Unit 6: The Normal Distribution - Subjective Questions

Definition: The normal distribution is a continuous probability distribution that is symmetric and bell-shaped.

Nature of the normal curve:

• It is unimodal, with one peak at the mean.
• The mean, median, and mode are equal.
• The curve is perfectly symmetrical around the mean.
• Most observations lie near the mean, while fewer observations occur at the extremes.
• The two tails extend indefinitely in both directions and approach, but never touch, the horizontal axis.
• The total area under the curve is equal to , representing total probability.

The normal curve is widely used as a mathematical model for psychological test scores and other naturally occurring measurements.

The normal probability curve has the following important characteristics:

• Symmetry: The two halves of the curve are mirror images of each other.
• Central tendency: The mean, median, and mode coincide at the center.
• Bell shape: The curve rises gradually, reaches a maximum at the mean, and falls gradually.
• Asymptotic tails: The tails approach the baseline but never intersect it.
• Unit area: The entire area under the curve is .
• Maximum height: The greatest height occurs at the mean.
• Points of inflection: The curve changes curvature at one standard deviation above and below the mean, that is, at and .
• Spread: The standard deviation determines the width of the curve; a larger standard deviation produces a flatter and wider curve.

The mean and standard deviation determine the location and spread of a normal curve.

• Mean (): It determines the center or location of the curve. Increasing the mean shifts the entire curve to the right, while decreasing it shifts the curve to the left.
• Standard deviation (): It determines the spread or dispersion of scores. A small standard deviation produces a tall and narrow curve. A large standard deviation produces a short and broad curve.
• Shape: Although different normal curves may have different means and standard deviations, all retain the same basic bell-shaped form.

The general normal density function is:

Thus, the mean controls position and the standard deviation controls variability.

The standard normal distribution is a normal distribution with:

• Mean equal to .
• Standard deviation equal to .
• A unit area of under the curve.

Any raw score can be converted into a standard score using:

where is the raw score, is the population mean, and is the standard deviation.

The -score indicates how many standard deviations a score lies above or below the mean. A positive -score represents a score above the mean, while a negative -score represents a score below the mean. Standard scores allow researchers to compare scores from different distributions and to determine areas and probabilities using the standard normal table.

The empirical rule describes the approximate percentage of observations within standard deviation units of the mean in a normal distribution:

• Approximately 68% of scores lie between and .
• Approximately 95% lie between and .
• Approximately 99.7% lie between and .

This is often called the 68–95–99.7 rule.

Significance:

• It helps psychologists interpret test scores.
• It provides a quick method for identifying unusually high or low scores.
• It helps estimate the percentage of individuals falling within a particular range.
• It supports the identification of extreme scores that may require further investigation.

The rule applies most accurately when the distribution is approximately normal.

A sampling distribution is the probability distribution of a statistic, such as a sample mean, calculated from all possible samples of a fixed size drawn from a population.

The normal curve is used as a model for sampling distributions because of the Central Limit Theorem. According to this theorem, when sufficiently large random samples are repeatedly drawn from a population, the distribution of sample means tends to become approximately normal, even when the original population is not perfectly normal.

For the sampling distribution of the mean:

• Its mean is approximately the population mean: .
• Its standard deviation is the standard error: .
• Its shape becomes increasingly normal as sample size increases.

This model allows researchers to estimate probabilities, construct confidence intervals, and test hypotheses about population parameters.

The Central Limit Theorem states that the sampling distribution of the mean approaches a normal distribution as the sample size becomes sufficiently large, regardless of the shape of the population distribution, provided that the observations are independent and come from the same population.

Its main implications are:

• The average of repeated samples tends to cluster around the population mean.
• The variability of sample means decreases as sample size increases.
• The standard error of the mean is given by:

• Larger samples produce more stable and reliable estimates.
• Normal probability methods can be used even when the population distribution is somewhat non-normal.

In psychology, the theorem supports the use of inferential statistics when researchers draw conclusions about a population from sample data.

Population distribution:

• Includes the scores of all individuals or observations in the population.
• Its parameters are usually represented by and .

Sample distribution:

• Includes the individual scores obtained from one particular sample.
• Its descriptive statistics are commonly represented by and .

Sampling distribution:

• Includes the values of a statistic calculated from many possible samples of the same size.
• It describes how a statistic varies from sample to sample.
• For sample means, its standard deviation is called the standard error.

The sampling distribution is especially important because it connects sample results with population parameters and forms the basis of statistical inference.

Suppose a population has standard deviation and independent observations are selected to form samples of size . The variance of the sample mean is:

Taking the square root gives the standard deviation of the sampling distribution of the mean:

This quantity is called the standard error of the mean.

Effect of sample size:

• Increasing decreases the standard error.
• A larger sample produces sample means that are more closely clustered around the population mean.
• To reduce the standard error by half, the sample size must be increased fourfold.
• A smaller standard error indicates greater precision in estimating the population mean.

Thus, sample size affects the spread of the sampling distribution but not its expected center.

Skewness is a measure of the degree and direction of asymmetry in a distribution.

• A distribution is symmetric when its left and right sides are mirror images.
• A positively skewed distribution has a longer tail toward higher values.
• A negatively skewed distribution has a longer tail toward lower values.
• Skewness describes the direction and extent to which observations deviate from symmetry.

