MODULE 1 2.

Inferential Statistics

Introduction  Branch of statistics performed after

analyzing the data at hand in order to
Statistics – a branch of mathematics that deals primarily
make sound conclusions and
with collection, presentation, analysis, and interpretation
generalizations.
of quantitative data to draw sound conclusions in phases
d. Interpretation
of uncertainty.
Understanding the meaning of the results
A. Input of the tests conducted on the data and
a. Data applying it on the dynamics of the subject
B. Process matter on hand.
a. Collect
Types of Data
b. Present
c. Analyze Apply proper statistical procedures to present the data
d. Interpret accurately.
C. Output
1. Qualitative Data (AKA Categorical Data)
a. Information – for effective decision
 Expresses qualities, attributes, or
Importance of Statistics responses.
 They are normally in nominal form where
1. Data are everywhere
mathematical operations cannot be used
2. Statistical techniques are used to make decisions
and answerable only by two responses
that affect our daily lives
(e.g. gender).
3. Knowledge of statistical methods will help you
 We can “quantify” such data by
understand how decisions are made and give a
transforming them into dummy data – that
better understanding on how it can affect an
is, giving a numeric value to their presence
individual.
or absence (e.g. 0 for male and 1 for
Types of Statistics female.)
2. Quantitative Data (AKA Numerical Data)
1. Descriptive Statistics
 Obtained from measurement
 Branch of statistics concerned with
 It expresses an amount using a specified
describing the attributes, behavior,
level or type of measurement.
properties, or characteristics of data on
A. Discrete Variable (whole number) – can
hand.
assume only certain values, and there are gaps
a. Collection
between values.
Primary/First-hand data
B. Continuous Variable (decimals) – can assume
1. Experiments
any value within a specific range.
2. Questionnaires
Secondary Data Levels of Measurement
Includes but not limited to historical data
1. Nominal Data
from proper agencies
 The simplest form of data.
b. Presentation
 No scales, no measurements involved as
“Organization of data”
these are just counts to differentiate
1. Textual
classes or categories and hence, no
2. Tabular
mathematical operation can be performed.
3. Graphical
c. Analysis  They are mutually exclusive (an individual
Describing properties or behavior of the or object is included only in one category)
data with the intent of extracting relative and exhaustive (individual, object or
information by using various Statistical measurement is included must appear in
Tools (Ex. Measures of Central Tendency, one category). (Ex. Country, Race, Hair, Color)
Parametric and Non-Parametric Tests) 2. Ordinal Data
  These are the ones used in ranking or  Sampling error is expected. The researcher should
levels that indicates superiority from one establish first the reliability of such before drawing
level to another. any conclusion or generalizations from the sample
 It indicates a position in a series or order in reference to the population.
(Ex: 1st, 2nd, 3rd. Severe, High, Elevated, Guarded,
Low) Sampling Procedures
3. Interval
1. Non-Probability Sampling
 Possesses the characteristic of ordinal
Definition: Subjects are chosen without regard to
numbers and differentiates between
their probability. It does not involve random
classes or categories, normally of equal
selection and no sampling frame.
unit of measurement between each score.
Used when: pilot studies or small-scale project
 no true zero and can represent values
market
below zero. (Ex. Temperature, IQ Scores,
Advantages: Convenient, speedy, less costly
Personality Scores, and the likes)
4. Ratio Disadvantages: Less representative, lack of
 The highest form of measurement. accuracy due to selection bias, cannot be readily
used to draw conclusions about the population
 It has the properties of interval
A. Accidental/Convenience Sampling (AKA
measurement plus two more properties:
Haphazard or “Man on the Street Interview”
a. it has a meaningful zero point – the
Sampling Technique)
complete absence of the data being
measured and;  based on availability, proximity, and
b. the ratio between two numbers is convenience of the researcher or by
meaningful. the use of volunteers (self-selected
samples)
(Ex. Weight, height, age) B. Purposive Sampling
Summary  based on the purpose of the study;
hence they are practically a “pre-
1. Nominal: the data can only be categorized defined group” who knows where,
2. Ordinal: the data can be categorized and ranked whom and how to go with the
3. Interval: the data can be categorized, ranked, and research goals.
evenly spaced  useful in reaching targeted groups but
4. Ratio: the data can be categorized, ranked, evenly highly discouraged especially in
spaced, and has a natural zero. economic and business research due
Population and Sample to its biased nature.
 can still be useful in qualitative
Population (all items) assessments.
a. Modal Instance
 the entire set of individuals or objects on interest,
based on typical case or certain indicators
or the measurements obtained from all individuals
(e.g. income, age etc.) which makes it
or objects of interest.
highly subjective; used in informal studies.
 Such observations have common attributes, which
b. Expert Sampling
are classified or grouped to determine certain
sampling expert subjects on a certain topic
movements, which can be useful to a researcher.
