random factors. A typical study
aims to "reject the null hypothesis
Parametric test a type of statistical test which is more sensitive and robust than a non
parametric test, in order to be able to carry out a parametric test on your data, the data must be
interval/ratio, and should ideally be normally distributed
p Value: The probability of a finding occurring by chance alone given that the null
hypothesis is actually true. A p value < 0.05 is often considered significant.
Phenomenology the study of the lived experiences of people
Pilot study a small scale test run of the proposed larger research study under the same
conditions, useful as an means of ironing out problems before the study starts
Placebo the term given to an intervention that is actually a mock intervention, for example
giving a person a tablet which has no active ingredients but resembles the active drug in
outward appearance
Population: Every person who satisfies inclusion criteria for the study about to be performed.
Qualitative research Is research not based on numbers, useful for assessing opinions
attitudes, feelings and so on
Quantitative research Is research which is based on numbers
Research Question: The question contains the population, the manoeuvre the study
population, and the outcomes. The research question should specify one measurable outcome,
in addition to all conditions and any other important variables.
Randomisation: The process by which every member of a population has an equal chance to
be included in the sample.
Reliability the repeatability of a study

Randomised Controlled Trial (RCT): An experiment, in which the researcher randomly

assigns some patients to at least one manoeuvre and other patients to a placebo. When
properly done, an RCT can be used to determine cause and effect.
Sample the individuals who satisfied the inclusion criteria and who actually entered the
research study
Sensitising concept a term used to describe how certain concepts may guide the researched
towards a certain behaviour
Significance Level the probability of incorrectly rejecting the null hypothesis, i.e. saying that
there is a difference between two groups when actually there is no difference. Otherwise
known as the probability of Type I error. By convention, the level of significance is often set
to a p value of 0.01(99% significance level) or 0.05(95% significance level)
Theoretical saturation (data saturation) used in qualitative research, the point where no
new themes of concepts emerge form the data being collected
Thematic analysis on qualitative research the researcher attempts to categorise and analyse
the themes that are produced from interviews
Triangulation the process whereby the same data is obtained form various means as an
attempt to improve its validity
Type I Error: rejecting the null hypothesis when it is actually true.
Type II Error: accepting the null hypothesis when it is false.
Standard Error of the Mean A measure of variability. The standard error of the mean
quantifies how accurately the true population mean is known.
Standard Deviation A measure of variability or spread of data. The standard deviation
quantifies how much the values vary from each other. It is a measure of the spread of
individual observations around the mean value of the sample. A normal, unskewed curve will
have 34% of the cases between the mean and 1 standard deviation above or below the mean;
68% of cases between 1 standard deviation above and 1 below the mean; 95.5% of cases will
be within two standard deviations of the mean.
Validity the truth of the research

Internal Validity: Are the results of the study valid for the patient population
studied?
External Validity: This refers to whether or not the results are valid outside the
population that has been studied. For example are results from studies done on dogs
valid for cats?