Chapter 5 — Statistics

What is statistics?
Statistics is the branch of mathematics that deals with how to collect, organise, summarise and interpret numerical information. It provides tools and methods to turn a long list of numbers into useful statements such as: what is a typical value, how spread out the values are, and whether two groups are similar. In everyday life and in many subjects — science, economics, social studies or sports — statistics helps us to understand patterns and to make decisions based on data rather than guesswork.

Steps in a statistical study
There are clear stages in working with data. First, define the question or objective: what do you want to find out? Second, design how to collect the data: through a survey, experiment, measurement or observation. Third, collect the raw data carefully and record it accurately. Fourth, organise the data using tables or lists so it becomes easier to work with. Fifth, summarise the data using measures such as mean, median and mode and represent it using graphs. Finally, interpret the results in relation to the original question, noting any limitations.

Why organisation matters
Raw data are often messy. Listing values in order, creating frequency tables, or grouping them into classes clarifies structure and saves work. Organisation makes it faster to compute summaries and to spot mistakes, patterns or anomalies. Good organisation also makes graphs clearer and helps others understand your findings.

Data types and basic decisions
Data may be numerical (quantitative) like heights and marks, or non-numerical (qualitative) like colours or categories. Numerical data can be discrete (whole number counts) or continuous (measurements that can take any value in an interval). The type of data determines which measures and graphs are appropriate: you calculate averages for numerical data but not for nominal categories, and you draw histograms for continuous data but bar charts for categories.

Population and sample
A population is the entire set of items or people you are studying, for example all students in a school. A sample is a subset of the population chosen for practical study. Much of statistics is about using sample results to make statements about the population, but this requires careful sampling methods to avoid bias.

Common uses for class 9
At this level you will learn how to prepare frequency tables, calculate measures of central tendency and dispersion, and draw common graphs. These skills are directly useful for class projects, science practicals, school surveys and for understanding statistical information in everyday life or in the news.

Practical advice
Always check raw data for errors and record the units (e.g., cm, marks). When you prepare a report, include the method of data collection and any assumptions made. This makes your statistical work trustworthy and easy to verify.

Overview of data types
When you begin to work with statistics you must first understand what kind of data you have. Broadly, data are of two kinds: qualitative (categorical) and quantitative (numerical). Qualitative data describe qualities or categories, such as eye colour, type of vehicle or blood group. Quantitative data are numbers that measure something, like age, height or marks.

Discrete and continuous quantitative data
Quantitative data split into discrete and continuous types. Discrete data take separate values (often whole numbers), for example number of siblings or number of goals scored. Continuous data can take any value within a range; examples include height in cm, weight in kg or time in seconds. Continuous data are often measured and then rounded for recording, but they are still different from discrete counts because in principle they take many possible values.

Scales of measurement
Statistics uses four measurement scales that determine which summaries make sense:

• Nominal scale: Categories with no order, for example favourite fruit (apple, banana, mango). You can count frequencies and find the mode but not compute a meaningful average.
• Ordinal scale: Categories with a clear order but without equal intervals, such as ranks or ratings (poor, average, good). You can find median and percentiles but mean may not be appropriate.
• Interval scale: Numeric scale with equal intervals but no true zero, such as temperature in Celsius; differences are meaningful and means can be calculated, but ratios are not meaningful because zero is arbitrary.
• Ratio scale: Numeric with a true zero so both differences and ratios are meaningful, e.g., weight, height, and income. Most arithmetic summaries like mean and standard deviation are appropriate.

Why the scale matters
The measurement scale guides your choice of statistics. For nominal data you can only use counts and mode. For ordinal data median and percentiles are useful. For interval and ratio numeric data you can use mean, variance and standard deviation. Using the wrong summary gives misleading results: for example, computing a mean of colours or ranks is meaningless.

Practical classroom decisions
When you collect data, note the scale. If the data are measurements with many decimal places, consider whether you should treat them as continuous or round them and create class intervals. When data are naturally ordinal (e.g., exam grades: A,B,C), use medians for central tendency. If categories are many and unordered, present frequencies or a bar chart.

Examples of categorisation
In surveys you often convert qualitative answers into categories or codes for analysis; be careful to preserve order if it exists. For numeric measurements, if values cluster or cover a wide range you may choose grouped frequency tables for easier analysis. Always state how you treated the data when presenting results so others can judge appropriateness.