or criterion.
 The scope of such data can be from individuals to
C. Quota Sampling
even across-country cases. If all data are
 involves a pre-defined number of
considered, then the results can be said as
samples and normally used in two-
accurate.
criterion observations (e.g. gender).
Sample (items selected from the population)  The two sides can be proportional or
non–proportional. The selection will
 taking a particular part to represent the population
not stop unless the target is hit before
and facilitate easy processing in terms of lesser
stopping
cost and practicality.
D. Snowball Sampling (AKA Pyramiding)
 it involves samples based on recommendations It is a more complex form of cluster sampling,
or referrals from the prior sample. in which smaller groups are successively
2. Probability Sampling selected from large populations to form the
Definition: Subjects are chosen on the basis of sample population used in your study.
known probability (chance). The exact reversal of
Determining the Sample Size
NPS, it involves random selection and follows a
sampling frame Most of the surveys conducted are done on a sample basis
Used: Whenever possible because of financial and economic considerations, time
Advantages: Samples are representative of the and manageability of data involved. If the population size is
population thus, can be used to draw sound used when sampling from a finite population of (N)
conclusions about the population individuals, the sample size (n) may be obtained from the
Disadvantages: Costly, rigorous (hence, slow Slovin’s formula:
process) and poses considerable inconvenience on
the researcher N
n=
1. Simple Random Sampling 1+ N e 2
a sample selected so that each item or person n= sample size
in the population has the same chance of being
included. N= population size
a. Fishbowl Method
e= margin of error (the error we expect to commit in
uses the notion of “lottery” thus for each
getting the sample since it is an estimate parameter)
observation, the probability of being
selected is 1/N where N is the total MODULE 2
number of observations in the population
Collection of Statistical Data
b. Table of Random Numbers
uses an array of randomly arranged Types of Statistical Data
numbers to determine who will be
included in the sample. Primary Data
2. Systematic Sampling Government agencies, business establishments,
employs the use of intervals drawn by dividing organizations and individuals who carry original data or
the number of population by the number of have firsthand information relevant to a given problem.
target sample.
N Secondary Data
k=
n Data coming from secondary sources which include
If k is not a whole number, then round down. newspapers, magazines, journals, and published materials.
Random sampling is used in the selection of
the first invoice. Methods of Data Collection
3. Stratified Sampling
A. Direct or Interview Method
taking sample per strata or groups within a
 Interview is a face-to-face conversation between
population (e.g. the number of students per college in a
two persons in which the one soliciting information
certain university). The process is done by
(interviewer) and the one supplying the data
reversing systematic sampling:
(interviewee).
n
%k= x 100  Effective method when the elements in a sample
N
are not so numerous.
4. Cluster Sampling
 The researcher could either used personal or
taking sample from a group of the same or
telephone interview.
similar elements gathered or
 The advantage of the interview method is that the
occurring closely together (e.g. geographical
question can be repeated, rephrased, or modified
area etc.)
for better understanding.
5. Multistage Sampling (AKA Phase Sampling)
 It may be too costly or expensive. It also demands
Sampling in stages.
much time.
B. Indirect or Questionnaire Method
  Questionnaire is an instrument that contains  Presents data in paragraph form and becomes
prepared set of questions. This can be distributed effective when the objectives is to call the
either via mail or hand carry to the intended reader’s attention to some data that require
person and collects them by the same process. special emphasis.
 Structured question (AKA close-ended question) - B. Tabular Form
It is a type of question where all respondents must  presents data in rows and columns.
tick or check from a list of choices the best  more convenient and understandable than textual
appropriate response. method because the numerical information is
 Unstructured question – a type of question where displayed in a more concise and systematic
the purpose is to solicit opinion, ideas, or manner by using a vertical and horizontal lines
perception of a person. which describes the corresponding heading.
 This method of collecting data is more economical  a statistical table has four essential components:
than interview because it can involve a greater  Table heading – shows the table number and the
number of individuals in a population with the title. Table number serves to give the table an
same amount of funds. However, the researcher identity while the title briefly explains what are
cannot expect that all the questions mailed will be being presented.
retrieved since many respondents will simply  Body – shows the main part of the table which
ignore answering the questionnaire. contains the quantitative information.
C. Documents or Registration Method  Stubs – shows the opposite rows of the body and
