CCommons

Are units, dimensions and error analysis really easy marks, or do people lose them too?

The chapter looks gentle, yet marks leak through sign slips, missing factors and mixed units. What does it deserve, and where do the slips come from?

Commons Community Team

15 Sept 2026 · 1 min read

Units and dimensions look like the gentlest chapter in the Physics book, and many of you say you skimmed it in a day. Then a mock asks for the error in gg from a pendulum experiment, or a dimensional formula for a quantity you have never met, and the marks vanish.

The chapter has three parts: dimensional analysis (checking and deriving relations), measurement (significant figures, least count, vernier and screw gauge), and error propagation. Each is small. The difficulty is that they are needed everywhere, so a shaky foundation leaks marks in later chapters too.

Do you treat this as easy marks or as a chapter that deserves real time? Share what you do for it and which part keeps catching you.

WhatsAppPost on X

5 answers & replies

Sign in to join the conversation.

  • [+]Commons Physics DeskTeam· 1mo ago

    Easy to learn, and still easy to lose. Three habits cover the chapter.

    1. Principle of homogeneity. Every term in a valid equation has the same dimensions. Use it to check an equation, convert units or derive a relation. It cannot fix numerical constants like 2π2\pi, and anything inside a trig or log function must be dimensionless.

    2. Derive by matching. Suppose a pendulum's period depends on ll and gg: T∝lagbT \propto l^a g^b. Dimensions: [T]=[L]a[LT−2]b[T] = [L]^a[LT^{-2}]^b. Time gives 1=−2b1 = -2b, so b=−12b = -\tfrac{1}{2}. Length gives 0=a+b0 = a + b, so a=12a = \tfrac{1}{2}. Hence T∝l/gT \propto \sqrt{l/g}. The 2π2\pi must be known separately.

    3. A small set of dimensional formulae to know cold:

    • force [MLT−2][MLT^{-2}], energy and work [ML2T−2][ML^2T^{-2}], pressure [ML−1T−2][ML^{-1}T^{-2}]
    • Planck's constant has the dimensions of angular momentum, [ML2T−1][ML^2T^{-1}]
    • impulse and momentum share dimensions; so do torque and energy (different quantities, same dimensions)

    Dimensionless quantities worth recognising: strain, angle, refractive index, relative density.

    The limit: dimensional analysis cannot give you a numerical factor, and it cannot choose between two dimensionally correct forms.

  • [+]Commons Mock & PYQ DeskTeam· 1mo ago

    Marks vanish in error analysis because of loose rules, so keep three.

    • Sum or difference: absolute errors add. Δ(A−B)=ΔA+ΔB\Delta(A - B) = \Delta A + \Delta B. Yes, they add even for a difference.
    • Product or quotient: relative errors add.
    • Power: if Z=ApZ = A^p, then ΔZZ=∣p∣ ΔAA\frac{\Delta Z}{Z} = |p|\,\frac{\Delta A}{A}.

    Worked example: g=4π2lT2g = \dfrac{4\pi^2 l}{T^2} with l=100.0 cm±0.1 cml = 100.0\ \text{cm} \pm 0.1\ \text{cm} and T=2.00 s±0.01 sT = 2.00\ \text{s} \pm 0.01\ \text{s}.

    Δgg=Δll+2 ΔTT=0.001+2(0.005)=0.011\frac{\Delta g}{g} = \frac{\Delta l}{l} + 2\,\frac{\Delta T}{T} = 0.001 + 2(0.005) = 0.011

    So about 1.1 percent. The usual blunder is forgetting the factor 2 on the TT term.

    Other recurring items: significant figures (leading zeros do not count; trailing zeros after a decimal point do), the least count of a vernier or screw gauge, and mean absolute error from repeated readings.

    In mock analysis, put unit and error slips in a separate column from concept mistakes. They are cheap to fix and expensive to repeat.

  • [+]Commons Chemistry DeskTeam· 1mo ago

    Chemistry borrows this toolkit more than students expect, mostly in physical chemistry numericals.

    • Units of rate constants come straight out of dimensional thinking. For a reaction of order nn, the unit of kk is (mol L−1)1−n s−1(\text{mol L}^{-1})^{1-n}\,\text{s}^{-1}. Zero order: mol L−1^{-1} s−1^{-1}. First order: s−1^{-1}. Second order: L mol−1^{-1} s−1^{-1}.
    • The gas constant RR appears in different unit systems (J mol−1^{-1} K−1^{-1} and L atm mol−1^{-1} K−1^{-1}, for example). Mixing them is a classic way to get an answer wrong by a large factor. Check the units of pressure and volume in the question, then pick the matching RR.
    • Concentration units: molarity and molality are different things. Molarity depends on volume, so it changes with temperature; molality does not.
    • Significant figures at the end of a long calculation.

    A habit that works: write the unit next to every number in the first line of the numerical and let the units cancel on the way. If the final unit is wrong, the working is wrong, and you find out before you tick an option.

  • [+]Commons Biology DeskTeam· 1mo ago

    Biology has its own small unit zone, and it is easy to ignore until a question uses it. Know the ladder.

    UnitIn metres
    millimetre (mm)10−310^{-3}
    micrometre (μ\mum)10−610^{-6}
    nanometre (nm)10−910^{-9}
    angstrom (Å)10−1010^{-10}

    NCERT Biology quotes sizes of cells, organelles and viruses in these units. When a question compares, say, a cell with a virus, it is mostly a matter of counting powers of ten. Do not memorise a list of sizes from some random page; anchor each one to what the NCERT text actually states.

    The ladder also appears in microscopy: magnification is image size divided by object size, and the answer only makes sense when both are in the same unit. Same habit as in Physics: write the unit with every number.

  • [+]Commons Strategy DeskTeam· 1mo ago

    So, easy marks or not? Easy to learn, and still easy to drop. For time: finish the concepts early, since they feed every later chapter, then do a short mixed set of ten unit and error questions once a week for a month. After a month you should be able to do them without checking.

    A neat habit: keep one sheet where you write the dimensional formula of every quantity you meet in a new chapter. By the end of the year you have a complete table that you wrote yourself, which sticks better than one you read.

    Check the current NTA syllabus notice for exactly which parts of this unit are included. And tell us what trips you up: significant figures, error propagation, or deriving a relation? Post your slip below and see whether others share it.

More from Commons Community