Glossary
The key technical terms from the articles, briefly explained. In the texts they are underlined with dashes: hover or tap.
1RM (maximal strength)
The weight you can lift with good form exactly once in an exercise.
1RM stands for “one repetition maximum”. Loads are often given as a percentage of it, e.g. 60% 1RM. Because 1RM tests depend heavily on practice and technique, measured strength rises especially when you train that exact movement.
Bayesian analysis
A statistical method that states how likely an effect is, rather than just “significant or not”.
Bayesian meta-analyses give statements like “a small advantage is highly likely”. That's often easier to understand than p-values and suits fields with few, small studies such as sports science.
Calorie deficit
You take in less energy than you use. The body draws on its reserves.
A deficit is needed to lose fat. The larger it is, the more likely you also lose muscle. Plenty of protein and continued heavy training help limit that.
Carnosine
A compound in muscle that buffers acid. Beta-alanine increases carnosine stores.
During hard efforts of about 1 to 4 minutes, acid builds up in the muscle and limits performance. More carnosine can delay that somewhat.
Confidence interval
The range in which the true value most likely lies, usually with 95 % confidence.
A result like “1.6 g/kg” is only the best estimate. The confidence interval shows how uncertain it is. If it's wide, e.g. 1.0 to 2.2 g/kg, the estimate is shaky. If it includes zero, it's unclear whether there's an effect at all.
DXA and BIA
Two methods for measuring body composition: DXA uses X-rays, BIA a weak electrical current.
DXA is fairly accurate for fat and lean mass but can't separate water from muscle. BIA (e.g. body-fat scales) is cheap but reacts strongly to fluid intake and time of day.
Effect size (SMD, d)
A number expressing how large a difference is, independent of the unit of measurement.
Often reported as SMD or Cohen's d. Rough rule of thumb: around 0.2 is small, around 0.5 medium, around 0.8 large. “Statistically significant” only means an effect is unlikely to be due to chance. How much it matters in practice is what the effect size tells you.
Hypertrophy (muscle growth)
A muscle getting bigger, mainly through thicker muscle fibres.
In studies, hypertrophy is measured e.g. by ultrasound (muscle thickness), MRI (cross-sectional area) or muscle biopsy (fibre size). Whole-body lean mass measurements such as DXA are less precise because they also count water.
Lean mass / fat-free mass
Everything in the body except fat: muscle, bone, organs and water.
Many studies measure lean mass instead of muscle directly, e.g. via DXA or BIA. Because water counts too, changes such as with creatine can look bigger than the actual muscle gain.
Lengthened partials
Partial reps only in the stretched part of an exercise, e.g. the bottom half of a biceps curl.
The idea: a muscle loaded in a long position grows particularly well. Early studies are promising, but direct comparisons with full reps are still few.
Median
The middle value: half of the values lie above it, half below.
Unlike the average, the median is hardly distorted by outliers. The “66 days” in the habit study is a median; some people took much less or much longer.
Meta-analysis
A study of studies: the results of many individual studies are pooled statistically.
Because individual sports studies often have only 10 to 30 participants, their results vary a lot. A meta-analysis combines many studies to get a much more reliable estimate, e.g. an average effect of creatine on muscle growth. Special variants are meta-regression and network meta-analysis.
Meta-regression
A method within a meta-analysis that tests whether the size of an effect depends on a characteristic of the studies.
Instead of just computing an average effect, a meta-regression looks at, for example, how muscle growth relates to weekly sets or protein intake. This produces dose-response curves such as “more volume helps, with diminishing returns”. Because it compares studies rather than people, it shows associations, not definitive causes.
Muscle protein synthesis (MPS)
The building of new muscle proteins. Training and protein raise it for several hours.
MPS is easy to measure in short lab studies. But more MPS over a few hours doesn't automatically mean more muscle growth over months. That's why such studies count as limited evidence here.
Muscular failure
The point where, despite full effort, you can't complete another clean rep.
How close you get to failure is often given as RIR. True failure is demanding and increases fatigue. It's therefore riskier on heavy compound lifts and less of an issue on machines or isolation exercises.
Network meta-analysis
A meta-analysis that compares several options at once, including ones never tested head to head.
If A was tested against B and B against C, a network meta-analysis can also compare A with C. Example: low, moderate and high training volume in older adults, drawn from many studies that each had only two groups. This only works if the studies are reasonably comparable.
Position stand
An official statement by a professional society summarising the evidence on a topic.
Examples are the International Society of Sports Nutrition (ISSN) position stands on creatine, caffeine or beta-alanine. They summarise many studies and give recommendations. They are very useful, but not a meta-analysis in themselves, and societies can have conflicts of interest.
Pre-print
A study that's already public but hasn't yet been reviewed by independent experts.
Pre-prints make new results available quickly. Because peer review is still missing, numbers and conclusions can change before final publication. That's why they're always flagged as such here.
Publication bias
Studies with exciting, positive results are more likely to be published than studies finding no effect.
This makes an effect look bigger in the literature than it really is. Good meta-analyses check for it and adjust the effect downwards if needed. An example is the research on if-then plans, where corrected effects are considerably smaller.
Randomised controlled trial (RCT)
A study in which participants are assigned to groups at random, e.g. creatine vs. placebo.
Random assignment makes the groups similar on average. Differences at the end can therefore be attributed to the intervention with reasonable confidence. RCTs are considered the strongest study type for cause-and-effect questions.
Range of motion (ROM)
How far a joint moves during an exercise, from the stretched to the shortened position.
Partial reps use only part of it. Whether you work in the stretched or shortened range apparently makes a difference for muscle growth.
Replication
Repeating a study with the same design to check whether the result holds up.
If an effect doesn't survive a clean repetition, it was probably chance or an artefact. Pre-registered replications across many labs are especially informative, as in the test of “ego depletion”.
RIR (reps in reserve)
How many more reps you could have done at the end of a set. 0 RIR means failure.
2 RIR means you stop two reps short of failure. Many people overestimate their RIR, meaning they're further from failure than they think.
Stoicism
An ancient school of philosophy teaching you to focus on what you can influence yourself.
Founded around 300 BCE in Athens. Well-known Stoics include Seneca, Epictetus and Marcus Aurelius. For them, philosophy was above all a way of life: practice in dealing with thoughts, emotions and setbacks.
Systematic review
A review that searches for and evaluates all relevant studies on a question using fixed, predefined rules.
Unlike a regular review article, a systematic review doesn't pick studies by gut feeling. Search terms, databases and inclusion and exclusion criteria are set in advance. It is often the basis for a meta-analysis.
Training frequency
How many times per week a muscle or exercise is trained.
With equal weekly volume, a higher frequency just spreads the sets over more days. That's why studies need to look at frequency and volume separately.
Training volume
Usually measured as the number of hard sets per muscle per week.
Sets are usually counted if they end reasonably close to failure. Newer analyses count sets where a muscle only assists (e.g. triceps in the bench press) as half sets.
Virtue (in Aristotle)
A character disposition acquired through practice, e.g. courage or moderation.
For Aristotle, virtue is neither an innate trait nor pure knowledge. You acquire it by acting accordingly until it becomes second nature.