 makes use of important documents (e.g. number of usually to the left are labels. These are
households, birth rates, death rates and marriages that can be classifications or categories which are presented as
found in both private and government offices.) values of a variable.
 It is very economical not only in terms of cost but  Box heads – the caption that appear above the
also in terms of time and effort. column.
D. Observation Method  In addition to these components, footnotes may
 used when we want to conduct a study by way of be placed immediately below the main part of the
direct observation. table and a source note may be included to
 Obtaining data pertaining to behavior of an acknowledge the origin of the data which may
individual or a group of individuals at the time of appear below the title or below the footnote.
occurrence of a given situation. C. Graphical Form
 Subjects may be observed individually or  shows visual presentation of the data.
collectively depending on the objectives of the  provide us an easier way to identify patterns of a
investigator. set of data.
 One limitation of this method lies in the fact that in  the most effective way of presenting statistical
most cases, observation is made only at the time of data because important relationship is brought
occurrence of the appropriate events. about clearly.
E. Experiment Method  comparison and trends of quantitative values are
 This method of collecting data is used to find the readily available to enable ease of communication
cause-and-effect relationship. of results or information.
 Data needed to find out the cause-and-effect 1. Bar Graph – used to organized data/information
relationship may be obtained through a series of visually. It is helpful in comparing quantities. Bar
experiments. Secondary data can be obtained graph can be horizontal or vertical.
from: journals and periodicals, newspapers, tables, 2. Line Graph – the most practical and effective
unpublished or published research papers and device which shows a general trend, pattern or
thesis and dissertations. changes over a given time. It makes use of ordered
Methods of Data Presentation pairs and graph of ordered pairs in a coordinate
plane. The categories or time periods are
A. Textual Form (paragraph method) chronologically arranged on the horizontal axis and
 Combines text and figures in a statistical the relevant values are indicated in the vertical
report. axis.
 3. Pie Graph –consists of a circular region divided R=Range
into sections that do not overlap, and each section HO= Highest Observation
represents a part or percentage of the whole being LO= Lowest Observation
considered.
4. Picture Graph or pictograph – used to describe the 3. Determine the class size or class width . This is the
difference among a few quantities. It is very distance or gap between the lower limit and the
effective tool for attracting attention since it uses upper limit. It is obtained by dividing the range by
pictures or symbols to indicate the message of the the number of classes.
obtained numerical information. C=R/ K
C= Class Interval
Frequency Distribution
R= Range
 A tabular arrangement of the data by using K= Class Interval
categories or classes and their corresponding 4. Choose an appropriate lower limit for the first-
frequencies. class interval. This number shall be less than or
 The frequency of a particular observation is the equal to the lowest value in the data. It is more
number of times the observation occurs in a convenient to use a lower limit that is divisible by
category or class. the class width. Add the class width to obtain the
 The numbers whether arranged or not by next lower-class limit. Keep on adding the class
magnitude are called raw data. width to get all the other lower-class limits.
5. Find the upper-class limits. If the class size is
One of the most convenient rules which gives explicit rounded off the unit’s place, subtract 1 from the
guidelines for the number of classes to be use is the second lower class limits to arrive at the first upper
Sturge’s Rule. class limit. Subtract 0.1 from the result, if rounded
K=1.+ 3.3 log N off to the tenth place and subtract 0.01 if rounded
to the hundredths place.
K= number of class intervals 6. Determine the class boundaries. The class
boundaries are the true limits of a class interval
N= total number of observations
made up of the lower-class boundary and upper-
log N= logarithm of N to the base 10 class boundary. The class boundary is the midway
between the upper limit and lower the limit of the
 The ideal number of class intervals should be 5 to
next higher-class interval. If we are dealing with
15.
discrete data, the class boundaries are obtained by
 Less than 8 class intervals are recommended for a
adding 0.5 to the upper limit and at the same time
data with less than 50 observations/values.
subtracting 0.5 from the lower limit. If data are
 For a data with 50 to 100 observations/values, the
continuous data, the number to be added to the
suggested number should be greater than 8.
upper limit and the number of decimal places a
 Please note that the few number of class intervals
particular observation or case has.
will result to crowded data while too many
7. Find the Class Mark or Class Midpoint – this is
numbers of class intervals tend to spread out the
needed in computing the mean and some
data too much.
measures of variability. It is obtained by taking the
Steps in Constructing Frequency Distribution Table average of the lower and upper limit

