International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
A Study on the Educational Use of
Statistical Package for the Social Sciences
Zhenjie Sun*
Northwest University of Political Science and Law, Xi'an710063, Shaanxi, China
*
Corresponding author e-mail:
[email protected]ABSTRACT. This study was performed to suggest the effective utilization of statistical
package for educational research.More specifically, This study was approached in a
perspective of two direction. The one is the testing, the other is the prediction. In a
viewpoint of the teaching effect testing, this study is focused on the 'Post Hoc' and
'Planned Contrasts'. In a angle of our current living, this study is placed on the
index, the quotient, and multiple analysis. In a critical point of future education, this
study is put more emphasis on the factor analysis, time series analysis, the casual
relationship, the behavioral analysis, the development of alternative model, a
realistic possibility analysis of alternative future, estimation of factor in changing,
hypothesis testing, the stability testing of system.Not only the literature review was
used in this study, but also workshop articles and the results of discussion on the
educational issue in the world future society.The result of this study is expected to
help setting of the future oriented research direction, finding the needs of the more
new statistical techniques.
KEYWORDS: Futurology, Educational Statistics, Structure Equation Model, Multiple
Response, Post Hoc, Contrasts
1. Introduction
1.1 The necessity and purpose of research
We are now living in the fourth industrial revolution society. Big Data is
inextricably linked with the Fourth Industrial Revolution. The ability to extract
valuable information and analyze data is becoming important in the reality of having
to live with tens of billions of big data. The Fourth Industrial Revolution Society is
making a big difference in our lives, and changes in society will affect our job
structure. The World Economic Forum (2016) predicts a sharp drop in jobs in the
fields of business, administration, manufacturing and production by 2020 and a
significant increase in jobs in business, financial, management, and computer and
math. In addition, Mc Kinsey & Company reports that data analytics experts in the
Published by Francis Academic Press, UK
- 20 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
United States are expected to be 140,000 to 190,000 in 2021, and 1.5 million
managers will be able to make data-driven decisions. We expect the demand for
professionals involved in processing and analysis of big data to grow in the future,
which requires a change in awareness of the importance and need of Big Data
analytics technologies at the national level. Since changes in society and the work
district group inevitably affect the direction of education, our educational paradigm
should also be changed, and should be nurtured in intelligence information capacity,
a key competence in the Fourth Industrial Revolution society.
The importance of statistical analysis is growing in the field of education. In
education, which is generally developing with the trend of social science statistics,
the center of gravity of statistical use is shifting from verification to prediction.
Verification statistics are often referred to as predictive statistics when trying to
verify past events or social phenomena based on historical data, to use historical data
to predict future possibilities.
Therefore, in order to effectively spread it to the educational field in accordance
with social changes and the increasing demands for the use of statistical programs, it
is necessary to explore ways of educational use in-depth analysis of the needs of
education subjects. The purpose of this paper is to explore the teaching program of
statistical analysis software represented by SPSS, which is applicable to the field of
social science education.
2.2 Content of research
(1) To find ways to utilize the response and case ratio of Multiple Response in
education.
(2) The method of educational use of the pre- and post-mortem verification methods
is devised in the random analysis.
(3) Through consideration of the concept, background, merits and characteristics of
the integrated approach used to explain causality of phenomena or issues that are
occurring at the educational site, we seek educational use measures.
(4) Explore the types of techniques that study the future and complex uses, and
consider what ways to increase the likelihood of success of forecasting.
2. Educational Use of Percentage and Verification Statistics
2.1 Multiple Response Analysis and Percentage
Multiple responses are techniques used to analyze inspection tools designed to
enable more than one response to a single question: Categories that let you enter
only the selected variable values, and selective ones that differentiate between
choices. It is also known as a relatively simple yet reliable method of analyzing the
various needs of those surveyed in depth. However, it was not until recently widely
Published by Francis Academic Press, UK
- 21 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
used due to limitations in coding and statistical processing procedures, but in recent
years it has become easy to use with the development of statistical programs.
In particular, in the case of Categories, the choice may be limited by considering
the order, or the number of choices may be limited without considering the order. If
the sequence is considered, the use of weight cases should be used, If only two
inspection tools are to be selected in consideration of the sequence, the number of
cases in general is adjusted by multiplying the first position by two and the second
rank by one. However, if such cases are deemed too large, researchers can adjust
them to 1.5 times the first rank and 1 times the second rank on the basis of logical
justification.