Definition and purpose
Ungrouped data are individual observations listed as they were recorded. When the number of observations is small or values repeat, it is helpful to make a frequency distribution that lists each distinct value with the number of times it occurs. This reduces clutter, makes comparisons easier, and helps in computing mean, median and mode quickly.

How to construct a frequency table
First arrange values in ascending order. Then list each distinct value in one column and its frequency (count) in the next column. You can add more columns such as relative frequency (frequency divided by total) and cumulative frequency (running total). For discrete numeric data, include all values even if frequency is zero for completeness.

Tally method
When collecting counts, use tally marks grouped in fives (|||| ) to avoid counting mistakes. Convert tallies into final frequencies once counting is complete. This is especially useful during classroom surveys or while observing events like number of cars passing a point.

Cumulative frequency
Cumulative frequency shows how many observations are less than or equal to a given value. It is obtained by adding frequencies from the start up to each value. Cumulative frequencies are essential for locating medians and percentiles in ungrouped data. For example, if cumulative frequency reaches 50 at a value, that value is the median when total sample size is 100.

Relative frequency and percentages
Relative frequency = frequency ÷ total number of observations. Multiply by 100 to get a percentage. Relative frequencies help compare distributions of different sizes and are useful in drawing pie charts and bar graphs when proportions are required instead of absolute counts.

Using the table to compute measures
From an ungrouped frequency table you can compute mode as the value with highest frequency. For median, use cumulative frequency to find the middle position: for n observations, median position is (n+1)/2 or use cumulative frequencies to find where the middle observation lies. For mean use Σ(x_i * f_i)/Σf_i where x_i are distinct values and f_i their frequencies.

Practical considerations
Ungrouped frequency distributions work well for small to medium datasets and when precise values matter. For larger datasets with many distinct values, grouped tables are better to summarise and to construct histograms.

Why group data?
When there are many observations or when measurements are continuous and take many different values, listing each value is not practical. Grouping data into class intervals simplifies representation and makes patterns more visible. Grouped data are used to create histograms, frequency polygons and for estimating measures such as mean and median when raw data are not available.

Forming class intervals
Decide the number of classes depending on data size — for small samples 5–7 classes may be enough, while larger samples may need 10–15 classes. Calculate class width as (max − min)/number of classes and round to a convenient integer if data are in whole numbers. Try to use equal class widths to simplify interpretation and graphing. Choose lower and upper limits so classes are contiguous and mutually exclusive; no observation should fall into two classes.

Types of class limits and boundaries
There are two ways to show class limits: inclusive integer limits (e.g., 10–19) and continuous limits or class boundaries (e.g., 9.5–19.5). For measurement data which are continuous, use boundaries to avoid ambiguity at class borders. For discrete integer data, inclusive integer limits are simpler.

Class mark (mid-point)
Class mark is the average of lower and upper class limits and serves as a representative value for that class in calculations such as grouped mean and grouped variance. Compute it as (lower + upper)/2. Using class marks introduces approximation because individual values are replaced by the mid-point.

Frequency density for unequal widths
If classes have unequal widths, represent frequency by area in histograms so that area, not height, shows frequency. In such cases, compute frequency density = frequency ÷ class width and use density as bar height. For equal widths, simple frequencies can be used as heights directly.

Frequency table columns
A thorough grouped frequency table typically includes columns: class interval, class boundaries (if used), class mark, frequency, cumulative frequency, relative frequency and sometimes frequency density. This full table aids different analyses such as building ogives (using cumulative frequency) or histograms.

Loss of information and caution
Grouping reduces detail; two different raw datasets can produce the same grouped table. Keep class widths small enough to preserve important features, and state that grouped statistics are estimates. When presenting results, mention class width and method of grouping so readers can understand approximations.

Understanding the mean
The arithmetic mean is the familiar average. It gives a single number that summarises the centre of a distribution by balancing all observations. For ungrouped data the mean is the sum of all values divided by the number of values. It is useful when all values are numeric and especially when the distribution is roughly symmetric without extreme outliers.

Computation of mean
For n observations x1, x2, …, xn, calculate x̄ = (x1 + x2 + … + xn)/n. When values repeat, use frequencies: list each distinct value x_i with its frequency f_i and compute x̄ = Σ(x_i f_i)/Σf_i. This method speeds calculations and reduces repeated addition. Always include units in the mean (e.g., cm, marks) when presenting results.

