Statistics & Probability

Measures of central tendency:

• Mean = sum of all values ÷ number of values
• Median = middle value when arranged in order
• Mode = most frequently occurring value

Measures of spread:

• Range = highest − lowest
• Interquartile range (IQR) = Q3 − Q1 (middle 50% spread)

Graphs:

• Pie charts — each sector represents a proportion (sector angle = value/total × 360°)
• Bar charts — frequency on y-axis, categories on x-axis
• Histograms — area of bar = frequency (used for grouped continuous data)
• Frequency polygons — join midpoints of class intervals
• Scatter graphs — show correlation (positive, negative, or none). Draw a line of best fit.

Probability:

• P(event) = favourable outcomes ÷ total possible outcomes
• P ranges from 0 (impossible) to 1 (certain)
• P(A or B) = P(A) + P(B) for mutually exclusive events
• P(A and B) = P(A) × P(B) for independent events
• Tree diagrams help organise multi-stage probability problems.

Deep dive: build the idea, then use it

Conceptual model

Statistics is disciplined storytelling with data. Start with the question, population, sample, and variable; then choose a display and summary that do not hide important variation. Categorical data needs counts or proportions, while numerical data can show centre and spread. Sampling can be useful without being perfect, but a biased method cannot be repaired by a neat graph.

Worked example

Times (minutes) for five journeys are 18, 20, 21, 22, 49. Mean is 130/5=26; median is 21; range is 31. One delayed journey changes the mean strongly, so reporting median and range gives a fairer picture of a typical trip and its spread. If grouped intervals are used, a midpoint mean is an estimate because individual values are unknown.

Exam-method habit

State units and a meaningful title. On a histogram, use frequency density when class widths differ; area represents frequency. Compare distributions using both centre and spread, and do not infer causation merely because a scatter graph trends upward. When a conclusion goes beyond the sample, use cautious language such as “the sample suggests”.

Common errors to catch early

Using a bar chart for continuous grouped data; omitting a zero break without clear marking; calculating median from an unordered list; calling a convenience sample representative; and claiming two variables cause each other because they correlate.

Retrieval drill — close the notes, phir try karo

Classify shoe size, height, and favourite subject by data type. Find median and IQR of 3, 4, 6, 8, 10, 12, 17. Explain when a histogram is better than a bar chart. Write one limitation of an online voluntary survey.

A Pakistan-relevant use

A local sports club can survey members about preferred practice times using an anonymous, clearly worded form and a spread of age groups. The results can guide a trial timetable, then be checked against attendance rather than assumed to represent every family in the area.

Concise summary

Good statistics begins before calculation: ask how data was collected. Then show centre, spread, uncertainty, and the limits of the conclusion.

A reliable self-check routine

Before accepting an answer, say what each quantity, symbol, particle, or graph feature means. Then check its unit, sign, direction, size, or conservation rule. In a calculation, write the relationship first, substitute with units, calculate, and decide whether the result is sensible. In an explanation, make a chain: cause → mechanism → observed result. This is not extra decoration; it is how a reader can follow your thinking and how you catch a copied digit or an attractive-but-wrong statement. If the question gives a new context, do not hunt for a memorised sentence. Identify the model underneath it and apply that model. Short, precise working beats a long paragraph that never answers the command word.

Practice plan

Try one straightforward question without notes, one mixed question where you choose the method, and one question where you explain why an answer is reasonable. Mark the exact first step that felt uncertain. Revisit that step the next day for two minutes instead of rereading everything. Small retrieval loops make the topic stick, yaar.

Connect and transfer

This topic becomes stronger when you deliberately meet it in an unfamiliar wrapper. A diagram may be rotated, a calculation may use an awkward unit, a practical may describe an everyday object, or a question may provide more information than you need. Pause and sort the information into three columns: given, wanted, and relationship. That small pause prevents the common rush of putting every number into the first formula remembered. If the answer is qualitative, decide whether the task is asking for a prediction, a description, or a mechanism; these need different sentences. If it is numerical, estimate its order of magnitude before the calculator. If it involves a graph or table, describe the relevant trend using the actual variables before explaining it.

Teach the idea out loud in sixty seconds as though a friend missed the lesson. Avoid specialist words you cannot unpack. Then add those words back with their exact meaning. This exposes the difference between recognition (“that looks familiar”) and recall (“I can construct the answer”). Keep an error log with a corrected example, not a list of scores. For the next attempt, cover the correction and reproduce the decision that led to it. The goal is calm, repeatable reasoning—not racing through a page. When your final answer differs from a friend’s, compare the model and assumptions before comparing calculators. Often the useful learning is in the first different step.