The results of the multiple response analysis show that two new percentages are
presented. Response percentage means the ratio of the number of responses to the
option when the total number of responses is 100. The percentage of cases means
the percentage of responses to the option when the number of people who
participated in the survey is 100.
If a new curriculum is to be developed in the field of education, teachers,
students, and parents can be utilized to obtain in-depth analysis results, and it can
also be used to establish education plans and determine important policies.
Therefore, the percentage used in education is as follows: Percentage English
words have the meaning of '/cent' or '/100'. In other words, when the reference
quantity in the ratio is taken as 100, the number of percent of the comparison is
shown and the symbol uses '%'. Percent is a simple concept, but its usefulness is
very high. Percent is used primarily to clarify the relative size of two or more
numbers. First, if you make a reference number of 100 and then change the other
number to a number of ratios to 100, you can compare the relative sizes at a glance,
even within a limited range, even if they are different or heterogeneous.
However, many percentile data are intended for public deliberation, which is
expressed in percentage points (%P). If a percentage with the same standard is
directly compared over time, it may be added or subtracted with a normal number,
the difference between the two percentages is called 'percentage point'.
In summary, the percentage is either the ratio or the rate of change. And the
percentage point is the percentage change in the amount expressed as a percentage.
For example, if a baseball hitter enters the ninth batter's box and hits three times, his
batting average is "3/9=0.333..." However, the percentage is characterized by
multiplying these probability values by 100 and always setting the criterion as 100.
An example of a percentage is the rate of price rise. For example, at the end of
2017, suppose that prices rose by 5% at the end of 2018, and increased by 6% at the
end of 2019 compared with the end of 2018. 5% and 6% can be called the "price rise
rate", and the difference of 1% can be "1% p" To describe; can be interpreted as "the
price increase rate in 2019 increased by 1 percentage point compared to the 2018
price increase rate."
Published by Francis Academic Press, UK
- 22 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
2.2 A priori and a posteriori of variable analysis
ANOVA(ANalysis Of VAriances) is a general term for the methods used to
verify differences between averages when there are more than three groups to
compare. In this case, the independent variable of the input analysis must be the
nominal scale, and the dependent variable must be the interval or ratio scale of the
fractional variable. Based on an independent variable composed of three or more
groups, the analysis method used to verify the mean difference of a dependent
variable is called "one-way ANOVA". When more than two independent variables
are input to verify the average difference of a dependent variable, it is called
"multiple ANOVA".
The formula used for verification uses the F value calculated by dividing the
intergroup variation by the variation within each group. However, this F value only
validates the significant difference between groups, and it is not known specifically
which group and group there are significant differences. Therefore, post-verification
techniques are generally used to identify differences between these groups.
In general, it is difficult to be confident of the difference in mean between groups
before analyzing the data. For that reason, a large number of post-verification
methods are used, but if there is confidence in the difference in mean between
groups, the Planned Contrasts method may be used.
(1)Educational Application of Post-Verification Method
During the verification process of the hypothesis, researchers need to proceed in
a way to reduce type I error (α). However, reducing the probability of type I error
increases the likelihood of falling into type II error ( β ), resulting in lower
verification power (1-β) corresponding to confidence in the verification results.
Therefore, this regard in determining how after-the-fact screening need to take into
account.
LSD (the last-significant difference) performs T-verification of the mean values
of all groups, but does not reduce the likelihood of type I error. Bonferroni and
Tukey verification are strong in error control but somewhat weak in verification.
While Bonferroni's verification power is higher than Tukey's when the number of
groups to be compared, Tukey has an advantage over the larger Rock Bonferroni.
Tukey is also generally more qualified than Scheffe or Dunnett.
'Q of R-E-G-W' stands for 'Ryan-Einot-Gabriel-Welsch' not only for good
verification, but also for good control of Type I errors. However, it is not
recommended to use a different number of examples of groups to be compared.
The formula to be used for post-verification is determined by the difference in
the number of cases for the groups to be compared, whether the population is
equally distributed, and whether normal distribution can be assumed. While
relatively good verification results can be expected if the deviation is small, it is
difficult to expect good results if the parent variance is different or if the number of
cases differs between groups.
