Research Methods

Fall 2022
BAB4

Knar KHACHATRYAN, PhD

kkhachatryan@aua.am
Session 6: Agenda

I. Sampling
Why Use a Sample?

Availability of Lower cost

elements
Sampling
provides
Greater Greater
speed accuracy
 Sampling design
Availability

Example:
Firms must have evidence to substantiate the claim that safety
is a compelling marketing appeal for most vehicles.

So companies crash-test cars to test bumper strength or airbag

efficiency.

All the cars they test are destroyed, so a census would mean
destruction of all cars manufactured.
Sampling is also the only process possible if the population is
infinite.
 Sampling method
Assuming a sample is chosen, the researcher makes several
decisions when selecting a sampling method:

Probability sampling is based on the concept of random

selection of cases, a controlled procedure that assures that
each case is given a known, non-zero chance of selection.
• provides estimates of precision.
• offers the opportunity to generalize the findings to the
population of interest from the sample population.

Nonprobability sampling is arbitrary and subjective; when we

choose subjectively, we do so with a pattern or scheme in
mind.
Each member of the population does not have a known chance
of being included.
 Probability Sampling
Methods
Simple Random Sampling
• the purest form of probability sampling.
• Assures each element in the population has an
equal chance of being included in the sample
• Random number generators
Sample Size
Probability of Selection = Population Size

The AUA ugrad applicants population is of 20,000. If the

sample size is 300, the probability of selection is 1.5 percent
(300/20,000 = .015).
Systematic Sampling
• An initial starting point is selected by a random
process, and then every nth number on the list is
selected

• n=sampling interval
• The number of population elements between
the units selected for the sample

• Error: periodicity- the original list has a

systematic pattern

• If every 3rd person on the list is, say, “rich” or “senior”

or some other consistent pattern
 SYSTEMATIC SAMPLING
It is not 'simple random sampling' because different subsets
of the same size have different selection probabilities -
e.g. the set {4,14,24,...,994} has a one-in-ten probability
of selection, but the set {4,13,24,34,...} has zero
probability of selection.
 In this approach, every kth element in the population is
sampled, beginning with a random start of an element in
the range of 1 to k.
 K = skip interval =population size/ sample size
Stratified Sampling
• Sub-samples are randomly drawn from
samples within different strata that are
more or less equal on some characteristic
• Why?

• Can reduce random error

• More accurately reflect the population
by more proportional representation
 STRATIFIED SAMPLING

Draw a sample from each stratum

Most populations can be segregated

into mutually exclusive subpopulations,
or strata.
Cluster Sampling
• The primary sampling unit is not the
individual element, but a large cluster of
elements.
• Either the cluster is randomly selected
or the elements within are randomly
selected
• Why?

• Frequently used when no list of population

available or because of cost
• Ask: is the cluster as heterogeneous as the
population? Can we assume it is
representative?
Cluster Sampling example
• You are asked to create a sample of all
Management students who are working
in Dilijan during the summer term
• There is no such list available
• Using stratified sampling, compile a list
of businesses in Dilijan to identify
clusters

• Individual workers within these clusters

are selected to take part in study the
population is divided into internally
heterogeneous subgroups.
Stratified and Cluster Sampling

Stratified Cluster
 Population divided into  Population divided into
few subgroups many subgroups
 Homogeneity within  Heterogeneity within
subgroups subgroups
 Heterogeneity between  Homogeneity between
subgroups subgroups
 Random choice of cases  Random choice of
from within each subgroups
subgroup
Types of Cluster Samples
• Area sample:
• Primary sampling unit is a geographical area
• Multistage area sample:
• Involves a combination of two or more types
of probability sampling techniques. Typically,
progressively smaller geographical areas are
randomly selected in a series of steps
Nonprobability Samples

No need to
generalize

Limited
Feasibility
objectives

Time Cost
Nonprobability Sampling Methods

Convenience

Judgment

Quota

Snowball
 Non-Probability Sampling
Methods
Convenience/opportunity Sample
• The sampling procedure used to obtain
those units or people most conveniently
available
• Why: speed and cost
CONVENIENCE SAMPLING

 Use results that are easy to get

18
Judgment or Purposive Sample
• The sampling procedure where the
researcher selects units to be sampled
based on their knowledge and
professional judgment.

