OTM Correlation & Regression| MSLR | CAF DEC 2023

OTM – Only This Much

CORRELATION &
REGRESSION

MATH, LR & STATS

CA FOUNDATION DEC 2023

CA. PRANAV POPAT

SESSION LINK:
https://youtube.com/live/NOlAl4AIME8

JOIN TELEGRAM CHANNEL FOR ALL UPDATES

AND NOTES:
https://telegram.me/learnwithpranav

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav

 OTM Correlation & Regression| MSLR | CAF DEC 2023

OTM: Correlation and Regression

Past Trends

Attempt Practical Theory Total

May 2018 1 6 8
Nov 2018 3 2 5
Jun 2019 5 1 6
Nov 2019 3 1 4
Nov 2020 0 3 3
Jan 2021 1 4 5
Jul 2021 5 1 6
Dec 2021 2 2 4
Jun 2022 2 4 6
Dec 2022 4 1 5
June 2023 4 1 5

Bivariate Data

• When data are collected on two variables simultaneously, they

are known as bivariate data
Definition
• and the corresponding frequency distribution, derived from it, is
known as Bivariate Frequency Distribution
• It is the frequency distribution of one variable (x or y) across the
Marginal Distribution other variable’s full range of values
• Number of Marginal Distribution = 2
• It is the frequency distribution of one variable (x or y) across a
particular sub-population of the other variable.
• No. of Conditional Distributions = m + n
Conditional Distribution
m = no. of class interval of x
n = no. of class interval of y
• Number of Cells = m × n

For a 4 x 7 classification of bivariate data, the maximum number of conditional

MTP Nov 21 distributions is:
a. 11 b. 28 c. 35 d. None
Ans: a

For a p x q bivariate frequency table, the maximum number of marginal

MTP Oct 21 distributions is
a. p b. p+q c. 1 d. 2
Ans: d

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav

 OTM Correlation & Regression| MSLR | CAF DEC 2023

Scatter Diagram

• It helps us to find Nature and Relative Strength of Correlation

Theory about Scatter • It is useful for Non-Linear Correlation also
Diagram • It cannot be used to determine value
• Diagrams are time taking

If the plotted points in a scatter diagram lie from upper left to lower right, then
PYQ Nov 20 correlation is
a. Positive b. Negative c. Zero d. None
Ans: b

If the plotted points in a scatter diagram lie from upper left to lower right, then
PYQ Nov 20 correlation is
a. Find the type of correlation
b. Identify whether variables correlated or not
c. Determine the linear or non-linear correlation
d. Find the numerical value of correlation coefficient
Ans: d

If the plotted point in a scatter diagram lie from lower left to upper right then
PYQ June 22 correction is:
a. Positive b. Negative c. Zero d. None
Ans: a

Karl Pearson’s Correlation Coefficient

Cov(x , y)
Formula rxy =
(σx  σy )
(x − x )(y − y ) xy
Formula of Covariance Cov(x ,y) = or − x. y
n n
Property 1 The Coefficient of Correlation is a unit-free measure
Property 2 Value lies from -1 to +1
Change of Origin No impact
Change of Scale No impact of value, but if change of scale of
Property 3
both variables are of different sign then sign of
r will also change

500
A relationship r2 = 1 − is not possible. This statement is
PYQ May 18 300
a. True b. False c. Both d. None
Ans: a

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav

 OTM Correlation & Regression| MSLR | CAF DEC 2023

If the correlation coefficient between variables X and Y is 0.5, then the correlation
PYQ Nov 18 coefficient between the variables 2x − 4 and 3 − 2y is
a. 1 b. 0.5 c. -0.5 d. 0
Ans: c

Given that
X -3 -3/2 0 3/2 3
PYQ Jun 19 Y 9 9/4 0 9/4 9
The Karl Pearson’s Correlation Coefficient is
a. Positive b. Zero c. Negative d. None
Ans: b

What is the correlation coefficient from the following data:

x 1 2 3 4 5
PYQ Nov 19
y 5 4 3 2 6
a. 0 b. -0.75 c. -0.85 d. 0.82
Ans: a

For the set of observations {(1,2),(2,5),(3,7),(4,8),(5,10)} the value of Karl

PYQ Jan 21 Pearson’s coefficient of correlation is approximately given by
a. 0.755 b. 0.655 c. 0.525 d. 0.985
Ans: d

The coefficient of correlation between x and y is 0.5, the covariance is 16 and

PYQ Jan 21 variance of x is 16, then SD of y is
a. 4 b. 8 c. 16 d. 64
Ans: b

If the covariance between two variables is 20 and the variance of one of the
variables is 16. What would be the variance of the other variable?
Set B – Q3
a. ( σ y )  25
2 b. More c. Less than d. More than
than 10 10 1.25
Ans: a

Spearman’s Rank Correlation Coefficient

• find the level of agreement (or disagreement) between two

judges so far as assessing a qualitative characteristic (attribute)
Usage
is concerned
• Use in case of ranks
6d2
rR = 1 − 2
Formula (Regular) n(n − 1)
d = difference in ranks
Spearman’s Rank Correlation Coefficient (in case of tied values)
Formula (In case of Tie) 6 ( d2 + A )
rR = 1 − here A is adjustment value
n(n2 − 1)

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav

 OTM Correlation & Regression| MSLR | CAF DEC 2023

 ( t3 − t )
A= where t = tie length (calculate t value for each of the ties)
12

Determine the Spearman’s rank correlation coefficient from the given data
PYQ Jun 19  d2 = 30 and n = 10
a. 0.82 b. 0.32 c. 0.40 d. None
Ans: a