1. Decide on the number of class interval to use lower limit + upper limit
CM =
between 5-15. Too many class intervals result to 2
several empty class intervals while too few creates
8. Tally the row scores and indicate the frequency for
long details. Use the Sturge’s formula whenever
each of the class intervals.
possible.
9. Get the relative frequency. This gives us the
2. Compute the Range. This is the difference between
percentage of observations in a particular class of
the highest value and the lowest value in the set of
interest. This is obtained by dividing the frequency
data.
of the class by the total number of
R=HO−LO
frequency/observations.
 f Xi= observed value
rf = x 100
n
∑ = Summation notation
rf= relative frequency interval
Weighted Mean
f= frequency
There are some cases where values are given more
n= number of observations importance than others.
n
10. Add the frequencies and indicate the sum.
∑ xi
MODULE 3 & 4 x=
i=1
n
MEASURES OF CENTRAL TENDENCY
B. Median
Measures of central tendency commonly referred to us as
an average. The purpose of an average is to pinpoint the The median of ungrouped data arranged in an array
center in the set of observations. (increasing or decreasing order of magnitude) Is the middle
observation for an odd number of items or the arithmetic
3 measures of central tendency commonly used in mean of two middle values when the number of items in
business and economics – arithmetic mean, median and the distribution is even.
mode.
C. Mode
Statistics is any measure calculated from sample data,
thus, measures of location from a sample are statistic. The mode for the ungrouped data is defined as the value
that appears with the highest frequency.
Parameters – numerical values calculated from population
data. That is, the item that appears most often, usually denoted
by 𝒙̂ (read as x hat).
Measures of Statistics Parameter
Central (sample) (Population) It is generally used with nominal data. It can be easily
Tendency identified by inspection of an ungrouped set of data by
Mean 𝒙̅ 𝝁 getting the score or item which occurs most frequently.
Median mdn Mdn
When all values appear with the same frequency, the
Mode mo Mo
mode does not exist. A distribution with the only one
n = sample size, N = population size
mode is called unimodal while a distribution which has two
modes is bimodal; and for the same sets of data with the
Methods of Central Tendency of Ungrouped Data three or more modes is known as multimodal.
Ungrouped or raw data are those data which are not yet Measures of Relative Standing: Ungrouped Data
organized or arranged into frequency distribution.
Another set of measures that helps us describe data set
A. Arithmetic Mean are the measures of relative standing.
the sum of the values the variables divided by the number Quantiles are extension of the median concept; these are
of observations. The definition is the same for both sample the values which divide a set of data into equal parts. The
and the population, although we use a different symbol to measures of relative standing are the quartiles, deciles,
refer to each. and percentiles.
Population or Sample Mean: Percentile: The whole data set is equally divided into 100
N∨n parts.
∑ xi Decile: Data set is divided into 10 equally divided parts.
i=1
μ∨ x=
N∨n Quartile: Data set is divided into 4 parts. Note: At certain
𝝁 or 𝒙̅ = population or sample mean points, these three measures will have the same values

