Course Title: Scientific Research
Week 4: [Validity and levels of measurement in scientific research]
Instructor: [Mahmoud Salah]
Date: [11/10/2025]
License: CC BY-NC-SA 4.0
Introduction
We must be able to measure accurately in physiotherapy and the health sciences so that we
can understand patients' conditions, track progress, and make decisions regarding the effect of
treatment. A lot of what physiotherapists measure, though, is not directly observable: pain,
motivation, or functional independence, for example. We can't hold "motivation" in our hand;
we must measure them indirectly by means of instruments, tests, and scales.
What Is a Construct?
A construct is a theoretical concept and an abstract entity, e.g., pain, anxiety, strength, or
function. Constructs are the "invisible ideas" we try to understand and measure by means of
observable indicators.
For example:
Muscle strength is a conceptual variable with measurable outcomes as the force in a
dynamometer or by manual muscle testing.
Intensity of pain is a conceptual variable that can be operationalized as a numerical or visual
analogue scale (VAS).
Functional mobility can be operationalized using performance assessments such as the Timed
Up and Go (TUG).
Because constructs cannot be seen, researchers and practitioners must explicitly define them
prior to their measurability. A definition of what a construct is represents the first step in
developing valid and reliable measuring instruments. For example, before measuring
"rehabilitation motivation," a physiotherapist needs to clarify what it is—commitment to the
goal, exercise compliance, confidence, or readiness to participate.
Poorly validated or unreliable instruments may lead to inaccurate clinical judgments and
unsuccessful treatment planning when evidence-based and dependable assessment is
guaranteed utilizing instruments with sound psychometric characteristics in physiotherapy
practice.
Understanding Validity
Validity is the basis of accurate measurement. It is the extent to which an instrument in fact
measures what it is intended to measure—not something else.
Validity is especially important because tests of physiotherapy often decide diagnosis,
assessment of progress, and treatment response. Even highly consistent results from a test
will be meaningless if a test is not valid.
Types of Validity
Validity can be tested from different points of view depending on whether the measurement
precision aspect being considered is face validity, content validity, construct validity, or
criterion validity.
a. Face Validity
Face validity is the degree to which an instrument appears to measure what it should measure,
based on simple judgment. It is the easiest and most subjective form of validity.
For example, a balance confidence survey should have items directly related to the balancing
tasks such as standing on one leg or walking over uneven terrain. If the items don't seem to be
connected (e.g., diet or mood), it lacks face validity.
Although face validity is not tested statistically, it influences the level to which clinicians and
patients trust an assessment. A test that "looks right" reflects confidence and cooperation in
practice and research.
b. Content Validity
Content validity examines if the instrument measures all the important areas of the construct
one is measuring without leaving out key things. It verifies if the test material covers the
whole area of the concept.
For instance, if a physiotherapist develops a measure for functional mobility, items on the
measure have to be on walking, climbing stairs, standing, sitting, and balancing. If the
measure just samples walking, it does not include other essential mobility components, which
is unacceptable content validity.
Content validity is typically determined through expert judgment, whereby the researchers
and clinicians review each item to maximize clarity and appropriateness. This ensures that the
test completely captures the construct in question.
c. Construct Validity
Construct validity is the cornerstone of scientific measurement. It is the degree to which the
instrument is really measuring the theoretical construct it is supposed to be measuring.
Construct validity is cumulatively established by the accumulation of evidence that the
instrument is performing as it ought to relative to other measures or to theoretical standards.
A survey claiming to measure rehabilitation motivation would need to show:
Convergent validity – a high correlation with similar constructs such as adherence and self-
efficacy.
Discriminant validity – low or zero correlation with dissimilar constructs such as anxiety or
intelligence.
Structural validity– an internally consistent form (e.g., factor analysis validating that all items
fall within the hypothesized subscales).
Construct validity is thus defined through theory-constrained testing and statistical analysis
rather than observation. High construct validity in physiotherapy guarantees that a tool is
measuring the targeted concept of rehabilitation and not a different attribute such as fatigue or
mood.
d. Criterion Validity
Criterion validity evaluates to what degree scores on a new test coincide with a prevailing
gold standard or predict future outcomes.
It answers the question: How well does this measure agree with the best current available
indicator of the same construct?
It has two broad subtypes:
Concurrent validity – the new measure coincides with the gold standard simultaneously.
