OBTAINING DATA Methods on Obtaining Data

Statistics. science that deals with the I. Methods of Data Collection

collection, organization, presentation, - a systematic method of collecting
analysis, and interpretation of data in and measuring data from different
order be able to draw judgments or sources of information in order to
conclusions provide answers to relevant
> Descriptive Statistics - procedures questions.
that organize, summarize and describe - acquiring information published
quantitative data. literature, surveys through
> Inferential Statistics - making a questionnaires or interviews,
judgment or a conclusion about a experimentations, documents and
population based on the findings from a records, tests or examinations, etc.
sample > Investigator - conducts the
inquiry
Statistical Terms > Enumerator - helps in collecting
information
Population or Universe the totality of > Respondent - information is
objects, persons, places, things used in a collected from them
particular study. Data
> Primary - Data collected in the process
Sample any subset of population. of investigation. These are collected for
the investigator’s use from the primary
Data facts, figures and information source.
collected on some characteristics of a > Secondary - collected by some other
population or sample. (quanti/quali) organization for their own use but the
> Ungrouped/Raw data - not investigator also gets it for his use.
organized in any specific way. They are
simply the collection of data as they are 3 basic methods of collecting data
gathered. > Retrospective study - use the
> Grouped Data - raw data organized population or sample of the historical data
into groups or categories with which had been archived over some
corresponding frequencies. period of time.
> Observational study - researchers only
Parameter descriptive measure of a observe the subjects and do not interfere.
characteristic of a population/whole. > Experiments designed -
experimentation and observation of the
Statistic measure of a characteristic of a resulting data is the only way to solve
sample. them.

Constant property of a population or II. Planning and Conducting Surveys

sample which is common to all members
of the group. Survey method of asking respondents
some well-constructed questions.
Variable any characteristics, number, or - Face to face
quantity that can be measured or counted. - Self administer
A variable may also be called a data item.
When designing a survey, the following Non-probability sampling - judgment or
steps are useful: subjective sampling. This method is
1. Determine the objectives of your convenient and economical but the
survey: What questions do you inferences made based on the findings
want to answer? are not so reliable
2. Identify the target population > Convenience Sampling - researcher
sample: Whom will you interview? use a device in obtaining the information
Who will be the respondents? from the respondents which favors the
What sampling method will you researcher but can cause bias to the
use? Respondents. (easiest to reach)
3. Choose an interviewing method: > Purposive Sampling - selection of
face-to-face interview, phone respondents is according to the
interview, self- administered paper characteristic of interest made by the
survey/internet survey. researcher. Randomization is absent in
4. Decide what questions you will ask this type of sampling.
in what order, and how to phrase > Quote Sampling
them. (a) Proportional - major
5. Conduct the interview and collect characteristics of the population by
the information. sampling a proportional amount of
6. Analyze the results by making each is represented.
graphs and drawing conclusions. (b) Non proportional - less
restrictive. the minimum number of
In choosing the respondents, sampling sampled units in each category is
techniques are necessary… specified.

Sampling process of selecting units (e.g., III. Planning & Conducting Experiments
people, organizations) from a population
of interest. Experiment series of tests conducted in a
Sample must be a representative of the systematic manner to increase the
target population. understanding of an existing process or
explore a new product or process
2 Ways of selecting a sample
Design of Experiments (DOE) a tool to
Probability sampling - researcher develop an experimentation strategy that
chooses samples from a larger population maximizes learning using minimum
using a method based on the theory of resources.
probability.
> Simple Random Sampling - group of Methodology of DOE
subjects (a sample) is selected for study - ensures that all factors and their
from a larger group (a population). interactions are systematically
> Stratified Sampling - taking samples investigated resulting to reliable
from each stratum or sub-group of a and complete information
population. 5 stages of DOE
> Cluster Sampling - entire population 1. Planning - carefully plan for the
is divided into groups, or clusters, and a course of experimentation before
random sample of these clusters are going to the process of testing and
selected. data collection.
 2. Screening - used to identify the - 20 basketball players from teams
important factors that affect the of 5 players. Each player can play
process under investigation out of ANY position.
the large pool of potential factors.
3. Optimization - determine the best PERMU AND COMBI
setting of these factors to achieve - A hat with 3 marbles A, B and C
the objectives of the investigation. and you just wanna find out how
4. Robustness Testing - it is many ways are there if you have to
important to make the product or pick two marbles out of the three at
process insensitive to variations random and how many ways if you
resulting from changes in factors wanna pick two?
that affect the process but are
beyond the control of the analyst. Rule of Addition
Mutually exclusive:
FORMULAS P(A ∪ B)=P(A)+P(B)
- a standard die roll being 1 or 6?
Probability (cannot occur at the same time)
𝑛(𝑆) NOT mutually exclusive:
𝑃(𝐸) = 𝑛(𝐸) P(A ∪ B)=P(A)+P(B)−P(A ∩ B)
- a standard die roll being even or
Permutation (ORDER MATTERS) prime?
- probability that the student checks
𝑛!
𝑛𝑃𝑟 = (𝑛−𝑟)!
out a work of fiction, non-fiction, or
both?
- Suppose 10 horses. 1st, 2nd, 3rd
- A class of 20 students and
Rule of Multiplication
elections are being made for
Independent: doesn’t affect the probability
President and VP. How many
of the other event.
different ways could the candidates
P(A ∩ B)=P(A)P(B)
be picked?
- Two coins are flipped. What is the
- How many ways can you arrange
probability they both land heads
all the letters in the word “MATH”?
up?
- 8 runners are running a race. In
Dependent: affects the probability of the
how many ways can a gold, silver,
other event.
bronze medal be awarded?
P(A ∩ B)=P(A)P(B|A)

Combination (NO ORDER)

Rule of Subtraction
𝑛! “at least one” scenarios
𝑛𝐶𝑟 = 𝑟!(𝑛−𝑟)! P(A)+ P(A′) =1
- Suppose 10 horses run a race;
how many ways 3 horses can Discrete Probability Distributions
finish in 1st, 2nd, 3rd in ANY - Sum of all probabilities should be 1
order? - probabilities should be confined
- On a test, a student must select 6 between 0 and 1.
out of 10 questions. In how many - CANNOT be negative (-)
ways can this be done? P(x) = 1
Cumulative Distribution Functions
𝐹(𝑥) = 𝑃(𝑋 ≤ 𝑥) = Σ 𝑓(𝑡)

F(x) = all of the probability less than or

equal to x.
x = output

Cumulative Binomial Probability

- probability that the binomial
random variable falls within a
specified range
- greater than / equal to a stated
lower limit and less than or equal
Expected Values of Random Variables
to a stated upper limit
𝐸(𝑥) = Σ 𝑥𝑓(𝑥) Poisson Distribution
- weighted average of all possible 𝑥 −µ
values µ𝑒
𝑃(𝑥) = 𝑋!
μ = (np) average # of times that an event
occurs in a certain period of time or space.
e = 2.71828 (nasa calcu)
n = number of samples
p= probability of success

The Binomial Distribution

Probability of success (p) and failure (q)

(combination)

P(x) = Probability of x
n = # of trials Cumulative Poisson Probability
x = # of successes among trials - Poisson random variable is greater
p = success than some specified lower limit,
q = failure less than some specified upper
limit.