What makes a test worth trusting, how to build a battery that actually matches an athlete's sport, and the order that keeps an early test from wrecking the ones that follow.
What actually shows up on the exam
Testing only earns its place in a program if the numbers it produces actually change a decision. A strength coach who tests without a plan for the data is just burning a training session — so before picking a single test, it helps to be explicit about which of these jobs the testing is actually doing.
Applied — the test-train-retest loop
Good testing isn't a one-off event; it's a loop. A baseline test identifies a target, training addresses it, and a retest under the same conditions checks whether the intervention worked. Skip the retest and you're only ever guessing whether the program is doing anything. Change the conditions between test and retest — a different room temperature, a different tester, a different time of day — and you've broken the comparison before it started.
Three words get used almost interchangeably in casual conversation but mean specific, different things on the exam. Getting the hierarchy straight — test produces a measurement, and a measurement only becomes useful once it's evaluated — clears up most of the confusion.
| Term | Precise meaning |
|---|---|
| Test | A specific, standardized procedure used to collect data on one trait or performance quality (e.g., "a 1RM back squat, performed per protocol X") |
| Measurement | The actual process of gathering that data and expressing it as a number or category — the raw output of running the test |
| Evaluation | Interpreting what the measurement means — judging it against a standard, a norm, or the athlete's own prior score |
| Norm-referenced interpretation | Judging a score against how a broader reference population scored (e.g., "85th percentile among Division I linebackers") |
| Criterion-referenced interpretation | Judging a score against a fixed, predetermined cutoff that has to be met, regardless of how anyone else performed (e.g., "vertical jump ≥ 24 in. to qualify") |
Keeping norm- and criterion-referenced straight
Norm-referenced needs a Neighborhood — you can't rank someone without a group to rank them against. Criterion-referenced is about clearing a Cutoff — pass or fail, met or not met, with no group required at all.
Validity asks a single question: does the test actually measure the thing it claims to measure? A grip-strength test is a valid measure of grip strength; it's a poor, invalid measure of a golf swing, no matter how precisely it's administered. Validity isn't all-or-nothing, either — it's usually established piece by piece, through several different kinds of evidence.
| Type | What it asks | Example |
|---|---|---|
| Construct validity | Does the test genuinely capture an underlying, not-directly-observable trait? | A vertical jump is accepted as construct-valid for "lower-body power" because it's grounded in impulse and rate of force development, not just tradition |
| Content validity | Does the test cover the full breadth of the quality being assessed, in the judgment of subject-matter experts? | A "conditioning" battery built from a single 40-yard dash would be flagged by expert review as too narrow to represent conditioning as a whole |
| Criterion validity — concurrent | Does the test correlate with an accepted gold-standard measure taken around the same time? | A field-based sub-maximal step test scored against a lab treadmill V̇O₂max test administered that same week |
| Criterion validity — predictive | Does the test's score forecast a future outcome? | A pre-season change-of-direction score correlating with which athletes suffer non-contact knee injuries that season |
Where people lose points
Reliability and validity aren't interchangeable, and the relationship only runs one direction. A test can be reliable without being valid — a broken scale that reads 10 lb heavy every single time is perfectly consistent and completely wrong. But a test can never be valid without also being reliable — if it can't produce a consistent number, there's no stable value left for it to accurately measure. Reliability is the floor validity has to stand on.
Concurrent vs. predictive, fast
Con-CURRENT = checked against the gold standard right now, in the current moment. Predictive pre-dicts something that hasn't happened yet. If the criterion measure was taken on the same day, it's concurrent; if the criterion is a future event, it's predictive.
Reliability is consistency — if nothing about the athlete's true ability has changed, does the test keep producing the same number? A reliable test can still be measuring the wrong construct entirely, which is exactly why validity has to be checked separately.
