In the context of handling missing data, 'imputation' refers to the process of filling in missing data with substituted values. These values can be determined in a variety of ways such as using measures of central tendency (mean, median, mode), predictive models, or other techniques.

A correlation coefficient close to -1 implies that the two variables have a strong negative relationship. This means that as one variable increases, the other decreases and vice versa.

The difference between skewness and kurtosis is that skewness measures the asymmetry of a data distribution around its mean, whereas kurtosis measures the "tailedness" of a data distribution. So, skewness is about the symmetry, and kurtosis is about the tails of the distribution.

In a box plot, data points plotted as circles often represent outliers.

The key difference between 'removal' and 'transformation' of outliers is that removal discards outliers from the dataset, while transformation modifies the values of outliers to reduce their impact.

In statistics, an outlier is an observation that lies an abnormal distance from other values in a random sample from a population. In simple terms, an outlier is a value that is significantly different from other similar values.

Replacing missing values with the median of the existing values is known as 'median' imputation. This technique is useful for skewed distributions as the median is less affected by outliers than the mean.

This would also be NMAR (Not Missing at Random) because the missingness (income level) depends on the value of the unobserved data itself (i.e., people with higher or lower incomes may be more likely to omit this information).

Discrete data can only take certain values (usually integers) and there are gaps between the values.

When dealing with a large number of outliers in a data set, the "Interquartile range (IQR)" is the most suitable measure of dispersion. This is because it measures the statistical spread between the 25th and 75th percentiles, thus excluding outliers.