Compute rank correlation coefficient between Eco and Stats Marks

ICAI Eco 80 56 50 48 50 62 60
Example Stats 90 75 75 65 65 50 65
a. 0.2053 b. 0.15 c. 0.40 d. None
Ans: b

While computing rank correlation coefficient between profits and investment for
10 years of a firm, the difference in rank for a year was taken as 7 instead of 5 by
mistake and the value of rank correlation coefficient was computed as 0.80. What
Set B Q12
would be the correct value of rank correlation coefficient after rectifying the
mistake?
a. 0.3 b. 0.945 c. 0.25 d. 0.28
Ans: b

Coefficient of Concurrent Deviations

A very quick, simple and casual method of finding correlation when we

Usage
are not serious about the magnitude of the two variables
 2c − m 
rc =    
Formula  m 
where c is number of concurrent deviations (same direction)
m is number of pairs compared (equals to n-1)

For 10 pairs of observations, number of concurrent deviations was found to be 4.

MTP Jun 22 What is the value of the coefficient of concurrent deviation?
a. 0.2 b. 1/3 c. -1/3 d. − 0.2
Ans: c

1
If concurrent coefficient is and number of concurrent deviations is 6. Find
PYQ Jun 22 3
the number of pairs of data?
a. 9 b. 8 c. 10 d. 11
Ans: c

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav

 OTM Correlation & Regression| MSLR | CAF DEC 2023

Regression Basics

Estimation of one variable for a given value of another variable on the

Meaning basis of an average mathematical relationship between the two
variables
• Estimation of Y when X is given
Requirements
• Estimation of X when Y is given
Perfect • When linear relationship exists between two
Correlation variables, correlation is perfect.
• Perfect Correlation is represented by a linear
equation and this equation can be used for
regression purpose directly.
General Points
• Same equation can be used in both ways
Imperfect • In case of imperfect correlation there is no
Correlation definite line and equation
• We will use method of least square to estimate
both regression lines
Estimation of Y • Use Regression line of Y on X
when X is given • Equation Format:
Y − Y = byx (X − X)
Formula of Regression byx is regression coefficient of Y on X
Equations/ Lines Estimation of X • Use Regression line of X on Y
when Y is given • Equation Format:
X − X = bxy (Y − Y)
bxy is regression coefficient of X on Y

Regression Coefficient SDy cov(x,y)

byx = r. and byx =
of Y on X
( SDx )
2
SDx
Regression Coefficient
Regression Coefficient SDx cov(x,y)
bxy = r. and bxy =
of X on Y
( SDy )
2
SDy

Change of Origin/ Scale for Regression Coefficients: Origin no impact,

Scale impact of both magnitude and sign.
change of scale of y
Property 1 bvu = byx 
change of scale of x
change of scale of x
buv = bxy 
change of scale of y
Two regression lines (if not identical) will intersect at the point [means]
Property 2
( x, y )
Correlation Coefficient is the GM of regression coefficients:
Property 3 rxy =  bxy  b yx
Note: rxy , bxy , byx all will have same sign

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav

 OTM Correlation & Regression| MSLR | CAF DEC 2023

If the two lines of regression are x + 2y − 5 = 0 and 2x + 3y − 8 = 0 then the regression

PYQ Nov 18
line of y on x is
PYQ Nov 19
a. x + 2y − 5 = 0 b. 2x + 3y − 8 = 0 c. Both d. None
Ans: a

PYQ Jul 21 If byx = -1.6 and bxy = -0.4, then rxy will be
PYQ Dec 22 a. 0.4 b. -0.8 c. 0.64 d. 0.8
Ans: b

If the two regression lines are 3x = y and 8y = 6x , then the value of correlation
PYQ Nov 18
coefficient is
a. 0.5 b. -0.5 c. 0.75 d. -0.80
Ans: a

If the regression line of y on x is given by y = x + 2 and Karl Pearson’s coefficient

of correlation is 0.5 then ( σ y / σx ) is
2
PYQ Jun 19
a. 9 b. 2 c. 4 d. 3
Ans: c

If the slope of regression line is calculated to be 5.5 and the intercept is 15 then
PYQ Jul 21 the value of Y is if x = 6
a. 88 b. 48 c. 18 d. 78
Ans: b

Probable Error

1 − r2
Formula Probable Error in correlation: 0.6745 
N
• Correlation is calculated using sample, value for sample may
differ from population, this difference is probable error
Use
• If there is significant probable error, there is no evidence of real
correlation
Limits of Sample
Correlation Coefficient r  PE

Case Conclusion
If r is less than PE There is no evidence of correlation
How to check evidence of
If r is greater than six The presence of correlation is
Correlation using PE
times of PE certain
Since r lies from -1 to +1 PE can never be negative

2
Find the probable error if r = and n = 36
PYQ Jun 19 10
a. 0.6745 b. 0.06745 c. 0.5287 d. None
Ans: b

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav

 OTM Correlation & Regression| MSLR | CAF DEC 2023

Coefficient of Determination and Non-Determination

Coefficient of Determination
(r )
2
Accounted Variance/ Explained Variance xy

1 − (r )
Coefficient of Non-Determination 2
Unaccounted Variance/ Unexplained Variance xy

If the two regression co-efficient are 4 and 0.16 the percentage of unexplained
MTP Nov 18 variation is:
a. 64 b. 36 c. 54 d. 46
Ans: b

If the coefficient of correlation between two variables is 0.7 then the percentage
MTP Nov 18 of variation accounted for is
a. 49% b. 30% c. 51% d. 36%
Ans: a

Dil Se Re ❤️ Instagram: @ca_pranav Telegram @learnwithpranav