• A researcher wants to know what it takes

to graduate summa cum laude in
college, the only people who can give
the researcher first hand advise are the
individuals who graduated summa cum
laude.
Quota Sample
• The sampling procedure that
ensure that a certain characteristic
of a population sample will be
represented to the exact extent
that the investigator desires

• For example, an interviewer may be told

to sample 200 females and 300 males
between the age of 45 and 60. This
means that individuals can put a
demand on who they want to sample
(targeting)
Snowball sampling
• The sampling procedure in which the
initial respondents are chosen by
probability or non-probability methods,
and then additional respondents are
obtained by information provided by the
initial respondents
Sampling
Design in
the
Research
Process
 Steps in Sampling Design
 Define the target population and a case (describe those entities—collectively
and individually—that possess the desired information about the chosen
variables and its parameters).
 Define the population parameters of interest (summary descriptors—
proportion, mean, variance—of study variables).
 Identify and evaluate the sample frame (list of cases within the target
population) or create one.
 Define the number of cases needed (choose between a census or sample;
choose the size of any sample).

 Define the appropriate sampling method (the type of sample to be used).

 Define the sampling selection and recruitment protocols (choose standardized
procedures or custom-design ones).
 Define Target Population & Case

Common Types of
Target Populations in
Business Research
 Define Target Population & Case/UA
People (individuals or groups: e.g., employees, customers, suppliers).

Organizations or institutions (companies, trade associations, professional

online communities, unions).

Events and happenings (e.g., trade association meetings, presentations to

financial analysts, industry conventions, employee picnics).

Objects or artifacts (e.g., products, machines, production waste or

byproducts, tools, maps, process models, ads).

Settings and environments (e.g., warehouses, stores, factories,

distribution facilities). – supermarket merchandising

Texts (e.g., annual reports, productivity records, social media posts, emails,
memos, reports). – political marketing
 Population parameters
Population parameters are summary descriptors of variables
of interest in the population.

Sample/descriptive statistics are:

• Descriptors of those same relevant variables computed from
sample data.
• Used as estimators of population parameters.
• Depending on how measurement questions are phrased,
each may collect a different level of data.
• Each different level of data also generates different sample
statistics.
• Thus, choosing the parameters of interest will dictate the
sample type and its size.

HSD within that those that have migrant members.

 Population parameters

Example:

We want to draw conclusions about the extent of dining club use that could
be expected from married students versus single students, residential
students versus commuter students, and so forth.

Such questions have a strong impact on the nature of the sampling frame
we accept (organize the list organized by these subgroups, or within the list
each characteristic of each element would be noted), the design of the
sample, and its size.

Within this case identify 4 types of variables (N, O, I and R).

Data Types Review
Metro U: Population Parameters
What Is a Valid Sample?

Accurate Precise
 Sampling design
Accuracy
Accuracy is the degree to which bias is absent from the sample.
An accurate (unbiased) sample is one in which underestimates offset
overestimates.

Systematic variance is “the variation in measures due to some known or

unknown influences that ‘cause’ the scores to lean in one direction more
than another.”

Homes on the corner of the block, for example, are often larger and more
valuable than those within the block. Thus, a sample that selects only
corner homes will cause an overestimation of home values in the area.

Increasing the sample size can reduce systematic variance as a cause of

error.
Even the large size won’t reduce error if the sample frame from which cases
are drawn is biased.
 Sampling design
Precision

No sample will fully represent its population in all respects.

To interpret research findings, we need a measure of how closely the

sample represents the population.

The numerical descriptors that describe samples may differ from those that
describe populations because of random fluctuations inherent in the
sampling process.

This is called sampling error (or random sampling error) and reflects the
influence of chance in drawing the sample members.
When to Use a Larger Sample?

Population
variance

Number of Desired
subgroups precision

Confidence Small error

level range
 Sample size
• The greater the dispersion or variance within the population, the larger
the sample must be to provide estimation precision.

• The greater the desired precision of the estimate, the larger the sample
must be.

• The narrower or smaller the error range, the larger the sample must be.

• The higher the desired confidence level in the estimate, the larger the
sample must be.

• The greater the number of subgroups of interest within a sample, the

greater the sample size must be, as each subgroup must meet minimum
sample size requirements.