Short-cut (assumed mean) method
For arithmetic ease, pick an assumed mean A close to the central values. Let d_i = x_i − A. Then x̄ = A + Σ(f_i d_i)/Σf_i. Because d_i are smaller numbers, multiplication and addition are easier and there is less chance of arithmetic error. This trick is especially helpful in examinations when class marks or values are large.

Properties of mean
The mean uses all data values and is unique for a given dataset. It lies between the minimum and maximum value. For symmetric distributions the mean equals the median. The mean is additive: the mean of combined groups can be computed from group means and sizes. However, the mean is sensitive to outliers; one very large or very small value can shift the mean significantly.

When to prefer mean
Use mean when measurements are on an interval or ratio scale and when there are no extreme outliers. Mean is also the basis for further measures such as variance and standard deviation. It is the most commonly reported average when data are fairly symmetric.

Limitations and checks
If data are skewed or contain outliers consider using median instead. After computing mean, check whether it is within expected range and whether it lies near central frequency classes. If mean seems odd compared to most data points, recheck calculations and watch for data entry errors.

Median: definition and method
The median is the middle value of ordered data and is a robust measure of central tendency. To find the median, arrange data in ascending order. If n is odd, the median is the value at position (n+1)/2. If n is even, the median is the average of values at positions n/2 and (n/2)+1. For example, for 5 numbers the median is the 3rd value; for 6 numbers the median is average of 3rd and 4th values.

Median with frequency tables
When data are presented with frequencies, find the cumulative frequencies and locate the position of the middle observation, that is (N+1)/2 for ungrouped lists or N/2 in some textbook conventions; use the cumulative frequency that first reaches or exceeds the median position and identify the corresponding value or class.

Why median is useful
Median is not affected by extreme values. In skewed distributions or when outliers are present (for example incomes where a few are very high), median often provides a better idea of a typical value than mean. Median is also suitable for ordinal data where precise numeric differences are not defined.

Mode: definition and interpretation
Mode is the most frequent value in a dataset. It is simple to find from a frequency table: the value with the highest frequency is the mode. A distribution can be unimodal (one mode), bimodal (two modes) or multimodal. Mode is the only measure of central tendency that can be used for nominal data, where numerical order is meaningless.

Mode in grouped data
For grouped data estimate the modal class as the class with highest frequency. Then a formula can give an approximate modal value within that class by interpolating, especially if neighbouring class frequencies differ. The grouped modal estimate is useful for continuous measurements summarised by classes.

Comparison of median and mode
Median gives a centre by position, mode gives the most common value. For symmetric distributions mean = median = mode. When skews occur mean moves toward the long tail, median shifts less and mode remains at the peak. Report more than one measure if possible to give a fuller picture of the data.

Practical tips
In exam answers show the ordered list and positions when giving the median. For mode, state whether distribution is unimodal or otherwise. When data are grouped, mention that median and mode are estimates and show the interpolation formula if used.

Estimating mean from grouped data
Grouped data hides individual values but gives class intervals and frequencies. To estimate the mean, use the class mid-points (class marks) as representative values. If x_i are class marks and f_i their frequencies then grouped mean x̄ ≈ Σ(f_i x_i)/Σf_i. This is an approximation that improves when class widths are small and data points are evenly distributed within classes.

Direct method step-by-step
1. Compute class mid-point for each class: x_i = (lower limit + upper limit)/2. 2. Multiply each x_i by its class frequency f_i. 3. Add these products to get Σ(f_i x_i). 4. Divide by total frequency Σf_i to get the grouped mean. Show these steps in exams and label units clearly.

Assumed mean (short-cut) method
To reduce calculation work choose an assumed mean A equal to a convenient class mark, often that of the central or largest class. Compute d_i = x_i − A and then Σ(f_i d_i). The grouped mean is x̄ = A + Σ(f_i d_i)/Σf_i. Because d_i are smaller than x_i, multiplications and sums are easier and errors less likely.

Step-deviation method for equal class widths
When class widths are equal with width h, define u_i = (x_i − A)/h which are often small integers. Then Σ(f_i x_i) = Σf_i [A + h u_i] = AΣf_i + hΣ(f_i u_i). So x̄ = A + h[Σ(f_i u_i)/Σf_i]. This method simplifies arithmetic further and is especially useful for many classes with large mid-points.