N= total number of observations Percentiles for Ungrouped Data

 p cfD1= cumulative frequency before decile 7 class
L p=(n+1)
100
fD1= frequency of decile 7 class
Lp= percentile location
i= class interval
p= percentile interest
Steps in Identifying the Decile 7 class.
n= number of observations
1. Multiply data set by 7 and divide the product by
The following guidelines will help identify percentile 10.
location. 2. The decile 7 class is the class that has a sum of
frequencies greater than the result of step 1.
1. If Lp is whole number, the percentile location is the
Lth in the ordered set of observations. Percentiles: Grouped Data
2. If Lp is not a whole number, the percentile location
The formula for percentiles divides the set of observations
between the Lth and (L + 1)st, by taking the
into 100 equal divisions. Same scenario, this will be also
difference between the Lth and (L+1)st location
patterned in the median formula. Example P85:
and multiply the result by the decimal portion of

( )
Lp. ln
−cf P 85
Measures of Relative Standing: Grouped Data 100
P85=lb 85 + i
f P 85
Quartiles: Grouped Data
Steps in Identifying the Percentile 85 class.
The formula for quartiles will patterned from the median
formula. Example Q1: 1. Multiply data set by 85 and divide the product by
100.

( )
1n 2. The percentile 85 class is the class that has a sum
−cf q 1
4 of frequencies greater than the result of step 1.
Q1=lb Q1 + i
f q1
MODULE 5
lbQ1= lower boundary of quartile 1 class
Measures of Variability (Dispersion): Range
n= number of observations
A. Range
cfq1= cumulative frequency before quartile 1 class

fq1= frequency of quartile 1 class

The simplest measure of dispersion is the range.

i= class interval It is the difference between the largest and the smallest
values in a data set.
Steps in Identifying the Quartile 1 class.
Ungrouped data – finding the difference between the
1. Divide the number of observations by 4. highest and the lowest value.
2. Go over the entries in the less than cumulative
frequency column. The class that has a sum of Grouped data – the range is determined by subtracting the
frequencies greater than the n/4 is the quartile 1 lower boundary of the lowest class interval from the upper
class. boundary of the highest-class interval of a frequency
distribution. (Class boundaries are considered the true
Deciles: Grouped Data limits)
The formula for deciles will also be patterned from the Interquartile Range
median formula. Example D7:
Quartile divides the distribution of numerical into four
ln equal parts.
−cf D 1
10
D7=lb D 1 +( )i The first or lower quartile lies on the 25% of the total
f D1
number values, while the third or the upper quartile is on
lbD1= lower boundary of decile 7 class the 75%.
n= number of observations
The interquartile range (IQR) – finding the difference 2. Compute the value of the mean
between the value of the third quartile (𝑄3) or upper
3. Find the individual absolute value of each deviation from
quartile and the first quartile (𝑄1) or lower quartile.
the mean
IQR = Q3 – Q1
4. Find the sum of the absolute value in Step 3 and
Semi-Interquartile Range
5. Substitute the values in the formula and solve.
The semi-interquartile range (SIQR) or quartile deviation
For the grouped data, the mean deviation or average
(QD) indicates the variation or dispersion of the values
deviation is determined by the following
covering the middle 50% of the distribution of the data is
found by getting half of the value or distance between or procedures below:
distance between the third quartile or upper quartile and
the first quartile or the lower quartile. 1. Compute the mean 𝑥̅of the distribution

SIQR or QD = (Q3 – Q1)/2 2. Subtract the mean from each of the midpoints and write
the absolute values of the results under the column 𝑥 − 𝑥̅
ICR (Grouped Data)
3. Find the product of items under column f and items