For a perfectly normal distribution, the coefficient of skewness is approximately . Skewness affects the relationship among the mean, median, and mode and may influence the selection of statistical procedures.

Positive skewness:

• The distribution has a long tail extending toward larger values.
• A few unusually high scores pull the mean to the right.
• The usual relationship is:

Negative skewness:

• The distribution has a long tail extending toward smaller values.
• A few unusually low scores pull the mean to the left.
• The usual relationship is:

In a perfectly symmetric normal distribution, the mean, median, and mode are equal. The direction of skewness is determined by the direction of the longer tail, not by the location of the peak.

A normal distribution has skewness close to . Skewed distributions depart from the normal model and may require data transformation or non-parametric methods.

Identifying skewness is important because it reveals whether a distribution departs from the normal model.

• It helps researchers determine whether the mean is an appropriate measure of central tendency.
• In a highly skewed distribution, the median may represent the typical score better than the mean.
• Skewness can indicate floor effects, ceiling effects, or restricted measurement ranges.
• It affects the validity of statistical procedures that assume normality.
• It helps researchers detect unusual observations and possible data-entry errors.
• It guides the choice between parametric and non-parametric tests.

For example, reaction-time data are often positively skewed because a small number of participants may take exceptionally long to respond.

Kurtosis refers to the degree of peakedness or flatness of a distribution and, more generally, to the heaviness of its tails compared with a normal distribution.

It provides information about how concentrated scores are near the mean and how frequently extreme scores occur.

• A distribution with normal kurtosis resembles the normal curve.
• A leptokurtic distribution is more sharply peaked and has heavier tails.
• A platykurtic distribution is flatter and has lighter tails.

Kurtosis is often expressed using excess kurtosis, for which the normal distribution has a value of . Using this convention, positive values indicate leptokurtosis and negative values indicate platykurtosis.

Mesokurtic distribution:

• Has a shape and tail weight similar to the normal distribution.
• Its excess kurtosis is approximately .

Leptokurtic distribution:

• Has a relatively sharp central peak.
• Possesses heavier tails than the normal distribution.
• Contains a greater tendency for extreme scores.
• Its excess kurtosis is positive.

Platykurtic distribution:

• Has a flatter and broader central region.
• Possesses lighter tails than the normal distribution.
• Contains fewer extreme scores than a normal distribution.
• Its excess kurtosis is negative.

Kurtosis should be interpreted together with graphical displays and other descriptive statistics rather than as the only test of normality.

Skewness and kurtosis describe different aspects of a distribution.

• Skewness measures asymmetry. It indicates whether one tail is longer or heavier than the other and identifies the direction of departure from symmetry.
• Kurtosis measures tail weight and the relative concentration of scores around the center. It indicates whether a distribution is more peaked or flatter than the normal curve.

For a normal distribution:

• Skewness is approximately .
• Excess kurtosis is approximately .

A distribution may have zero skewness but differ in kurtosis, or it may have normal kurtosis but be skewed. Therefore, the two measures provide separate information about distributional shape.

The normal curve represents probabilities through areas under its graph. Since the total probability of all possible outcomes is , the total area under the curve is or .

The mean divides the curve into two equal areas:

• of the area lies below the mean.
• lies above the mean.

The standard deviation divides the curve into meaningful intervals. In a normal distribution:

• The area between and is about .
• The area between and is about .
• The area between and is about .

By converting raw scores to -scores, researchers can use areas from the standard normal table to determine probabilities and percentile ranks.

A raw score is expressed relative to the mean by subtracting the mean and then scaling the difference by the standard deviation:

The derivation is interpreted as follows:

• gives the deviation of the score from the mean.
• Dividing by expresses that deviation in standard deviation units.
• The resulting -score has a mean of and a standard deviation of .

To locate a score:

• If , the score is at the mean.
• If , the score is standard deviations above the mean.
• If , the score is standard deviations below the mean.

The -score permits comparison of scores from different tests and allows the use of standard normal probability tables.

The normal curve is an appropriate model for a sampling distribution when several conditions are reasonably satisfied:

• Random sampling: The sample should be selected in a way that represents the population.
• Independence: Observations should not influence one another.
• Adequate sample size: Larger samples make the sampling distribution more nearly normal, especially when the population is skewed.
• Finite variance: The population should have a meaningful and finite variance.
• Appropriate statistic: The theorem applies particularly well to means and sums, and with suitable conditions to other statistics.
• No severe outliers: Extreme observations can delay the approach to normality and distort estimates.

When these conditions are met, researchers can use normal-theory methods to estimate sampling probabilities and conduct statistical tests.

The normal curve is useful in psychological measurement because many characteristics, such as ability, aptitude, and some personality measures, are approximately normally distributed in large populations.

Its usefulness includes:

• Interpretation of scores: Individual scores can be compared with the group mean.
• Percentile estimation: Areas under the curve indicate the percentage of people below or above a score.
• Standardization: Raw scores can be converted to -scores, -scores, or other standard scores.
• Identification of unusual scores: Scores several standard deviations from the mean may be considered uncommon.
• Test construction: Norm-referenced tests often use a normal model to establish reference groups.
• Statistical inference: Many psychological procedures rely on approximately normal sampling distributions.

However, researchers must check whether the observed data actually approximate normality before applying normal-theory procedures.