Accuracy and interpretation
Remember grouped mean is an estimate; the true mean would require raw data. Choose class widths and boundaries thoughtfully: wider classes increase approximation error. When presenting results state that you used class mid-points and indicate the degree of rounding.

Worked approaches and checks
It is good practice to compute grouped mean by both direct and assumed mean methods for verification. Also check the mean lies within the range and near the class with high frequency. These checks help catch arithmetic mistakes.

Range as a simple measure of spread
Range is the difference between the maximum and minimum observations. It is the simplest measure of spread and gives a quick sense of how widely values are distributed. However, range depends only on two values and is sensitive to outliers; a single extreme value can make the range large even if most data are close together.

Understanding quartiles
Quartiles split ordered data into four equal parts. The first quartile Q1 is the value below which 25% of observations lie, the second quartile Q2 is the median (50th percentile), and the third quartile Q3 is the value below which 75% of observations lie. Quartiles are useful because they describe the spread of the central portion of the data and are less influenced by extreme values than the range.

How to find quartiles for ungrouped data
Arrange data in ascending order and find the median to split the data into lower and upper halves. If n is odd, exclude the median when forming halves; if n is even, split the list evenly. The median of the lower half is Q1 and the median of the upper half is Q3. Some conventions differ slightly on inclusion of the median for odd n—use the convention taught in your class and state it in answers.

Interquartile range (IQR)
IQR = Q3 − Q1. It measures the spread of the middle 50% of data and is robust against outliers. A small IQR indicates data are clustered near the median; a large IQR shows the central half is more spread out. IQR is often used to detect outliers: observations lying more than 1.5×IQR below Q1 or above Q3 are commonly considered outliers in many analyses.

Estimating quartiles from grouped data
For grouped data use interpolation within the quartile class. Compute positions kN/4 where k = 1 for Q1 and k = 3 for Q3, with N total frequency. Find the class where the cumulative frequency first equals or exceeds this position. Use the formula Q_k ≈ L + [(kN/4 − cf) / f] * h where L is lower class boundary, cf is cumulative frequency before the quartile class, f its frequency and h class width. This estimates quartile values within the class.

Use in boxplots
IQR and quartiles are the basis of box-and-whisker plots which display minimum, Q1, median, Q3 and maximum. Boxplots visually compare central tendency and spread between datasets and highlight outliers. They are very handy for quick comparisons between groups.

Practical advice
When reporting quartiles mention method used and whether interpolation was applied for grouped data. Use IQR together with median when summarising skewed data because it gives a more robust picture than mean and standard deviation alone.

Definition and motivation
Mean deviation (also called average deviation) measures the typical distance of observations from a chosen centre, usually the mean or median. It is calculated as the average of absolute deviations, which avoids the cancellation that occurs when simple deviations (which can be positive or negative) are summed. Mean deviation gives an intuitive measure of spread in the same units as the original data.

Formula for ungrouped data
For data x1, x2, …, xn and chosen centre c (often the mean x̄ or the median), mean deviation MD = (1/n) Σ|x_i − c|. Using absolute values ensures each term contributes positively and we obtain an average distance.

Choice of centre
Mean deviation about the median is often smaller than that about the mean for skewed distributions because the median minimises the sum of absolute deviations. In symmetrical distributions the MD about mean and median are similar. Always state which centre you used when reporting MD.

Grouped data calculation
For grouped data replace individual values by class mid-points x_i and use MD = Σ[f_i * |x_i − c|] / Σf_i where f_i are class frequencies and c is chosen centre (grouped mean or grouped median). Compute absolute deviations for each class mark, multiply by frequency and divide by total frequency. This gives an estimate of MD for the grouped dataset.

Comparison with variance and standard deviation
Mean deviation uses absolute values; variance uses squared deviations. Squaring gives greater weight to larger deviations and is algebraically convenient for theoretical work, while absolute deviations are easier to understand and interpret. Standard deviation is more widely used in advanced work because it has desirable mathematical properties, but mean deviation remains useful for simple interpretation and when robustness to squaring is wanted.

Practical calculation tips
When computing MD, choose a centre near the bulk of data to reduce arithmetic. For grouped data use class mid-points consistently. Show intermediate steps in exams: table of class marks, deviations, absolute deviations, products with frequencies and final summation. Round the final answer to the required precision and include units.

When to use MD
Use MD to give an understandable average distance from the centre, particularly when communicating results to non-technical audiences or when you want a robust measure less influenced by squaring. It complements median and IQR in descriptive statistics.