• Cost considerations influence decisions about the size and type of

sample and the data collection methods.
Sources of Error
Sample size: Statistical Precision

Must know:
 Variability of population and individual stratum
 Acceptable level of sampling error
 Needed level of confidence
 Type of distribution (if non-normal)
 Sample Size Formula

2
 zs 
n 
E
where:
n = sample size
z = confidence interval in standard error units
s = standard error of the mean
E = acceptable magnitude of error
 Sample Size Formula:
Example #1

Suppose a survey researcher, studying

expenditures on lipstick, wishes to have a
95% confident level (Z) and a range of
error (E) of less than $2.00. The estimate
of the standard deviation is $29.00.
 Calculation: Example #1

 1.9629.00 
2 2
 zs 
n    
E  2.00 
2
 56.84 
   28.42 2
 808
 2.00 
 Sample Size Formula:
Example #2

Suppose, in the same example as the one

before, the range of error (E) is acceptable
at $4.00. By how much is sample size is
reduced?
 Calculation: Example #2

1.9629.00
2 2
 zs 
n     
E  4.00 
2
56.84
   14.212
 202
 4.00 
By doubling the acceptable range of error, from +/-
$2 to +/- $4, the necessary sample size shrinks
from 808 to 202
 Calculating Sample Size
99% Confidence

2
 
2
(2.57)(29)  (2.57)(29)
n  n 
 2   4 
74.53   
2 2
 74.53
  
 2   4 
 [37.265] 2
 [18.6325] 2

1389  347
Going from a 95% confidence level to a 99% confidence
level almost doubles the required sample size.
 Determining Sample Size
Formulas:
Means n = (ZS/E) 2
Proportionsn = Z2 pq/ E2
Percentiles n = pc (100 – pc) Z2/ E2

Z at 95% confidence = 1.96

Z at 99% confidence = 2.58
 Sample Size (Mean)
Exercise 1
 We are about to go on a recruitment drive to hire some
auditors at the entry level. We need to decide on a
competitive salary offer for these new auditors. From
talking to some HR professionals, I’ve made a rough
estimate that most new hires are getting starting salaries
in the $38-42,000 range and the average (mean) is
around $39,000. The standard deviation seems to be
around $3000.

 I want to be 95% confident about the average salary and

I’m willing to tolerate an estimate that is within $500
(plus or minus) of the true estimate. If we’re off, we can
always adjust salaries at the end of the probation period.

 What sample size should we use?

Differences Between
Group Means
Differences between Groups when
Comparing Means

 Interval or ratio scaled variables

 t-test
 When groups are small

 When population standard deviation is unknown

 z-test
 When groups are large
 Null Hypothesis about Mean
Differences between Groups

  1 2

OR
  0
1 2
The null hypothesis always assumes that the means between the groups are
identical.
In other words, if you took the mean for one group and subtracted it from the
mean of the other group, then the result would be 0 if the null hypothesis is not
rejected (i.e., there’s no difference between group means).
 t-Test for Difference of
Means

1   2
t
S X1  X 2
X1 = mean for Group 1
X2 = mean for Group 2
SX1-X2 = the pooled or combined standard error
of difference between means
Distributions for Single Measures
 Symmetric Distribution

If the data on one variable is distributed symmetrically, like in this bell-shaped

curve, then the mean, the mode, and the median are identical.
 The values of the mode,
median, and mean will
differ.

One way to remember

positive versus negative
Skewed skew is by looking at
the direction of the tail.
Distributions
With positively skewed
data, the mode is less
than the median, which
in turn is less than the
mean.
 Normal Distribution
 Bell shaped
 Symmetrical about its mean
 Mean identifies highest point
 Almost all values are within +3 standard
deviations
 Infinite number of cases--a continuous
distribution
 Standardized Normal Distribution

 Area under curve has a probability density

= 1.0
 Mean = 0
 Standard deviation = 1

Normally distributed data can be standardized by

considering the area under the bell curve as a probability
density of 1.0. This sets the mean equal to ‘0’ and the
standard deviation equal to ‘1’.
Standardized Normal Curve

-2 -1 0 1 2 z
 Standardized Normal
is Z Distribution
Normally distributed data can be converted from raw scores
into Z scores, which are scaled from +3 to -3. That score
indicates the extent a score is above or below its mean.

–z +z
 Standardized Scores
Used to compare an individual value to the
population mean in units of the standard
deviation

x
z

 Linear Transformation of Any Normal
Variable into a Standardized Normal
Variable

 

  X

Sometimes the Sometimes the

scale is stretched scale is shrunk

x
z
-2 -1 0 1 2