Published by Francis Academic Press, UK
- 23 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
'Hochberg GT2' and 'Gabriel' were designed to be used when there were
differences in the number of cases in comparison groups, both of which were
generally better validated but were similar when the number of cases was
significantly different. In addition, 'Hochberg GT2' should not be used for data that
is considered difficult to assume equal variances in the population.
For these reasons, there are formulas designed for use where it is difficult to
assume equal variances in the population, such as 'T2 in Tamhane', 'T3 in Dunnett',
'Games-Howell' and 'Dunnett's C''Tamhane's T2' is a generally well-informed
verification method, and 'Dunnett's T3' and 'Dunnett's C' have very good control of
the first-class error. 'Games-Howell' is a very good method of verification if the
sample size is large enough and the intergroup sample size differs greatly.
In sum, if a group with equal variances can be assumed and wants to perform
post-mortem on a group with the same sample size, use the formula "Q" (R-E-G-
W)and "Tukey" with good control over and verification of Type I errors, use
'Bonronferi' to increase control of Type I errors, and in case there is a slight
difference in the sample size. The use of the 'Games-Howell' formula is
recommended if it is deemed difficult to assume equal variances.
(2)Educational Application of planned comparison
If a researcher has sufficient knowledge and theoretical background for the
research being pursued, the method of performing statistical verification with only
intergroup preparedness of his interest is called the Planned Comparison assay.
There are two ways to compare. One is the group comparison method of "one-to-
one" called pairwise comparisons, and the other is the group comparison method of
"one-to-one," called complex comparisons. Assuming that (Table 1) uses three
teaching methods, A, B, and C, and then pre-sections are performed in the process of
conducting a random analysis to verify their effectiveness, P1, P2, P3 corresponds to
a pair comparison, and C1, C2, and C3 are composite bridges.
There are two ways to compare. One is a "1 to 1" collective comparison method
called "pairwise comparison", and the other is a "1 to many" collective comparison
method called "complex comparison" . (Table 1) Use three teaching methods A, B,
and C to conduct experimental research. In order to verify the effect, perform ex-
ante analysis in the process of variable analysis, then P1, P2, and P3 are equivalent
to the comparison between them. C1 , C2, C3 is equivalent to compound comparison.
All contrast coefficients should be added to zero. Therefore, three combinations
of 1, -1, and 0 can be made in a even comparison, and if you take the contrast
coefficient C1 of a composite comparison, the purpose of the preliminary
comparison verification is to prepare for [Professor Law (A)] versus [Professor Law
(B)+Professor Method (C)]. The comparison coefficient is then calculated as the
ratio of the number of cases in [Professor Law (A)] to the number of cases in
[Professor Law (B)+Professor Method (C)], and added to them all, it is zero (1-
14/30-16/30=0).
Published by Francis Academic Press, UK
- 24 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
Table 1 Case of Contrast Coefficients for planned comparison
The teaching The teaching The teaching
Comparison group
method A method B method C
Test variables 10 13 15
Case number 20 14 16
Comparing
1 -1 0
coefficient P1
Pairwise Comparing
1 0 -1
comparison coefficient P2
Comparing
0 1 -1
coefficient P3
Comparing
1 -14/30 -16/30
coefficient C1
Complex Comparing
-20/36 1 -16/36
comparison coefficient C1
Comparing
-20/34 -14/34 1
coefficient C1
If none of the six contrast coefficients given in <Table 3> are zero, omnibus
hypothesis, which is set to zero, will be rejected and the study will accept that at
least one contrast will not be zero.
3. Educational application of future prediction methods
3.1 Trends in statistical analysis
A brief look at the latest trends in the statistical program shows that although
previously occurring techniques such as verification of hypotheses and based on
ANOVA have been mainstream until now, from a few years ago, the Predictive
Analysis (Predicted Analysis) technique has been reinforced. The Social Science
Statistical Package (SPSS), developed for social science research, was given the new
name PASW (Predictive Analysis Software) after being acquired by IBM, and
gradually expanded its use to the natural sciences and applications. However,
considering the user's awareness, the name was restored to IBM SPSS.
