Data Collection

and Sampling
Methods of Data Collection
Data can be collected
➢Directly using
•Questionnaires
•Interviews
PRIMARY DATA
•Experiments
•Direct observations
➢Indirectly through
•Existing documents/records SECONDARY
DATA
 Methods of Communication
Factor Self-Administered Telephone On-line Survey Personal Interview
Questionnaire Interview
Cost Inexpensive Quite Expensive Quite Expensive Very Expensive
Speed Time-consuming Fast Moderately Fast Time-consuming
Response Rate Poor Average Average Very Good
Interviewer Bias None Likely None Highly Likely
Quality of May be vague Good May be vague Very good
Response
Type of Limited Limited Limited Wide-range
Information
SAMPLING &
SAMPLING
TECHNIQUES
 Kinsa???
SELECTION OF THE PARTICIPANTS OF THE STUDY

Pila???
COMPUTATIONOFTHESAMPLESIZE

Unsaon???
.

. METHODS OF DATACOLLECTION
.
Sampling
Techniques
(Basic Concepts)
 Selection of participants
Process of collecting information
• Census
• Sampling

There is no
need for a
hypothesis
. testing
.
Sampling
.
Sample Survey
.
Census
. Complete Enumeration
.
 Sampling is the process of
selecting observations (a sample)
to provide an adequate description
and inferences of the population.

Sample
⦿Set of units that is selected
from population
⦿ Represents the whole
population 𝑥ഥ
. ⦿ To draw inference Sample
. mean
.
𝜇
. Population
. Mean
.
 Sample
Census
Survey
.

.
 What
you want What you
to talk Population actually
about observe in
the data

Sampling
Process
Sampling Sample
Frame Inference
.

.
 Why Do Sampling?
There are several reasons for researchers do
sampling rather than conducting a census. Four
important reasons are as follows:
✓ Low cost of sampling
✓ Less time consuming in sampling
✓ Scope of sampling is high
.
✓ Accuracy of data is high
.

.
 Sampling Error
Errors
The error attributed to the variation
present among the computed
values of the statistic from the
different possible samples
consisting of 𝑛 elements.

Nonsampling Error
. The error from the other sources
. apart from the sample fluctuations
.

.
 Nonsampling Errors

1. Measurement Error
the difference between the true value of the variable and
the observed value used in the study

2. Error in the Implementation of the

Sampling Design
This occurs when we do not adhere to the procedures
.

. and requirements as specified in the sampling design.

. • Selection error; Frame Error; Population Specification Error
.

.
 Nonsampling Errors
Measurement Error

Instrument Error
Error in the design of the questionnaire
Response Error
Can be classified as response bias and non-response bias
(It can be minimized by pretesting to detect possible errors
such as the presence of vague or confusing questions.)
Response bias happens when the respondent
does not give truthful answers.
Processing Error Nonresponse bias happens the respondent
Incorrect input of data recorded in the does not know the answer or simply refuses
questionnaire into the computer to answer certain questions.
(It can be reduced through a proper training of encoders
and by incorporating quality control checks.)
Interviewer Error
.
Surrogate Information Error This error occurs when the enumerator asks
.
Use of the wrong variable to measure the the questions in a manner wherein the
.
concept under study. enumerator infuences the reply of the
. (The concept under study is emotional quotient but the respondent.
questions included in the questionnaire measure IQ instead of
. EQ.)
.
 Minimize Sampling Error
By coming up with a good sampling design
(selecting appropriate sample selection procedure, including an adequate number of elements in the sample, etc.)

Minimize Nonsampling Error

.
By constructing a good questionnaire, and by providing the necessary
.
training of data collectors, data editors and coders, and so on.
.

.
Sampling
Techniques
(Random and Nonrandom Sampling Techniques)
 Recall
Survey
Method of collecting data on the variable of interest by asking people questions.

Census
When data came from asking all the people in the population

Sampling or Sample Survey

.

.
When data came from asking a sample of people selected from a well-defined population
.

.
 Random Sampling (Probability Sampling) is a
method of selecting a sample wherein each
element in the population has a known, nonzero
chance of being included in the sample;
otherwise, it is a Nonrandom Sampling (or
Nonprobability Sampling).
.

.
Random
Sampling
Techniques
 1. SIMPLE RANDOM SAMPLING
All units of the frame are
given an equal
probability.
❑ Random number
generators
❑ Lottery

.
 2. SYSTEMATIC RANDOM SAMPLING
❑ Order all units in the sampling frame
❑ m = N/n
❑ Pick the number from 1 to m. The number is labelled as k.
❑ Then every kth number on the list is selected

.
 3. STRATIFIED RANDOM SAMPLING
❑ Population is divided
into two or more
homogeneous groups
called strata
❑ Samples are randomly
selected from each
stratum

.
 4. CLUSTER SAMPLING
❑ The population is
divided into natural
groups (clusters).
❑ Randomly pick some
clusters from all the
clusters.
❑ Completely
enumerate all
. samples from chosen
. clusters.
.