Whenever an actual number differs between a baseline test and a retest, the gap can be traced to one of four sources — and only one of those four is the thing you were actually trying to measure.
| Source | What it reflects | Example |
|---|---|---|
| True difference | A genuine change in the underlying ability | Four weeks of a strength block genuinely raising 1RM back squat |
| Biological variability | Normal day-to-day fluctuation in physiology, unrelated to training | Hydration, sleep, or menstrual cycle phase shifting a bodyweight or power reading |
| Environmental factors | Conditions surrounding the test itself, differing between sessions | Testing outdoors at 95°F versus in a climate-controlled facility |
| Measurement error | Imprecision from the instrument, the protocol, or the person running it | A different tester reading a skinfold caliper, or a stopwatch started late |
Applied — shrinking measurement error before it happens
Most measurement error is preventable, not random: a written protocol everyone follows the same way, the same tester across sessions whenever possible, calibrated equipment, and a familiarization trial so a "learning effect" doesn't get mistaken for a real training effect. The less error a testing protocol lets in, the more of the remaining score change you can actually trust as a true difference.
A test battery that ignores what the sport actually demands is just a fitness assessment wearing the sport's jersey. Good selection starts from a needs analysis of the sport itself, then works outward to which tests would actually capture that.
| Dimension | What it groups tests by | Example split |
|---|---|---|
| By energy system | Which metabolic pathway the effort leans on | ATP-PC (a single max jump) vs. glycolytic (a 300-yd shuttle) vs. oxidative (a distance time trial) |
| By movement pattern | The plane, direction, and joint action involved | Linear sprint speed vs. multidirectional agility vs. vertical vs. horizontal jump |
| By athletic attribute | The physical quality being isolated | Maximal strength, power, speed, agility, aerobic capacity, body composition, flexibility |
| By sport specificity | How closely the test mimics the sport's actual demands | A generic 1.5-mile run for a distance runner vs. a repeated-sprint protocol for a field-sport athlete |
A well-rounded battery samples across all four dimensions rather than stacking multiple tests that all measure the same underlying quality. A team that tests only sprint speed three different ways has one data point, not three.
| Characteristic | Representative tests |
|---|---|
| Maximal strength | 1RM back squat / bench press, isometric mid-thigh pull, grip dynamometry |
| Power | Vertical jump (jump-and-reach), standing broad jump, medicine ball throw |
| Speed | 10–40 m sprint with splits, flying 20 m sprint |
| Agility | Pro-agility (5-10-5) shuttle, T-test, Illinois agility test |
| Aerobic capacity | 1.5-mile timed run, 20 m multistage shuttle (beep test), graded treadmill V̇O₂max |
| Body composition | Skinfold calipers, bioelectrical impedance analysis, DEXA / hydrostatic weighing |
| Flexibility | Sit-and-reach, goniometric joint range of motion |
The theoretically ideal battery loses to reality if it can't actually be run. Selection has to weigh the same constraints every time:
Applied — matching the battery to the sport
A soccer team's battery should lean on repeated-sprint and multidirectional agility tests, because that's the actual movement and energy-system demand of a match. A powerlifter's battery should lean on maximal-strength tests in the competition lifts. Running a V̇O₂max treadmill protocol on a powerlifter tells you almost nothing useful about their sport — it's a valid aerobic test that's simply the wrong test for that athlete.
Even a well-chosen, well-validated test can produce misleading numbers if it's run under the wrong conditions or in the wrong order. Administration is where a lot of otherwise-good testing protocols quietly fall apart.
| Condition | Effect on results |
|---|---|
| Temperature — hot | Raises cardiovascular strain and core temperature faster, inflating heart rate and lowering endurance-test performance independent of actual fitness |
| Temperature — cold | Slows muscle contractile speed and can blunt power and sprint output until the athlete is properly warmed up |
| Altitude | Reduced available oxygen lowers aerobic test performance (V̇O₂max, endurance times) versus sea level, even with zero change in fitness |
| Surface | A softer or less consistent surface (grass vs. rubberized track vs. hardwood) changes ground contact time and can shift sprint, jump, and agility scores |
| Time of day | Core temperature and neuromuscular readiness fluctuate naturally through the day, so afternoon testing tends to edge out early-morning testing on power and strength measures |
None of these need to be eliminated — they just need to be held constant between a baseline test and any retest. Testing indoors in October and outdoors in July isn't a comparison; it's two different tests wearing the same name.
