Advanced Statistics

The Nature of Probability and Statistics

Statistics • Students and professional people may be called

on to research in their fields since statistical
• the science of conducting studies to collect,
procedures are basic to research. To accomplish
organize, summarize, analyze, and draw
this, they must be able to design experiments;
conclusions from data.
collect, organize, analyze, and summarize data;
• came from the Latin word status and Italian
and possibly make reliable predictions or
word statista means state and statesman.
forecasts for future use. They must also be able
History of Statistics to communicate the results of the study in their
own words.
• The origin of descriptive statistics can be traced
• Students and professional people can also use
to data collection methods used in censuses
the knowledge gained from studying statistics to
taken by the Babylonians and Egyptians between
become better consumers and citizens. For
4500 and 3000 B.C. In addition, the Roman
example, they can make intelligent decisions
Emperor Augustus (27 B.C. – A.D. 17) conducted
about what products to purchase based on
surveys on births and deaths of the citizens of
consumer studies, about government spending
the empire, as well as the number of livestock
based on utilization studies, and so on.
each owned and the crops each citizen harvested
yearly. Area of Statistics
• Inferential Statistics originated in the 1600s,
when John Graunt published his book on • Descriptive Statistics – A branch of statistics that
population growth, Natural and political consists of the collection, organization,
Observations Made upon the Bills of Mortality. summarization, and presentation of data.
About the same time, another • Inferential Statistics – A branch of statistics that
mathematician/astronomer, Edmund Halley, consists of generalizing from samples to
published the first complete mortality tables. populations, performing hypothesis testing,
(Insurance companies use mortality tables to determining relationships among variables, and
determine life insurance rates) making predictions.
• A pioneer in census taking was Pierre-Simon de
Population and Sample
Laplace. In 1780, he developed the Laplace
method of estimating the population of a • Population – The totality of all subjects
country. possessing certain common characteristics that
• During 19th Century, Karl Pearson was the first to being studied.
use the insight provided by the descriptive • Sample – A group of subjects selected from a
analysis based on moments and histograms to population.
draw inferences on the underlying family of • Parameter – A characteristic or measure
distributions. obtained by using all the data values for a
• In 20th century, Ronald Fisher introduced a specific population.
systematic methodology of data collection • Statistic – A characteristic or measure obtained
through design of experiments as well as some by using the data values from a sample.
basic ideas in the theory of estimation.
• Jerzy Neyman and Egon Pearson developed the Variables and Type of Data
theory of formal hypothesis testing.
• A Variable is a characteristics or attribute that
• John Tukey invented a variety of techniques for
can assume different values.
looking more effectively at data.
• Data is the measurements or observations for a
Students Study Statistics variable.
• Variables whose values are determined by
• Students, like professional people, must be able chance are called random variables.
to read and understand the various statistical • Data set is a collection of data values.
studies performed in their fields. To have this • Data value or datum is a value in a data set.
understanding, they must be knowledgeable • Qualitative Variable – A variable that can be
about the vocabulary, symbols, concepts, and placed into distinct categories, according to
statistical procedures used in these studies. some characteristics or attribute.
 • Quantitative Variable – A variable that is 2. Mailed Questionnaire Survey
numerical in nature and that can be ordered or • Can be used to cover wider geographic
ranked. area than telephone surveys or
• Discrete Variable – A variable that assumes personal interviews since mailed
values that can be counted. questionnaire survey are less
• Continuous Variable – A variable than can expensive to conduct.
assume all values between any two specific • Respondents can remain anonymous
values. A variable obtained by measuring. if they desire.
• Hypothetical Variable – Have no real existence • A low number of responses and
except as imagined and rationally established by inappropriate answers to questions.
the mathematicians. Given certain events and • Some people may have difficulty
their characteristics. reading or understanding the
questions.
Measurement Scales 3. Personal Interview Survey
• Nominal Level of Measurement – A • Have the advantage of obtaining in-
measurement level that classifies data into depth responses to questions from the
mutually exclusives (non-overlapping) person being interviewed.
exhaustive categories in which no order or • The interviewers must be trained in
ranking can be imposed on the data. asking questions and recording
• Ordinal Level of Measurement – A responses, which makes the personal
measurement level that classifies data into interview survey more costly that the
categories that can be ranked, however, precise other two survey methods.
differences between the ranks do not exist. • The interviewer may be biased in his
• Interval Level of Measurement – A or her selection of respondents.
measurement level that ranks data and in which • Data can also be collected in other ways, such as
precise difference between units of measure 1. surveying records
exist. Zero is arbitrary. 2. direct observation of situations
• Ratio Level of Measurement – A measurement • Qualities of Good Survey Questions
level that possess all the characteristics of 1. Evokes the truth.
interval measurement and a true zero. It also has 2. Asks for an answer on only one dimension.
true ratio between different units of measures. 3. Can accommodate all possible answers.
4. Has mutually exclusive options.
Data Collection 5. Flows well from the previous question.
6. Does not make erroneous assumptions.
• One of the most common methods of data
7. Does not imply a desired answer.
collection is through the use of survey.
8. Does not use emotionally loaded or vaguely
• Survey can be done by using a variety of
defined words.
methods.
9. Does not ask the respondent to rank more
1. Telephone Survey
than five items in each series.
• Have an advantage over personal 10. Puts personal questions at the end of the
interview surveys in that they are less survey.
costly.
• People may be more candid in their Sampling Techniques
opinions since there is no face-to-face
• Researchers use samples to collect data and
contact.
information about a particular variable from a
• A major drawback to the telephone
large population.
survey is that some people in the
• Using samples saves time and money and
population will not have phones or will
enables the researchers to get more detailed
not answer when the calls are made.
