Вы находитесь на странице: 1из 55

Principles of

Sampling

Lecture Handouts
St. Luke’s College of Nursing
SY. 2017- 2018

Tiger’s Nest Monastery, Bhutan


Scenario
You are the Chief Nurse of Valere Medical
Center and you were tasked by the Medical
Director to know the age and interview all
active staff nurses to determine if there are
any relationship between generation gap and
work-related conflicts.

Active nurses : 200


Deadline to submit data : After 1 week
Sampling
Scenario
You are a company nurse of the newly established
Valere Construction Estates and tasked to help the
HR department to do an experimental study about
cement dust exposure and stomach ulcer, due to
reported cases from 30 construction workers with
stomach ache and other related symptoms (fever,
paleness) just from the past month alone.

Current Projects : 10 (whole country)


Construction Workers per site : 200
Deadline to submit data : After 3 month
Samples per site: 50 workers , 3 in Luzon, 1 in Visayas, 1 in Mindanao
Sampling (Mendoza et al.)
The act of studying or examining only a
segment of the population to represent
the whole

Whatever findings we get


for this segment of the population, we
Sampling (Mendoza et al.)
Drawing a blood SAMPLE from a patient

We make certain or initial diagnosis of the patient’s


condition based from just a few drops of blood
*no need to examine the whole 5L of the patient’s blood
Population (Mendoza et al.)
Population (_)
The group from which
representative information is n
desired and to which inferences
will be made
*holds the True Value of desired Target
data
N

Population (_) e.g.


Population from which a sample TP = SLCN Student Nurses
will actually be taken SP = All student leaders
*holds the Estimated Value
only
TP = 80 yo+ with A.D.
SP = Out/In-Patients of SLMC
General Principle (Triola)
1. Samples should be so that the erratic
behavior that is characteristic of very small samples will
not disguise the true effects of different interventions
N True Value (age in years)
20 22 19 21 18 22 22 21 23
18
20 21 22 20 21 20 22 24 20

Average = 21

n Data Gathered (age in years)


3 19 18 20
9 19 18 20 24 22 23 22 20 21

Which n size (3 or 9) would yield a closer average with the true


value average? Which n size would be more reliable?
General Principle (Triola)
Although it is important to have a sample that
is sufficiently large…

2. it is more important to have a sample in


which data have been chosen in some
appropriate way. If sample data are not
collected in an appropriate way, the data may
be so completely useless that no amount of
statistical analysis can save them
Sampling Designs
Non-Probability Sampling

Probability Sampling
Non-Probability Sampling
The probability of each member of the
population to be selected in the sample is
difficult to determine or cannot be
specified (Mendoza et al.)
Convenience
Purposive
Haphazard
Quota
Snowball
Non-Probability Sampling
Convenience Sampling
Use methods that are (Triola)
Non-Probability Sampling
Purposive Sampling
A “representative” sample of the
population is selected based on an
or on some pre-
specified criteria (Mendoza et al.)

e.g. Dean of the college


Informal leader in the community
Head of the household
Health officer of a Rural Health Unit
Non-Probability Sampling
Haphazard Sampling
Whatever items come at hand or
(Mendoza et al.)

e.g. The research department of a hospital is


until 5pm only, therefore they can only
administer a guided interview to nurses
who are in the 6-2 and 2-10 shifts, and
cannot to nurses on 10-6 shift since they
are during the time of
study.
Non-Probability Sampling
Quota Sampling
There is a quota to meet
(Mendoza et al.)
Non-Probability Sampling
Snowball
Frequently used when studying “hidden
populations” who are .
(Mendoza et al.)

Rule: The first person identified to be a member


of the target population will be interviewed for
the study and will be asked
Non-Probability Sampling
Snowball 4th
1st 3rd
5th
2nd
Probability / Random Sampling
Each unit in the population has a known chance of
being included in the sample (Mendoza et al.)

We expect all components of the population to be


approximately proportionately represented
(Triola).
Simple
Systematic
Cluster
Stratified
Probability / Random Sampling
Sampling
The number of subjects is selected in such a way that
every possible sample of the same size has the same
chance of being chosen (Triola)
Probability / Random Sampling
Sampling
Randomly select a starting point and then select
every kth (interval) element in the population (Triola)

where:
k = N/n N = Total Population
n = Sample Size
Probability / Random Sampling
Sampling
e.g.
N = 12
n=4

k = 12/4
=3

Randomly pick the first sample from the first 3 on the


list (for example you got #2), then count by 3’s to get
the succeeding samples until you reach your desired
number (n=4)
Probability / Random Sampling
Sampling
Divide the population
area into sections (or
clusters), then
randomly select
some of those
clusters, and then
choose all members
from those selected
clusters (Triola)
Probability / Random Sampling
Sampling
Subdivide the
population into at least
two different subgroups
(or strata) that share the
same characteristics
(such as gender or age
bracket), then draw a
sample from each
subgroup. (Triola)
Introduction to
Descriptive Statistics
Lecture Handouts
St. Luke’s College of Nursing
SY. 2016 - 2017
STATISTICS
PARAMETERS
Parameters
N = Total Number of Population
n = Total Number of Sample
Descriptive Statistics
Summarize the data:
Descriptive Measures

Tabular Presentation

Graphical Presentation

Textual / Narrative
Descriptive Measures
• Measures of the Middle (Central
Tendency)
• Measures of Variability (Dispersion)
• Measures of Position
• Grouped Data
Central Tendency
Mean

Median

Mode
Mean
Arithmetic Mean =

Formula:
Mean
n=7
Variable = age

= y.o.

