Standard Scores Comprehensive Notes

Standard Scores

Standard scores are valuable for understanding individual performance relative to the mean.
Raw scores provide limited context, especially in standardized tests where performance is evaluated against others.

Teachers need to know where students stand: at, below, or above the mean.
Standard scores facilitate informed discussions about student performance.
They represent performance relative to the mean and standard deviation of a normal distribution.

Norming involves testing a large, representative sample of individuals.
Intelligence tests, for example, are often normed within specific age groups.
The distribution of scores typically follows a normal distribution, with most scores near the mean.
Norming is a lengthy and costly process, periodically repeated for tests like the ACT and SAT to adjust for score inflation over time.
Norms change over time in industrialized nations due to various factors, causing means to gradually increase.

Raw scores have limited utility.
Common types include z-scores, T-scores, and stanines.
Z-scores and T-scores are frequently encountered in hypothesis testing and regression analysis.
Standard scores convert raw scores to represent overall performance relative to the mean and standard deviation.

A z-score indicates how many standard deviations a person is above or below the mean.
Z-score distributions have a mean of 0 and a standard deviation of 1.
A z-score of +1 means the person is one standard deviation above the mean.
Raw scores can be easily converted to z-scores.
Z-scores facilitate the determination of percentile ranks.

The z-score is calculated using the following formula:

Zi = \frac{Xi - \mu}{\sigma}

Where:

A z-score of 0 indicates the person scored at the mean.
A positive z-score indicates the person scored above the mean.
A negative z-score indicates the person scored below the mean.
The magnitude of the z-score indicates the distance from the mean in standard deviations.

If a person scores at the mean, their z-score is 0.
If a person scores above the mean(8) with mean(5) and standard deviation(3), their z-score is +1.
If a person scores below the mean(4) with mean(5) and standard deviation(3), their z-score is -1.

Z-scores can be expressed in decimal units.
A z-score of 0.333 means the person is one-third of a standard deviation above the mean.

Z-scores allow for easy plotting of individuals on a normal distribution.
Given the mean and standard deviation, you can calculate z-scores and plot individuals accordingly.

SPSS can automate the calculation of z-scores for a distribution of scores.
By selecting "save standardized values as variables", SPSS will compute z-scores for each individual score.
SPSS calculates z-scores by comparing each individual score to the mean and standard deviation of the distribution.

Using a z-score chart, you can determine the percentile rank associated with a given z-score.

T-scores are similar to z-scores but eliminate negative numbers.
They have a different scale but convey the same information about relative performance.

Stanines break the distribution into nine parts.
A stanine of 5 is equivalent to a z-score of 0 and a T-score of 50, representing the 50th percentile.

Percentile ranks can be derived from standard deviation units.
A score of one standard deviation above the mean corresponds to approximately the 84th percentile.
A T-score of 40 means that approximately 16% of participants scored below that score.

Standard scores can be created using any mean and standard deviation unit.
As long as the distribution is normal and the mean and standard deviation are known, scores can be easily interpreted.

To calculate the z-score, you need the individual's score, the mean, and the standard deviation.
Using z-score charts, you can determine the percentage of scores superseded by a given score.