WK 1 - BIOSTATS

Why are statistics important in science and health?

They allow you to understand, interpret, and analyse data meaningfully.

How do statistics help you understand real-world information?

They let you interpret data-based claims in news, politics, and health (e.g., comparing COVID vs flu deaths).

How do statistics help you avoid being misled?

They help identify misleading graphs, cherry-picked data, and biased samples.

Why do researchers need statistical understanding?

To judge if results are significant and interpret variability.

Why are statistics essential for scientists and health professionals?

To summarise, interpret, and present their own experimental results accurately.

What are the two main types of data?

Categorical (qualitative) and numerical (quantitative).

What is categorical data?

Data grouped by name or category, not measured numerically (e.g., blood type, gender).

What is numerical data?

Data that can be measured and expressed as numbers (e.g., age, height).

Typical graphs for categorical data?

Bar graphs, column graphs, pie charts.

Typical graphs for numerical data?

Histograms, boxplots, line graphs.

What is nominal data?

Categories with no logical order (e.g., eye colour, blood type).

What is ordinal data?

Categories that have a ranked order (e.g., grades, pain scale).

Can you calculate a mean for ordinal data?

No — you can rank it but not measure equal spacing.

How do you summarise categorical data?

With counts or percentages.

What is continuous data?

Data that can take any value within a range (e.g., height, weight, temperature).

What is discrete data?

Whole-number counts of items or events (e.g., number of children).

What is count data?

A subtype of discrete data representing the number of times something occurs (e.g., hospital visits).

What is rate or proportion data?

Measures relative to time or population (e.g., infection rate, growth rate).

What’s the key feature of ordinal data?

Order matters, but spacing between ranks doesn’t.

What’s the key feature of numerical data?

Both order and spacing are meaningful.

Example of ordinal data?

Grades: HD > D > C > P > F.

Example of numerical data?

Temperature, height, or age.

Can you calculate averages for ordinal data?

No, only ranks; means are meaningless.

What are the three measures of central tendency?

Mean, median, and mode.

When is the mean best used?

With symmetrical (normal) data and no outliers.

When is the median best used?

With skewed data or when outliers are present.

When is the mode useful?

For categorical data or data with multiple peaks.

Which measure is most affected by outliers?

The mean.

What does “spread” describe in data?

How variable or consistent the values are.

What is the range?

Maximum value minus minimum value.

What is the interquartile range (IQR)?

The range of the middle 50% of data (Q3 – Q1).

What is standard deviation (SD)?

The average distance of each point from the mean (used for normal data).

What percentage of data lies within \pm1 SD of the mean in a normal distribution?

About 68%.

What about \pm2 SD?

About 95%.

What type of data uses a histogram?

Numerical data (frequency distributions).

What type of data uses a column or bar graph?

Categorical data.

What does a line graph show?

Trends or changes over time.

What does a boxplot show?

Median, interquartile range, range (whiskers), and outliers.

What does a violin plot add to a boxplot?

The shape of the data distribution.

Why use column graphs with data points?

To show all individual values, improving transparency.

What does the median line in a boxplot show?

The middle value of the data.

What does the box in a boxplot represent?

The interquartile range (IQR), or middle 50% of data.

What do whiskers show?

The overall spread of most of the data.

What are outliers in a boxplot?

Points that fall more than 1.5\times IQR from Q1 or Q3.

What can box width sometimes indicate?

Sample size (wider = more data).

What does a normal (symmetrical) distribution look like?

Bell-shaped; mean = median = mode.

What does a positively skewed distribution look like?

Tail to the right; mean > median > mode.

What does a negatively skewed distribution look like?

Tail to the left; mean < median < mode.

Which measure best represents skewed data?

The median.

What is an outlier?

A data point far from the rest of the dataset.

What causes outliers?

Data entry error, measurement error, or true biological variation.

How can outliers affect statistics?

They distort the mean and standard deviation.

How should outliers be handled?

Check for errors; if valid, report them and use median/IQR for summary.

Skew direction = ?

Direction of the tail (right = positive, left = negative).

Which measure is resistant to outliers?

The median.

Histogram vs Column Graph?

Histogram = numerical data; Column = categorical data.

SD vs IQR?

SD for normal data, IQR for skewed/non-normal data.

What do boxplots display?

Median, IQR, whiskers, and outliers.

When to use non-parametric tests?

When data are skewed or contain outliers (non-normal).