information about a particular subject.
• Many people now have unlisted
• Samples cannot be selected in haphazard ways
numbers and cell phones.
because the information obtained might be
• Even the tone of the voice of the
biased.
interviewer might influence the
• To obtained samples that are unbiased, give
response of the person who is being
each subject in the population an equally likely
interviewed.
chance to being selected.
Non-Probability Sampling ➢ subjects are selected by using chance
• a sampling technique where the samples are method or random numbers.
gathered in a process that does not give all the ➢ a sample for which every member of
individuals in the population equal chances of population has an equal chance of being
being selected. selected.
• convenience sampling, consecutive sampling, ➢ lottery method
quota sampling, judgmental sampling, and • Systematic Sampling
snowball sampling ➢ subjects are selected by numbering each
• Convenience Sampling subject of the population and then selecting
➢ the most common of all sampling techniques every kth subject.
➢ the samples are selected because they are ➢ when using systematic sampling, one must
accessible to the researcher be careful about how the subjects in the
➢ subjects are chosen simply because they are population are numbered.
easy to recruit ➢ numbering is not always necessary.
➢ this technique is considered easiest, • Stratified Sampling
cheapest and least time consuming ➢ Subjects are selected by dividing the
• Consecutive Sampling population into groups, called strata,
➢ very similar to convenience sampling except according to some characteristic that is
that it seeks to include ALL accessible important to the study, then sampling from
subjects as part of the sample each group, stratum.
➢ this sampling technique can be considered ➢ Samples within the strata should be
as the best of all non-probability samples randomly selected.
because it includes all subjects that are ➢ Proportional Allocation – The sizes of the
available that makes the sample a better samples from the different strata are
representation of the entire population proportional. Sample sizes for proportional
• Quota Sampling allocation
➢ a non-probability sampling technique
𝑵𝒊
wherein the researcher ensures equal or 𝒏𝒊 = ∙𝒏
𝑵
proportionate representation of subjects
depending on which trait is considered as for i = 1, 2, 3, …, and k
basis of the quota
where n = n1 + n2 + n3 + … + nk
• Judgmental Sampling
➢ more commonly known as purposive • Cluster Sampling
sampling ➢ Subjects are selected by using an intact
➢ subjects are chosen to be part of the sample group that is representative of the
with a specific purpose in mind population.
➢ the researcher believes that some subjects ➢ The population is divided into groups called
are more fit for the research compared to clusters by some means such as geographic
other individuals area or schools in a large school district, etc.
• Snowball Sampling Then randomly selects some of these
➢ usually done when there is a very small clusters and uses all members of the
population size selected clusters as the subjects of the
➢ the researcher asks the initial subject to samples.
identify another potential subject who also ➢ Cluster sampling is used when the
meets the criteria of the research population is large or when it involves
➢ it is hardly representative of the population subjects residing in a large geographic area.
• Multistage Sampling
Probability Sampling
➢ the taking of samples in stages using smaller
• a sampling technique in which sample from a and smaller units at each stage
larger population are chosen using a method ➢ a complex form of cluster sampling because
based on the theory of probability it is a type of sampling which involves
• simple random sampling, stratified sampling, dividing the population into groups (or
cluster sampling, multistage sampling, and clusters)
systematic random sampling
• Random Sampling
Uses of Statistics • Changing the Subject
o Another type of statistical distortion can
• Education – Statistical tools are to get occur when different values are used to
information on enrolment, finance, physical
represent the same data.
facilities which are needed for a good o Ask yourself, “Which measure best
administration and management. represent the data?”
• Government – It provides the various • Detached Statistics
government offices with organized records on o A claim that uses a detached statistic is one
the movement of population, cost of living,
in which no comparison is made.
wages, taxes and material resources which are o Ask yourself, “Compare to what?”
necessary for intelligent policymaking.
• Implied Connections
• Business and Economics – Statistics plays an o Many claims attempt to imply connections
important role in the exploration of new markets between variables that may not actually
for a product forecasting of business trends,
exist.
control on the quality of goods produced and o Be careful when you draw conclusions form
improvement of personnel relations. claims that use words such as may, in some
• Medicine – Methods for the statistical design of people, and might help.
scientific experiments are valuable to the • Misleading Graphs
researchers. Causes and effects of factors which o Statistical graphs give a visual
affect experiments are best evaluated using
representation of data that enables viewers
statistical techniques. to analyze and interpret data more easily
• Psychology – Statistics is used to systemize, than by simply looking at numbers.
analyze, interpret data on intelligent scores, However, if graphs are drawn
aptitudes, personality trait, ratings and inappropriately, they can misrepresent the
attitudes. data and lead the reader to false
• Sociology and Population Dynamics – Studies on conclusions.
population movement, births, deaths, urban • Faulty Survey Questions
planning labor movement help in solving the o When analyzing the results of a survey using
problems of man and his society. questionnaires, you should be sure that the
• Sports – Figures on shooting averages, fouls, questions are properly written since the way
team assists in basketball, give the audience a questions are phrased can often influence
clear picture of the performance of their sports the way people answer them.
heroes. These statistical figures help in
maintaining interest in various sports. Computers and Calculators

Misuses of Statistics • TI-83 Plus or TI-84 Plus graphing calculator

• MINITAB and Microsoft Excel
• Suspect Samples • Statistical Package for the Social Sciences (SPSS)
o The first thing to consider is the sample that and Perfect Statistics Professionally Presented
was used in the research study. (PSPP)
o Sometimes researchers use very small
samples to obtain information.
o It is necessary to see how the subjects in the
sample were selected.
o When results are interpreted from studies
using small samples, convenience samples,
or volunteer samples, care should be used in
generalizing the results to the entire
population.
• Ambiguous Averages
o Selecting the one measure of average which
lends the most evidence to support their
position.