Introduction to variance and why squaring is used
Variance and standard deviation are standard measures of spread. To obtain a meaningful summary of how data deviate from the mean, we square deviations to avoid cancellation of positive and negative differences. Squaring also gives larger deviations greater weight, which can be useful when large departures from the mean are important to detect.

Definitions for complete data
For n observations x1, x2, …, xn with arithmetic mean x̄, the population variance is defined as σ^2 = (1/n) Σ(x_i − x̄)^2 and the standard deviation is σ = sqrt(σ^2). Standard deviation has the same unit as the original data, making it directly interpretable as an average spread around the mean.

Computational formula
A practical formula reduces arithmetic: σ^2 = [Σx_i^2 / n] − x̄^2. This avoids computing each deviation and squaring it separately when Σx_i and Σx_i^2 are available. Use this formula carefully to avoid rounding errors; compute sums with enough precision.

Properties and interpretation
Standard deviation indicates typical distance of observations from the mean in the squared-weighted sense. Smaller σ means data are closely clustered around the mean; larger σ means they are more dispersed. In a symmetric bell-shaped distribution, about 68% of observations lie within one σ of the mean (this idea is for normal distributions taught later), so σ is a key descriptive measure.

Limitations
Because deviations are squared, outliers have amplified effect on variance and standard deviation. A single extreme value can greatly increase σ. For skewed data or when outliers are present, complement σ with median and IQR for a fuller picture.

Worked calculation and exam technique
In exams show steps: compute mean, compute deviations or use computational formula, calculate σ^2 and take square root for σ. State units and round final answers appropriately. If asked, compare σ with range and IQR to comment on spread. For practice, compute σ using both direct and computational formula to build confidence.

When to use population formula
In class 9 we usually treat given data as a complete set and use the population formula with divisor n. The sample formula with divisor (n−1) appears in higher classes when making statistical inferences from samples to populations.

Estimating variance from grouped data
When data are grouped into class intervals you do not know each exact observation, so variance and standard deviation must be estimated. Replace values in a class by the class mark (mid-point) x_i and use frequencies f_i. Then grouped variance is estimated by σ^2 ≈ Σ[f_i * (x_i − x̄)^2] / Σf_i, where x̄ is the grouped mean. Standard deviation σ is the square root of this estimated variance.

Steps to compute grouped variance
1. Compute class marks x_i. 2. Calculate grouped mean x̄ as Σ(f_i x_i)/Σf_i. 3. For each class compute (x_i − x̄)^2 and multiply by f_i. 4. Sum these products and divide by total frequency to get σ^2. 5. Take the square root for σ. Present your working clearly in columns to avoid errors.

Use of assumed mean and step-deviation methods
To ease arithmetic choose an assumed mean A and compute d_i = x_i − A or u_i = (x_i − A)/h for equal class widths h. Then use transformed formulas to compute variance with smaller numbers: compute Σf_i u_i and Σf_i u_i^2. The grouped variance can then be expressed as σ^2 = h^2 * [Σf_i(u_i^2)/Σf_i − (Σf_i u_i/Σf_i)^2], which reduces calculation of large squares.

Unequal class widths
If classes have unequal widths still use class marks for variance estimation. Frequency density affects histograms but not the variance formula if class marks and correct frequencies are used. Be cautious because unequal widths can increase estimation error compared to equal narrow classes.

Interpretation and accuracy
Grouped variance is an approximation; its accuracy improves with narrower classes and even distribution of values within classes. Compare grouped σ with range and IQR to check consistency. If grouped σ is surprisingly large or small, recheck class marks and computations and consider whether grouping introduced distortion.

Presentation and exam tips
In answers state that you used class mid-points and indicate any assumptions. Show intermediate sums Σf_i, Σf_i x_i, Σf_i x_i^2 or their equivalents in step-deviation form. Round final standard deviation sensibly and include units. If permitted, mention that grouped statistics approximate true values.

Concept of cumulative frequency
Cumulative frequency (CF) at a class gives the number of observations up to and including that class. It is built by adding class frequencies successively. CF tells us how many data points fall below a certain boundary and is essential for reading percentiles, medians and quartiles from grouped data without resorting to raw lists.