( )
1n under column 𝑥 − 𝑥̅
−cf q 1
4
Q1=lb Q1 + i 4. Add the products in Step 3 to obtain the value of ∑ 𝑓(𝑥 −
f q1
𝑥̅)
lbQ1=lower boundary of quartile 1 class 5. Divide the sum obtained in Step 4 by n.
n= number of observations Measures of Variability (Dispersion): Variance
cfq1= cumulative frequency before quartile 1 class Variance is defined as the average of the squared
deviations from the mean.
Fq1= frequency of quartile 1 class
The square root of this variance is known as standard
i= class interval
deviation.
Measures of Variability (Dispersion): Mean Absolute
The variance for a sample data is denoted by S2 (read as S
Deviation (MAD)
squared or the square of S) while the symbol for variance
Mean Deviation or average deviation is defined as the of the population is 𝜎 2 (read as sigma squared)
average of the absolute deviations of the individual values
Ungrouped Data
of a set of numerical data from either mean, the median or
mode. Among the three, the mean is the most preferred To determine the variance of an ungrouped data, let us
and commonly used measure of central tendency for follow the steps below:
computing the deviation or average deviation
1. Arrange the values according to magnitude lowest to
Ungrouped Data highest or vice versa
n
2. Calculate the mean
∑❑
i=1
3. Obtain the individual deviations from the mean
Grouped Data
4. Square each deviation and write the results under
n column |𝒙 − 𝒙̅ |𝟐
∑❑
i=1 5. Find the sum squared deviations
For the ungrouped data, the mean deviation or average 6. Divide the sum in Step 5 by n-1 for sample data or by n
deviation is determined by the following for population data.
procedures below: Measures of Variability (Dispersion): Standard Deviation
1. Arrange the values from lowest to highest or vice-versa
Standard deviation is the most important measure of Forecasting in Different Departments
variability.
 Accounting – new product/process cost estimates,
By knowing the standard deviation, we can determine the profit projections, cash management
position of the scores in a frequency distribution in relation  Finance – equipment/equipment replacement
to the mean. needs, timing, and amount of funding/borrowing
needs
A standard deviation of a small value means that the
 Human resources – hiring activities, including
values in a distribution are scattered or spread out near
recruitment, interviewing, and training; layoff
the mean and vice versa.
planning, including outplacement counseling.
S= √ S2  Marketing – pricing and promotion, e-business
strategies, global competition strategies
S=
2∑ f ∨x−x̅ ¿2  MIS – new/revised information systems, internet
n−1 services.
 Operations – Schedules, capacity planning, work
and
assignments and workloads, inventory planning,
δ=√ δ make-or-buy decisions, outsourcing, project
2

management.
Measures of Shape: Skewness  Product/service design - Revision of current
Skewness is defined as the degree of departure from features, design of new products or services
symmetry. Help managers plan the system, and the other is to help
A frequency curve that has a longer tail to the right than to them plan the use of the system
the left is said to be skewed to the right or described as Features Common to all Forecast
positively skewed distribution. Reversely, a distribution is
negatively skewed or skewed to the left if it has a tail which  Forecasting techniques generally assume that the
is longer to the left than to the right. same underlying causal system that existed in the
past will continue to exist in the future.
When the value of the skewness is zero (0) then the  Forecasts are not perfect; actual results usually
distribution is symmetric, indicating that the mean is equal differ from predicted values; the presence of
to the median randomness precludes a perfect forecast.
3( x−x ) Allowances should be made for forecast errors.
SK=  Forecasts for groups of items tend to be more
S
accurate than forecasts for individual items
FORECASTING because forecasting errors among items in a group
usually have a canceling effect.
It is a prediction, estimate or determination of what will
 Forecast accuracy decreases as the time period
occur in the future based on a certain set of factors.
covered by the forecast—the time horizon —
The value being forecast may be sales, interest rates, increases. Short-range forecasts must contend
funds, gross national product (GNP), technological status, with fewer uncertainties than longer-range
and others. forecasts, so they tend to be more accurate.

The factors on which a forecast is based may be any of the Element of a good Forecast
following: past data, opinion or judgement, company data,
1. Timely
or perceived pattern related to time.
2. Accurate
Answers the following questions: 3. Reliable
4. Expressed in Meaningful Units
1. What is the purpose of the forecast?
5. In writing
2. What are the dynamics and components of the
6. Simple to understand and use
system for which the forecast will be made?
7. Cost-effective
3. How important is the past in estimating the
future? Categories of Forecasting in Time Horizon
Short-term Forecast – It covers one day to one year and
are used mainly for short-run control such as employment,
purchasing, scheduling, sales and production rates.
Intermediate-term Forecast – a period ranging from one
season to one or two years and is used for production
schedules, revenues, cash flow and budget planning.