Fatigue from an earlier test doesn't stay contained to that test — it bleeds into every test that follows it. The standard fix is to sequence the whole battery from least fatiguing and most skill/CNS-dependent to most fatiguing: resting and anthropometric measures and flexibility first, then agility and power (both need a fresh nervous system to be meaningful), then maximal strength, then sprint and muscular-endurance work, and finally anaerobic and aerobic capacity testing, which leaves the athlete too fatigued for anything reliable to follow.
| Situation | Guidance | Why |
|---|---|---|
| Between attempts of the same max-effort test (jump attempts, 1RM trials) | ~1 min for jumps/throws; 3–5 min for near-maximal strength lifts | Lets the phosphagen system resynthesize enough ATP-PC for another genuine maximal effort |
| Between two different max-strength or power tests in one session | At least 5 min | Residual neuromuscular fatigue from the first test can blunt output on the second even after ATP-PC has refilled |
| Before an aerobic capacity test that follows other testing | Full recovery — often scheduled as its own separate session | Aerobic tests are sensitive to any pre-existing fatigue, which suppresses the result and misrepresents true aerobic fitness |
| Order across the whole battery | Non-fatiguing / skill-dependent tests → maximal-effort tests → fatiguing capacity tests, in that order | A fatiguing test run early corrupts every test that comes after it, not just itself |
A common mix-up
It's tempting to run the 1.5-mile aerobic test first "to get it out of the way" while everyone's fresh. That's backwards — the aerobic test should be run last precisely because nothing needs to come after it. Run it early and every strength, power, and agility number recorded afterward is contaminated by residual fatigue, not a true reflection of those qualities.
Which term describes the process of interpreting a raw score against a standard to decide what it means?
C — evaluation. The test is the procedure, the measurement is the raw data it produces, and evaluation is the judgment made once that data is compared to something.
A strength coach tells an athlete they need a vertical jump of at least 24 inches to make the roster, regardless of how any other athlete scores. This is an example of:
B. A fixed cutoff that has to be met, independent of a comparison group, is criterion-referenced. Norm-referenced would instead rank the athlete against a distribution of scores.
A new field-based step test is scored against a lab treadmill V̇O₂max test given to the same athletes that same week. This is evidence of:
C. Checking a new test against a gold-standard measure taken around the same time is concurrent validity. If the comparison were to a future outcome, it would be predictive instead.
Which statement about the relationship between reliability and validity is correct?
C. Reliability is a floor validity has to stand on. A consistently wrong measurement (reliable, not valid) is possible; an accurate but inconsistent one is not.
An athlete's grip-strength score is 4 kg higher on a retest than at baseline, with identical equipment, tester, and time of day, and no training occurred in between. The most likely explanation is:
D. With protocol, tester, equipment, and environment all held constant and no training stimulus, day-to-day biological fluctuation (hydration, sleep, etc.) is the best explanation for the change.
Two different coaches independently administer the same sit-and-reach test to the same athlete on the same day and get similar scores. This demonstrates:
B — intertester reliability. Consistency between two different testers is intertester (interrater) reliability. Consistency from the same tester across trials would instead be intratester reliability.
A program is designing a test battery for competitive powerlifters. Which choice best reflects the principle of sport specificity?
C. Powerlifting is a maximal-strength sport judged on specific barbell lifts — a battery should measure that directly rather than substituting an aerobic-capacity test that has little bearing on competition performance.
Within a single testing session, which sequence best follows the recommended least-fatiguing-to-most-fatiguing order?
B. Power and skill-dependent tests come first while the nervous system is fresh, maximal strength follows, and the most systemically fatiguing test — aerobic capacity — goes last so it doesn't contaminate everything after it.
Roughly how much rest should separate two near-maximal attempts on the same strength test?
B. 3–5 minutes gives the phosphagen (ATP-PC) system enough time to resynthesize so the next attempt is a genuine maximal effort rather than one blunted by residual fatigue.
A team retests its athletes' sprint times at altitude after establishing baseline times at sea level, and every athlete's time gets slower. What does this most likely reflect?
C. Altitude is an environmental condition, not part of the athlete's underlying speed. Testing conditions weren't held constant between baseline and retest, so the comparison doesn't isolate a true training effect.
This chapter is original study material covering topics also addressed in Essentials of Strength Training and Conditioning, 5th edition (NSCA/Human Kinetics). It is an independent work, not a reproduction or paraphrase of that text. This guide is not affiliated with, endorsed by, or sponsored by the National Strength and Conditioning Association (NSCA). CSCS® and Certified Strength and Conditioning Specialist® are registered trademarks of the NSCA.