Constructing an ogive
An ogive is a graph of cumulative frequency against class boundaries (usually the upper class boundaries). To make an ogive compute CF at each class upper limit, plot points (upper limit, CF) on graph paper, including a starting point at the lower boundary with CF=0 for continuous data. Join the plotted points with straight lines to form the cumulative curve. The ogive rises monotonically and shows accumulation of observations across classes.

Using the ogive to read median and quartiles
To find the median draw a horizontal line at N/2 (N total frequency) across the ogive. Where this horizontal line meets the curve, drop a vertical to the x-axis; the x-coordinate is the estimated median. For quartiles draw horizontal lines at N/4 and 3N/4 and read off Q1 and Q3 respectively. This gives an interpolated estimate within a class for grouped data and is a practical alternative to algebraic interpolation.

Interpolation formula
If you prefer algebraic interpolation rather than reading a graph, use the grouped median formula: Median = L + [(N/2 − cf)/f] * h where L is lower boundary of the median class, cf cumulative frequency before that class, f class frequency and h class width. Similarly use Q_k = L + [(kN/4 − cf)/f] * h for quartiles. Both methods produce similar estimates; the ogive gives a visual check.

Practical tips for drawing
Use correct class boundaries (for continuous data use boundaries like 9.5, 19.5). Choose a scale so the curve fits the graph neatly and label axes and units. When reading values from the ogive use a ruler for horizontal and vertical lines to improve accuracy. If classes have unequal widths still plot CF at upper boundaries; the curve still accumulates frequency correctly.

Interpreting the ogive
Ogives help locate median and percentiles quickly and show skewness: a steep rise in early classes means many low values; a long tail at the right indicates positive skew. Use ogives when you need to report approximate percentiles or when comparing distributions by overlaying their cumulative curves.

Histograms explained
A histogram is a graphical representation of grouped continuous data. The horizontal axis shows class intervals and the vertical axis shows frequency or frequency density. For equal class widths draw adjacent (touching) bars whose heights equal class frequencies; the area of each bar is proportional to the class frequency. For unequal widths use frequency density (frequency ÷ class width) as the bar height so that area, not height, reflects frequency. Histograms are powerful for visualising the shape of a distribution — whether it is symmetric, skewed, uniform, or has multiple peaks.

How to draw a histogram
1. Decide class intervals and their boundaries. 2. Choose an appropriate scale for the vertical axis so the largest bar fits on the graph. 3. For each class draw a rectangle whose base is the class interval and whose height is frequency (equal widths) or frequency density (unequal widths). 4. Ensure there are no gaps between bars for continuous data. 5. Label the axes, include units and add a title.

Frequency polygon
A frequency polygon is formed by plotting points at class mid-points with heights equal to class frequencies (or densities) and joining them with straight lines. To close the polygon add an extra mid-point before the first class and after the last class with frequency zero. Frequency polygons emphasise the trend or shape of distribution and are convenient for comparing multiple distributions on the same axes because lines are easier to overlay and read than bars.

When to use histogram or polygon
Use histograms when you want to show detailed distribution and the area-proportional bars make the frequency differences visually clear. Use frequency polygons to compare two or more distributions on the same plot or to see smooth trends. A frequency polygon is also useful when you want to estimate the mode visually by locating the highest point.

Reading shape, centre and spread
From histograms and polygons you can visually estimate central tendency (peak or centre of the highest bars), spread (width of the distribution), skewness (direction of long tail) and modality (number of peaks). For example, a right-skewed histogram has a long tail to the right and the mean typically lies to the right of the median.

Accuracy and good practice
Choose sensible class widths: too wide classes hide details while too narrow classes produce noisy histograms. For unequal width classes ensure bar areas are correctly scaled. Always label axes, mention class boundaries and avoid 3D effects that distort perception. In exams show clear working and mark mid-points when constructing frequency polygons.

Bar graphs for categorical data
Bar graphs show comparisons between distinct categories. Place categories along the horizontal axis and frequencies or counts along the vertical axis. Bars should be of equal width and separated by gaps to indicate that categories are distinct and not continuous. Use bar graphs for nominal and ordinal data like favourite subjects, modes of transport or rating categories. Bar graphs make it easy to compare category sizes visually and to identify the most or least frequent categories.

Designing clear bar graphs
Label each bar with the category name and consider adding frequency labels on top of bars for clarity. Choose a linear scale on the vertical axis that begins at zero to avoid misleading impressions. Use consistent colours and a legend if multiple series are shown. For stacked bars show component parts but be careful as stacked bars make comparisons of individual categories across groups harder.