3.2 The Significance of Future Research Methods
The ultimate purpose of studying the future is to explore and create a predictable
future, and to test and validate it, so that individuals, institutions, countries and
international organizations can make the smartest decisions for the future, along
with warnings. More specifically, you can use your personal successful workplace
life, the ability to respond to change, your ability to choose your future, the ability to
make decisions, to guide the future of your children and young people, to prevent
Published by Francis Academic Press, UK
- 25 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
disasters, to seize opportunities, and to understand modern society Promote self-
esteem, expand one's range of activities, etc. as the significance of future research.
These future studies have several philosophical premises: First, we do not know
the future, but we can measure the possible range of the future ahead of us, and
secondly, the possibilities and conditions of the future can be changed by policy,
which is predictable in making such policy decisions.Third, just as the sunrise can
be predicted accurately rather than the main index, prospects and possibilities for the
future are also predictable at any rate if approached in phases, and fourth, cross-
validating in two or more ways can be more predictable than the single approach
used to predict the future.Finally, humans are more interested in future prospects
than in past experiences.
3.3 Success Factors for Forecasting the Future
(1) Increase the level of understanding of decision makers or leaders in future
research methods.
(2) Decision makers, experts in the organization and futurologists jointly pursue
research.
(3)If the decision maker lacks sufficient knowledge and understanding of the
contentious issues, it must have efforts and will to overcome them.
(4)Subjective alternatives that may be of practical help to decision makers should
not be daunted by claims for objectivity.
(5)Members of the research team should consist of people with a variety of areas of
interest and experts on the status of issues.
(6)Decisions should be made on the basis of the information gathered.
(7)At least one study method that can be applied to the overall process of the issues
under study needs to be utilized.
(8)Clarify to the decision maker the economic benefits expected from implementing
the alternative.
(9)Clearly present viable technologies and alternatives for decision makers to
overcome the fear of success.
(10)It presents a variety of clear and accurate alternatives for decision makers to
choose from a political, social, and cultural perspective that aligns with their goals
and strategies.
(11)Provide knowledge to overcome distrust that it will be impossible to realize.
Knowledge at this time refers to the actual specific examples that have been found in
the surrounding institutions or countries.
(12)If there is any information or data that is inaccurate, unreliable, or potentially
confusing, then present a clear alternative.
Published by Francis Academic Press, UK
- 26 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
(13)Focus is placed on creating alternatives to overcome stereotypes through links to
similar corporate and national situations at home and abroad.
3.4 Types of future research methods
Categorizing the methods employed by futurists in future research from a
quantitative, qualitative, prescriptive, and experimental perspective is shown in
Table 2. Among them are Future Polygon, Scenarios, State of the Future Index and
Real-Time Delphi, Environmental Scanning, Futures Wheel, Cross Impact Analysis,
etc.
Emerging Issues Analysis (EIA) was first introduced by Graham Moliter in
1977 and recently compiled by Dr. Seo Yong-seok of the University of Hawaii. The
EIA takes advantage of its prediction that many of the existing problems and
conflicts could provide information about the future that is feasible, starting from the
assumption that they did not exist at any point in the past, and at the same time
providing a reflective framework for interpreting the various assumptions that have
been established in the past. So EIA is a technique that studies the early generation
of issues, and these patterns typically appear in the form of logistic distributions in
the form of a S-curve.
Trend analysis focuses on something that has already emerged and is going on.
It's been a long time since it first appeared, but it's called trend analysis to capture
the issue before it becomes a universal or universal issue. As such, the TIA focuses
on predictable history or development processes, while the EIA focuses on issues
that have just begun to arise. Thus the emerging issue develops into a trend, and this
trend has already been started and taken over by someone, and if it is high-tech, it
must be patented or paid a penalty with. On the other hand, the emerging issue has
no owner, so if I preoccupy it, I can have a patent and receive a proper royalty. This
is why emerging issues are more important in the economic sector than in the trend.
However, compared with the view of Confucian classics, from the perspective of
futurology, EIA contains the more important significance of the starting point of
research.