*therefore, the mean or average age is y.o.


Median (Md)
Value of middle most observation

Formula:
Median (Md)
Odd
n=7
Variable = age
Md = =7+1 =
2

*therefore, the value of the __th observation is the


median, which is ___ y.o.
Median (Md)
Even
n = 10
Variable = age
Md = +
= 5th obs.+ 6th obs. = 16 + 17
2 2 2
=
*therefore, the median is between the values of the 5th and 6th
observation, which is y.o.
Mode (Mo)
Value of observation which occurs most
e.g. Ages of the twenty-four patients died due to
dengue from the past month

Age

*therefore, the mode or the most frequent value in the


set is y.o.
Measures of Variability
Range

Variance

Standard Deviation

Coefficient of Variation
Measures of Variability

Gives information as to the tendency of values to


be close to each other, or Homogenous

e.g. Which batch is more homogenous in terms of age?

Student Kalinga Valere


# Age
1 17 20
2 18 24
3 19 28
4 20 30
5 21 33
Range
Difference between
e.g.
Range

e.g. Which batch is more homogenous in terms of age?

Student Kalinga Valere


# Age
1 17 20
2 18 24
3 19 28
4 20 30
5 21 33
Range 4 13
Variance
Mean of squared deviations
Variance and the arithmetic
mean are not in the same
units, therefore standard
deviation is (more often)
used instead (Rosner)
Standard Deviation
Square root of variance

Formula:
Student Kalinga Valere
# Age
1 17 20
2 18 24
3 19 28
4 20 30
5 21 33
SD 1.58 5.09
Measures of Variability

Gives information as to the tendency of values to


be close to each other, or Homogenous

e.g. Which batch is more homogenous in terms of age?


Student Kalinga Valere
# Age *therefore,
1 17 20 batch
2 18 24 is more
3 19 28 homogenous
4 20 30 than
5 21 33 in terms of age
Range 4 13
SD 1.58 5.09
Coefficient of Variation
Useful in comparing variation between
different sets of data and is dimensionless

Formula: Standard deviation divided by


mean multiplied by 100

CV = s/ x 100

Limited in use for ratio scale


Coefficient of Variation
Student Valere
# Age (y.o.) Height (cm)
1 17 166
2 18 168
3 19 169
4 20 170
5 21 172
Mean 19 y.o. 169 cm
SD 1.58 y.o. 2.23 cm
CV 8.31 1.31

*therefore, for batch Valere, their is more homogenous


than their
Measures of Position
Percentile

Decile

Quartile

Measures that aid in determining the relative position of


a particular value in an array of observation
Percentile
A percentile tells the percent of scores that
are lower than a given score. Denoted by the
letter P

cent = 1/100 or 1%
= divides the observation equally by 1%
P1
1% 100%
(minimum) P50 (maximum)
50%
(median)
Decile
Divides the data set into ten groups. Denoted
by the letter D

deca = 10/100 or 10%


= divides the observation equally by 10%
D1 D2 D 3 D4 D5 D6 D7 D8 D 9
10% 10% 10% 10% 10%
1% 100%
10% 10% 10% 10% 10%
(minimum) (maximum)
50%
(median)
Quartile
Quartiles divide the data set into 4 groups.
Denoted by the letter Q

quarter = 100/4 or 25%


= divides the observation equally by 25%

25% 25% 25% 25%


1% 100%
(minimum)
Q1 Q2 Q3 (maximum)

(median)
Measures of Position
Therefore;
• Q1 corresponds to P25
• Q2 corresponds to P50 and D5 and the median
• Q3 corresponds to P75
• D1 corresponds to P10
• D2 corresponds to P20
• D3 corresponds to P30
• D3 corresponds to P40… (so forth and so on)
• D9 corresponds to P90
Measures of Position
Application:

Locate the Q1, median, and the D8


BMI Frequency % Cum %
Underweight 4 26.67 26.67
Normal 6 40 66.67
Overweight 3 20 86.67
Obese 2 13.33 100
TOTAL n = 15 100 -
Grouped Data
Grouped Data (frequency table)
Grouped Data
Grouped Data (frequency table)
Grouped Data
Grouped Data (frequency table)
Score Frequency
Below 75 4
76 - 80 14
Frequency Tables 81 - 85 2
86 - 90 8
91 - 95 5
96 - 100 1
Components of Frequency Tables
• Class Boundaries
• Class Limits
• Class Mark
• Frequency
• Percent Frequency
• Cumulative Frequency
• Percent Cumulative Frequency
Other References

Lecture presentation on Summarizing and Presenting


Vital Statistics and Health Parameters for Biostatistics by
H. Ho, 2010.

Lecture notes on Measures of Central Tendency,


Dispersion and Location for Biostatistics 201 from UPM
CPH DEBS, 2013.

Вам также может понравиться