Pie charts for showing proportions
Pie charts display how a whole is divided among categories. The circle represents the total and each slice (sector) shows a category’s proportion. The central angle for a category equals (frequency ÷ total frequency) × 360°. Pie charts are best when there are only a few categories and when the aim is to show composition rather than precise comparison.

Constructing a pie chart
1. Compute total frequency N. 2. For each category compute fraction f_i/N and multiply by 360° for angle. 3. Using a protractor draw the sectors in order around the circle, labelling each sector with category name and percentage. Include a legend if space is limited. Maintain accurate angles for correct representation.

When to prefer one over the other
Use bar graphs when comparing categories or when values need to be compared precisely. Use pie charts to show how a total is divided when there are few categories and when the viewer should focus on proportions. Avoid pie charts with many small slices; they become difficult to read. Also avoid using 3D effects which can distort angle perception.

Practical presentation tips
Always include a title, and label axes or provide a clear legend. Choose colours and shading to make graphs accessible; ensure printed or photocopied versions remain clear in black-and-white. For exams draw neat bars using a ruler and use a protractor accurately for pie charts to obtain full marks.

What is a stem-and-leaf plot?
A stem-and-leaf display organises numerical data by splitting each value into a stem (leading part) and a leaf (usually the last digit). It preserves the individual data values while showing the distribution at a glance. This method is helpful for moderately sized datasets where exact values matter and you want both order and frequency visible.

Constructing the display
Decide how to split numbers: for two-digit numbers use the tens digit as the stem and the unit digit as the leaf. List stems in a vertical column in ascending order. For each data value, write the leaf next to its stem, arranging leaves in ascending order. Include a key such as 12 → 1 | 2 to clarify the splitting convention. The display is quick to make by hand and easy to read.

Advantages over grouped tables
Stem-and-leaf plots retain original data values, unlike grouped frequency tables which replace values by class mid-points. This makes stem-and-leaf useful when you need exact values and a visual summary together. You can quickly read medians, modes and spot clusters or gaps. It is especially useful in classroom settings for small surveys and exercises where students want to see both raw data and distribution.

Deriving frequency tables
From the stem-and-leaf display you can count leaves beside each stem to produce a simple frequency table for the corresponding class intervals. This is useful when you later need grouped summaries or graphs; stem-and-leaf is therefore a bridge between raw lists and grouped tables.

Modifications and larger datasets
For larger datasets use split stems (e.g., 10s by 2) or change the leaf unit to tenths or hundreds to manage data range. For decimals choose stems as the integer part and leaves as first decimal digits. When numbers have many digits you may round or truncate to make the display practical but always state the method used.

Limitations
Stem-and-leaf is not suitable for very large datasets or for data with many digits unless modified. It is also less suitable when you require smooth visualisation like a histogram, though you can produce both from the same data. For exams, practise neat layout and include a clear key so markers can read your display easily.

Comparing central tendency
When comparing two or more datasets the first step is to compute measures of central tendency: mean, median and mode. Mean is appropriate for symmetric, numerical data without extreme outliers. Median is preferred for skewed distributions because it is resistant to extremes. Mode identifies the most common value and is useful when the most frequent category matters. Use at least two measures to get a fuller view; reporting both mean and median often clarifies whether a dataset is skewed.

Comparing spread
To compare variability use range, interquartile range (IQR) and standard deviation. Range gives a quick sense of extremes but is sensitive to outliers. IQR compares the spread of the middle 50% and is robust; standard deviation measures overall spread and is commonly used when distributions are roughly normal. Use boxplots to compare medians and IQRs visually or overlay frequency polygons/histograms for more detailed comparison.

Graphical comparisons
Place histograms with the same class intervals and scales side by side for visual comparison. Frequency polygons plotted on the same axes highlight differences in peaks and tails. Boxplots are excellent for comparing medians, IQRs and spotting outliers across groups. Ensure axes are the same scale for fair comparison and label graphs clearly.

Context and sample size
Always consider sample sizes: small samples give more variable summary measures and larger samples provide more reliable estimates. Also check whether samples come from comparable populations — for example comparing test scores across different exam papers may be invalid if paper difficulty differs. Mention sample sizes in your conclusions and be cautious making broad claims from small samples.