Example: A new portable strength meter for delivering isokinetic dynamometer-type results.
Predictive validity– the tool accurately foretells future occurrences.
Example: A preoperative assessment of mobility that predicts the postoperative rate of a
patient's recovery.
There is a special category called cross-cultural validity, which deals with determining
whether a translated or an adapted version of an instrument is equivalent both meaningfully
and psychometrically to other languages or cultural groups.
Why Validity is Important in Physiotherapy?
All the procedures in physiotherapy affect treatment. An improper test can give incorrect
conclusions regarding a patient's improvement and give suboptimal clinical outcomes. For
instance, if balance testing following a stroke is overwhelmed by lower limb strength and not
by balance control, it might inappropriately give an improvement when strength—but not
balance—had increased.
Effective tools enable physiotherapists to measure what really matters so that they can inform
decisions, monitor rehabilitation precisely, and report facts clearly to other healthcare
professionals and patients.
* Validity must be shown to exist for the target population, culture, and setting in which it is
intended to be used.
* The use of valid and reliable instruments increases measurement accuracy, research quality,
and clinical decision-making in physiotherapy.
Measurement Levels
When you're doing research or checking on patients, you should show how you're measuring
stuff. That's what level of measurement is all about.
How you measure things changes the data you get and the kind of math (stats) you can use on
it. A long time ago, this psychologist, Stanley Smith Stevens, figured out four main ways to
measure: nominal, ordinal, interval, and ratio. They give you different info, so understanding
them means you can pick the best tests.
1. Nominal (Categories)
Think of nominal data as putting things into different boxes. They're just different, with no
real order.
Examples:
* Injury type: could be a sprain, strain, broken bone, or dislocated joint.
* Treatment type: could be hands-on, using machines, exercises, or water-based.
* Body part: arm, leg, or back.
* Patient gender: male, female, or other.
If you give these boxes numbers (like 1 = sprain, 2 = strain), it's just for sorting; it doesn't
mean 2 is better than 1.
2. Ordinal (Ranked)
Ordinal data put things in order, but the spaces from one to the next might not be the same.
So, you know something's better than something else, but not *by how much*.
Examples:
* Pain level: mild, medium, or bad.
* Muscle strength (MMT): 0 (can't move) to 5 (normal).
So, a muscle strength of 4 is better than 3, but the jump from 3 to 4 might be different from
the jump from 4 to 5.
3. Interval (Equal Steps, No Real Zero)
Interval data use numbers with equal spaces, so you can say how different things are. But
zero doesn't mean there's nothing there.
Examples:
* Water temperature: in °C or °F.
The jump from 30°C to 35°C is just like the jump from 35°C to 40°C. But 0°C doesn't mean
no temperature.
4. Ratio (Equal Steps and a Real Zero)
Ratio data are like interval, but zero does mean nothing. That lets you say twice as much or
half as much.
Examples:
* Range of Motion (ROM): measured in degrees (0° = stiff joint).
* Muscle strength: from a handheld dynamometer in kilograms or newtons.
* Walking speed: like meters a second, or how far they got in six minutes.
* Reaction time: in seconds (0 = instant).
* Weight or limb size: in kilograms or centimetres.
Common Scales in Physio
We measure stuff like pain, balance, and confidence with scales. They're based on the four
types above, but tweaked for us.
a. Binary Scales
Binary scales give only a yes/no.
Examples:
* Can they stand without help? (Yes/No)
* Are they using a walking stick? (Yes/No)
* Does it hurt to exercise? (Yes/No)
Give these answers a number (like Yes = 1, No = 0) and add them up to get a score.
b. Likert Scales
Likert scales are those agree/disagree things in surveys.
Examples:
* I feel good when keeping upright outside.
(Strongly Disagree – Disagree – Neutral – Agree – Strongly Agree)
* I'm stoked with how my rehab is progressing.
(Very Unhappy – Unhappy – Neutral – Happy – Very Happy)
References:
1. Bhattacherjee, A. (2012). Social Science Research: Principles, Methods, and
Practices (Revised edition). University of South Florida.
BY-NC-SA 3.0).
Available at: [Link]
2. Physiopedia. (n.d.). Psychometric Properties. Physiopedia. Available at:
[Link]
License
This document is licensed under a Creative Commons Attribution. 4.0 International
License.