Normal Probability Curve
According to -
J.P. Guilford - Normal probability curve is a well-defined, well-structured, mathematical curve having a distribution of the scores with mean, median and modes are equal.
The curve itself is a mathematical conception, it doesnot occur in nature, it is not a biological or psychological curve.
The graph of the probability density function of the normal distribution is a continuous bell shaped curve, symmetrical about the mean, and is called normal probability curve.
Graph of normal probability curve-
Characteristics of normal probability curve (NPC)
It is a unimodel curve, i.e., it has only one mode.
It is a bell shaped curve combining both concave and convex curves.
For this curve, mean, median and mode are the same.
The curve is perfectly symmetrical, i.e., left and right halves are the mirror images.
This curve is bilateral, i.e., 50% area of the curve lies to the left side and the other 50% lies to the right side.
The curve is asymptotic to the x-axis, i.e., it approaches but never touches the x-axis.
The maximum ordinate (y-coordinate) occurs at the centre and is equal to 0.3989.
The points of Influx (convex to concave) occur at point ±σ (±1 standard deviation unit).
The curve extends on both sides -3σ distance on the left to +3σ distance on the right.
Application of normal probability curve
It is used in behavioural researches because most of the variables of social sciences are normally distributed if a large sample or whole population is taken.
It is used for standardising tests.
Difficulty value of test items can also be compaired by normal probability curve.
It is also used in assigning grades.
Raw scores can be converted into standard scores.
It is also used for developing z-scores or norms of the test.
It is used to calculate the number of cases who achieve more or less than a given particular score.
It is used to calculate the number of cases who achieve within the given two limits of scores.
It is used to calculate the limits which include a given percentage of cases.
It is used to calculate the relative difficulty of test items.
It is used to divide a normal distribution into various categories on the basis of the ability level.
It is used to compare two normal distributions in teams of their mutual overlapping.
Divergence in normality
Generally two types of divergence occur in the normal curve, i.e., Skewness and Kurtosis.
1. Skewness -
According to -
H. C. Garrett - A distribution is said to be skewed when the mean and median fall at different points in the distribution and the balance (centre of gravity) is shifted to one side or the other to left or right.
There are two types of skewness which appear in the normal curve -
A. Negative skewness -
The scores are massed to the high end (right side) and are spread out more gradually towards the low end (left side) of the curve.
The value of median will be higher than that of the value of the mean.
E.g.- If question paper is easy, then most students gets good marks.
B. Positive Skewness -
The scores are massed to the low end (left side) and are spread out more gradually towards the high end (right side) of the curve.
The value of median will be lower than that of the value of the mean.
E.g.- If question paper is hard, then most students gets low marks.
2. Kurtosis -
The team kurtosis refers to the divergence in the height of the curve, specially in the peakness.
There are two types of divergence in the peakness of the curve.
A. Leptokurtosis -
The frequency curve is more peaked than to the normal distribution curve.
The value of Ku (vain of kurtosis) is less than 0.263 (Ku < 0.263)
B. Platykurtosis -
The frequency curve is more flatter than to the normal distribution curve.
The value of Ku is greater than 0.263 (Ku > 0.263).
C. Mesokantic -
This is a term used for the normal distribution curve.
The value of Ku is equal to the 0.263 (Ku = 0.263)
Factors causing divergence in the normal distribution curve
1. Selection of the sample
If the sample size of the individuals is small or sample is biased one, divergence is possible in the normal distribution curve.
Scores from small and homogeneous groups yield narrow and leptokurtic distribution.
Scores from small and highly heterogeneous groups yield platykurtic distribution.
2. Unsuitable or poorly made tests
If the measuring tool or test is unappropriate, or poorly made, the asymmetry is possible in the distribution of scores.
If a test is too easy or too hard, normal distribution curve will have divergence.
3. The trait being measured is non-normal
Divergence will appear when there is a real lack of normality in the traits being measured, e.g., interest, attitude, deaths in old age or early childhood due to certain degenerative diseases etc.
4. Errors in the construction and administration of tests
Factors like the following during administrating the test may cause asymmetry in the distribution of scores
Unclear instructions
Error in timing
Lack of practice
Lack of motivation to complete the test.
Construction of unstandardised and poor item-analysed test.
Measurement of divergence
Measurement of skewness
Sk = 3(M-Md)/σ
where, M = Mean
Md = median
σ = Standard Deviation
Sk = [(P90 - P10)/2] - P50
where, P = Percentile
Measurement of kirtosis
Ku = Q / (P90 - P10)
where, Q = Quartile Deviation
P = Percentile