Common pitfalls
Comparing averages without considering spread can mislead. For instance, two groups may have the same mean but different spreads; one may have many extreme values whereas the other may be consistent. Beware of aggregated comparisons that hide subgroup trends (Simpson’s paradox). Always inspect subgroup tables when aggregate results are surprising.

Communicating results
When reporting comparisons, present both numerical summaries and graphs. State which measures were used and why, and comment on practical significance, not just numerical difference. For example say "Class A mean marks are 5 points higher than Class B, but Class B shows less variability" rather than simply presenting numbers without context.

Practical uses of statistics
Statistics are used to summarise and interpret data in many areas: education (exam scores), science (experimental results), health (disease rates), economics (prices, incomes), government (census data) and everyday life (weather averages). In school, statistics helps with projects and experiments where you need to present evidence and draw reasoned conclusions. Proper use of statistics turns raw numbers into information that supports decisions.

Sampling and inference
Often you study a sample and use its statistics to infer something about the whole population. The reliability of such inference depends on how the sample was chosen. Random, representative samples reduce bias, whereas convenience samples or voluntary responses can give misleading results. Mention the sampling method and sample size when interpreting results to make clear limitations.

Distinguishing correlation from causation
Two variables may show a relationship (correlation) but that does not prove that one causes the other. For example, ice-cream sales and drowning incidents may both rise in summer but buying ice-cream does not cause drownings; temperature is a common cause. Always look for possible confounding factors and avoid jumping from correlation to causal claims without additional evidence.

Graphical and numerical misrepresentation
Presentations can mislead. Common tactics include truncating axes to exaggerate differences, using inappropriate scales, or using pie charts with many slices that are hard to compare. Visual effects like 3D charts can distort perception. Always check axis scales, starting points and labels. Report percentages rather than absolute counts when appropriate to make fair comparisons across groups of different sizes.

Aggregate data pitfalls
Aggregated data can hide underlying subgroup patterns — a phenomenon known as Simpson’s paradox. For example, overall pass rates may favour one group but break down by subjects may show the opposite. When possible examine subgroup data before making broad conclusions and state any aggregation choices made.

Ethics and transparency
Honest reporting of data includes stating methods, rounding appropriately and disclosing limitations. Selective reporting, omitting inconvenient data or using misleading visuals is unethical. In projects and exams clearly document how data were collected, how classes were formed and which formulas were used so others can follow and check your work.

Final practical advice
Master computation and graphing skills, but also practise interpretation: always ask what the numbers mean in context and whether they answer the original question. Being critical of sources and methods will make your statistics useful and trustworthy.

Approach to solving problems
Start by carefully reading the question and identifying whether the data are grouped or ungrouped and which measures are required. Choose the appropriate method: direct computation for small ungrouped sets, grouped formulas and class marks for classed data, or assumed-mean/step-deviation methods to reduce arithmetic. Plan your steps: prepare a neat table, compute required sums (Σf, Σfx, Σfx^2), and then proceed to calculate mean, variance, median or other requested measures.

Show working clearly
Examiners look for method as well as final answers. Present tables with headings, show intermediate sums, and write formulas before substituting numbers. If you use the assumed mean or step-deviation method, state the chosen A and h and show how transformed values were computed. Clear layout earns marks even when final numerical errors occur.

Time-saving methods
Use computational formulas to avoid lengthy subtraction: for variance use Σx^2/n − x̄^2 when Σx^2 is easy to compute. Use an assumed mean to keep multiplications small. For medians in grouped data, apply the interpolation formula directly when graph plotting is slow, and for graphs use rulers and protractors to speed up drawing while keeping neatness.

Common mistakes to avoid
Watch for arithmetic errors in summation and sign mistakes with deviations. For grouped data always use correct class mid-points and boundaries, and ensure the total frequency matches the sample size. For histograms use continuous class boundaries and draw bars without gaps; for pie charts carefully calculate angles and use a protractor.

Checking answers
Perform simple checks: mean should lie between minimum and maximum and typically near the region with large frequency. Standard deviation should be less than or comparable to range; if σ seems larger than range recheck calculations. For grouped median ensure the estimated median class is correct by checking cumulative frequencies around the midpoint.

Practice and presentation
Practice a variety of problems under timed conditions to build speed and accuracy. Present numerical answers with appropriate units and rounding. If interpretation is required, write concise sentences linking the numerical result to the context, stating any assumptions or limitations. This clarity and precision will score well in examinations.