Table 2 Types of future research methods
Quantitati Qualitativ Normativ Explorat
Method
ve e e ory
Agent Modeling X X
Causal Layered Analysis X X
Chaos and Non-Linear Systems X X
Cross-Impact Analysis X X
Decision Modeling X X
Delphi Techniques X X X
Econometrics and Statistical Modeling X X
Environmental Scanning X X
Field Anomaly Relaxation X X
Published by Francis Academic Press, UK
- 27 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
Futures Polygon X X X X
Futures Wheel X X X
Genius Forecasting, Vision, and Intuition X X X
Interactive Scenarios X X X
Morphological Analysis X X
Multiple Perspective X X X
Participatory Methods X X
Prediction Markets X X
Relevance Trees X X
Robust Decisionmaking X X
Scenarios X X X X
Science and Technology Roadmapping X X X X
Simulation-Gaming X X
State of the Future Index X X X X
Structural Analysis X X X
Substitution Analysis
Systems Modeling X X
Technological Sequence Analysis X X
Text Mining X X X
Trend Impact Analysis X X
Visioning X X
Wild Cards X X X
Jerome C. Glenn‧
Theodore J. Gordon(2010). Future Research Methodology Ver. 3.0. The Millenium Project e-
Book, p.10.
4. Conclusion
The proportion of indices in our lives is increasing. Many of these indices on the
news, such as the Price Index, Stock Price Index, Exchange Rate Index,
Unemployment Index, Discomfort Index, Consumer Attitude Index, Travel Index,
and Ultraviolet Index, embody our lives as an indicator. The education sector is no
exception Intelligence index, sensitivity index, moral index, multi-intelligence index,
happiness index, stress index, window composition index, obesity index, standard
score, percent score, etc.
The index is used in English in various ways, such as index, indicator, quotient,
Z-score, Probability, and percentage. However, there is not much interest in the fact
that these indices are based on probabilities and percentages.
In addition, it will not be long before a highly civilized society passes by a
singularity. The human race is running toward a world where the long and long time
passed and countless events and inventions are doubled in less than a year. Perhaps
that is why it is not unusual to analyze the psychological needs of mankind.
Therefore, the ability to develop multi-response type inspection tools and process
statistics easily is expected to evolve in order to analyze individual in-depth needs
for psychological testing or learning.
In addition, it will not be long before a highly civilized society passes by a
singularity. The human race is running toward a world where the long and long time
Published by Francis Academic Press, UK
- 28 -
International Journal of Frontiers in Engineering Technology
ISSN 2706-655X Vol.1, Issue 1: 20-29, DOI: 10.25236/IJFET.2019.010102
passed and countless events and inventions are doubled in less than a year. Perhaps
that is why it is not unusual to analyze the psychological needs of mankind.
Therefore, the ability to develop multi-response type inspection tools and process
statistics easily is expected to evolve in order to analyze individual in-depth needs
for psychological testing or learning.
In addition, from a philosophical point of view, the term education implies the
concept of the 'future'. Therefore, education should always be the future, not the past.
And going into the educational scene, as long as schools exist, efforts to measure
learning effects and improve conditions based on them will continue. Therefore, it is
necessary to consider statistical methods to verify the learning effects.
In conclusion, statistical use in education can be summarized into verification
and prediction through in-depth analysis techniques based on an overall
understanding and insight into the index. It is expected that the concepts presented in
this paper will have a reasonable and positive effect on the various decisions taking
place at the educational site.
References
[1] Andy Field (2014). Discovering Statistics using SPSS (3rd. Ed.) California;
SAGE Publication Inc.
[2] Grell Jan M.(1992). Issue Analysis As An Extension of Environmental
Scanning.Futures Research Quarterly, vol.8, no.2, p.83.
[3] Huss William R.(1988). A Move Toward Scenario Analysis. International
Journal of Forecasting, vol.4, no.3.
[4] Biji Rob(1992). Delphi in a Future Scenario Study on Mental Health and Mental
Health Care, Futures, vol.2, no.3.
[5] Mitchel Vincent-Wayne(1992). Using Delphi to Forecast New Technology
Industries. Marketing Intelligence and Planning, vol.10, no.2.
[6] Nancy L. Leech, Karen C. Barrett, George A. Morgan(2008). SPSS for
Intermediate Statistics (3rd Ed.). New York; Psychology Press.
[7] Timothy C. Mack(ed.)(2011). World Future 2011 - Moving from Vision to
Action. World Future Society.
Published by Francis Academic Press, UK
- 29 -