Long-term Forecast - a forecast covers from two to five

years or more like market trends, technology, facilities
expansion and general policy

Steps in Forecasting Process

1. Determine the purpose of the forecast

2. Establish a time horizon
3. Obtain, clean and analyze appropriate data
4. Select a forecasting technique
5. Make the forecast
6. Monitor the forecast errors

Forecasting Techniques

 Qualitative technique – it is based on qualitative

data such as the collective opinion of the sales
force to forecast the future.
 Time Series Analysis – it is based solely on
historical data accumulated over time.
Forecasting Methods or Time Series Methods
 Causal Methods – a method that define as
relationships among independent and dependent  Naïve Forecast - A forecast for any period that
variables in a system related equations. equals the previous period’s actual value.
 Simple Moving Average – It is a forecasting method
Factors in Forecasting
simply eliminates is the effects of seasonal, cyclical,
 Seasonal – factor in which variations of time series and erratic fluctuations by getting the historical
associated with a period. data. The un-weighted average of a consecutive
 Trend – is the general movement or direction in number of data points.
the data.
 Cyclical – factors applicable in longer-term regular ∑ (most recent n data values)
fluctuations which make take several years to number of time periods
complete.  Weighted Moving Average - is a time series
 Irregular – They do not reflect typical behavior, forecasting method in which the most recent data
and their inclusion in the series can distort the are heavier compared to later data. This is
overall picture. desirable to vary the weights given to historical
 Random – residual variations that remain after all data forecast future demand or sales.
other behaviors have been accounted for.

F=
∑ ( N )( W ) (S)
∑W
F= forecast of time periods

N= number of time periods

S= actual values

W=weight given
  Exponential Smoothing – refers to family of The method of measuring the strength of association
forecasting models that are very similar to (correlation) among variables. It is a descriptive statistical
weighted moving average that weights the most method that measures the relationship between two
recent past data more than the distant past data. different variables.
 Regression Analysis – is a simple statistical tool
The correlation coefficient is measures on a scale that
used to model the dependence of a variable on
varies from +1 to -1. When one variables measures, as the
one or more explanatory variables.
other measure, the correlation is positive. When one
 Simple linear regression – is the least estimator of
decreases and the other increases, it is negative. Complete
a linear regression model with a single predictor
absence of correlation is represented by 0.
(one independent variable). The least square
model determines a regression equation by
minimizing the squares of the vertical distances
between actual and predicted values of Y.

F=
∑ ( N )( W ) (S)
∑W When to Use
Shortcut Method: Trend Formula (Excel) Parametric: (Pearson’s Coefficient) Where the data must
=Trend(known Y, known X, independent variable) be handled in relation to the parameters of populations or
probability distributions. Typically used with quantitative
Making Graphs data already set out within said parameters.
Nonparametric: (Spearman’s Rank) Where no assumptions
- Scatter Diagram
can be made about the probability distribution. Typically
- Chart Element (Trendline)
used with qualitative data but can be used with
- Display Equation on Chart
quantitative data if Spearman’s Rank proves inadequate

Interpretation

Positive Correlation – Any score from +0.5 to +1 indicates a

very strong positive correlation, which means that they
both increase at the same time. The line of best fit, or the
trend line, is places to best represent the data on the
graph. In this case, it is following the data points upwards
to indicate the positive correlation.

Negative Correlation – Any score from -0.5 to -1 indicate a

strong negative correlation, which means that as one
variable increases, the other decreases proportionally. The
line of best fit can be seen here to indicate the negative
correlation. In these cases, it will slope downwards from
the point of origin.

No Correlation – Very simply, a score of 0 indicates that

there is no correlation, or relationship, between the two
variables. The larger the sample size, the more accurate
the result. No matter which formula is used, this fact will
stand true for all. The more data there is in putted into the
formula, the more accurate the result will be.

Correlation Between Interval Variables: Pearson’s

Product-Moment Correlation Coefficient

The pearson’s product moment correlation coefficient

CORRELATION ANALYSIS
(usually denoted by r) is one example of correlation
coefficient. It is the measure of the linear association
between two variable that are measured on interval of the
ratio scales

Shortcut:

=CORREL(Array 1 (X) and Array 2(Y))

Correlation Between Ordinal Variables: Spearman Rank

Order Correlation Coefficient

The spearman rank order correlation coefficient is used to

calculate the correlation of ordinal data which are
classified according to order or rank.

Rank them first