.
 5. MULTISTAGE SAMPLING
❑ Carried out in stages
❑ Using smaller and Stage 1. All regions are included
smaller sampling units Stage 2. For each region, you choose the
at each stage provinces (Simple Random Sampling)
Stage 3. For each province, you choose
the cities/municipalities (Simple Random
Sampling)
Stage 4. For each city/municipality, you
choose the barangays (Simple random
.
Sampling)
. Stage 5. Within the barangay, you employ the
. survey in the households using systematic
random sampling.
.

.
Non-Random
Sampling
Techniques
 1. CONVENIENCE SAMPLING
❑ Convenience sampling
involves choosing
respondents at the
convenience of the
researcher
❑ Very low cost
. ❑ Extensively used
.

.
❑ Restriction of Generalization
.

.
 2. JUDGMENTAL SAMPLING
❑ Researcher employs his or her
own "expert” judgment about
❑ There is an assurance of Quality
response
❑ Meet the specific objective
❑ Bias selection of sample may
.
occur
.

. ❑ Time consuming process

.

.
 3. QUOTA SAMPLING
❑ Nonprobability sampling
version of stratified
sampling
❑ Strata exist but nonrandom
selection of individual within
the group
.

.
❑ Researcher just set a quota
.

.
 4. SNOWBALL SAMPLING
❑ The research starts with a
key person and introduce the
next one to become a chain
❑ Low cost
❑ Useful in specific
circumstances & for locating
.
rare populations
.

. ❑ Projecting data beyond

.
sample not justified
.

.
Sample Size
Determination
 Sample Size Determination

𝑁 𝑛
How large should the sample
.
.
be in order to make an
.

.
accurate estimate?
.

.
 Total Population (𝑵)
This is the total number of population.
SAMPLE Level of Significance (𝑎)
SIZE 𝛼 = 0.01,0.05
At 𝛼 = 0.05, 𝑍𝛼/2 = 1.96.
At 𝛼 = 0.01, 𝑍𝛼/2 = 2.575.

𝑍𝛼2ൗ 𝑝𝑞 INFINITE Sample proportion ( 𝒑 ) – The sample

2
𝑛0 = POPULATION proportion is what you expect the results to
𝑑2 be. This can often be determined by using
the results from a previous survey, or by
running a small pilot study. If you are
𝑛0 unsure, use 50%, which is conservative and
𝑛= FINITE gives the largest sample size.
𝑛0 − 1 POPULATION 𝑞 = 1 − p.
.
1+
. 𝑁
Margin of Error (𝒅) – the difference
.
between the population parameter and the
. sample statistic (e.g. the distance between
. 𝜇 and 𝑥)ҧ
.
 Example
You are investigating the level of awareness of CMNS students
in CSU towards the accessibility law or the BP 344. Three (3)
programs were used as the target populations, namely; BS
Math (𝑁1 = 200), BS Bio (𝑁2 = 500), and BS Chem (𝑁3 = 800).
Since no data are available on the proportion of CMNS
students knowledgeable, you take the worst case scenario and
set p = 0.5 (and therefore q = 1-0.5 = 0.5). As this is a
preliminary study you are prepared to accept a margin of error
. of ± 5% so you set d = 0.05. How many students per program
. should you get for your sample?
.

.
 Example
You are investigating the level of awareness of CMNS students in CSU towards the accessibility law or the BP 344. Three (3) programs were used
as the target populations, namely; BS Math (𝑁1 = 200), BS Bio (𝑁2 = 500), and BS Chem (𝑁3 = 800). Since no data are available on the proportion
of CMNS students knowledgeable, you take the worst case scenario and set p = 0.5 (and therefore q = 1-0.5 = 0.5). As this is a preliminary study
you are prepared to accept a margin of error of ± 5% so you set d = 0.05. How many students per program should you get for your sample?

Given: 𝑁 = 1500, 𝑝 = 0.5, 𝑞 = 0.5, 𝑑 = 0.05, 𝛼 = 0.05.

So, 𝑍 𝛼Τ2 = 𝑍0.025 = 1.96.

𝑍𝛼2ൗ 𝑝𝑞
2
𝑍𝛼2ൗ 𝑝𝑞 1.96 2 (0.5)(0.5)
𝑛0 = Now, 𝑛0 = 2
= = 384.16 ≈ 𝟑𝟖𝟓
𝑑2
𝑑2 0.05 2

𝑛0 𝑛0 385
𝑛= 𝑛 −1
1 + 0𝑁
So, 𝑛 = 𝑛0 −1 = 385−1 = 306.53 ≈ 307
1+ 1+
𝑁 1500

Stratum Population (𝑵𝒊) Proportion (𝑷𝒊 =𝑵𝒊ൗ𝑵) Sample Needed (𝒏𝒊 = 𝒏 × 𝑷𝒊)
. BS Math 200 0.133 41
. BS Bio 500 0.333 103
. BS Chem 800 0.533 164
. Total 1500 100.00 308
. 200 = 307 0.133 = 41
= = 0.133
1500
.