45
Indonesian Journal of Science & Technology 9.
45-108 Indonesian Journal of Science & Technology Journal homepage: http://ejournal.
edu/index.
php/ijost/ How to Calculate Statistics for Significant Difference Test Using SPSS: Understanding Students Comprehension on the Concept of Steam Engines as Power Plant Meli Fiandini1.
Asep Bayu Dani Nandiyanto1*.
Dwi Fitria Al Husaeni1.
Dwi Novia Al Husaeni1.
Mushiban2 Universitas Pendidikan Indonesia.
Jl.
Dr.
Setiabudhi No.
Bandung.
Indonesia MTs.
N 8 Kebumen.
Jl.
Lapangan Manunggal Wero Gombong.
Kebumen.
Jawa Tengah *Correspondence: E-mail: nandiyanto@upi.
ABSTRACT
A significant difference test is used to evaluate certain treatments on the sample in two different observation One of the commonly used software is SPSS which is used to analyze data which helps researchers in calculating data so that it can be completed quickly.
However, there are still many students and researchers who are not experts in calculating data using SPSS software, especially significant difference tests.
This article aims to provide a step-by-step guide in calculating data using SPSS for statistical requirements and significant difference tests.
To understand the calculations well using SPSS, we demonstrate the requirements tests .
, normality and homogeneity test.
, parametric significant difference tests .
One Sample ttest.
Paired sample t-test, and Independent Sample t-tes.
, and non-parametric .
Wilcoxon test and Mann-Whitney We also added and demonstrated the steps for calculating data in the field of education with the variables analyzed being differences in student learning outcomes.
used the data when delivering the steam engine concept to students, showing how statistical calculation can understand students' comprehension.
Bibliometric analysis regarding statistics was also added.
This paper can be used as a guide in carrying out statistical tests using SPSS software.
A 2024 Tim Pengembang Jurnal UPI
ARTICLE INFO
Article History:
Submitted/Received 21 Aug 2023
First Revised 15 Sep 2023
Accepted 09 Nov 2023
First Available online 10 Nov 2023
Publication Date 01 Apr 2024
____________________
Keyword:
Average difference test.
Experimental demonstration.
Bibliometric.
Concept.
Islamic school.
Power plant energy.
SPSS.
Steam engine.
Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 46
INTRODUCTION
Statistics is a branch of science that is the backbone of data processing and analysis.
In the modern information age, where data has become more abundant and diverse, statistics plays an increasingly important role in uncovering patterns, trends, and meanings from that data (Wan, 2.
One type of statistic that is frequently and widely used is the difference test statistic (Morris et al.
, 2019.
Afifah et al.
, 2.
Using different test statistics allows researchers to explore and understand significant differences between different data groups, variables, or populations (Savalei & Rhemtulla, 2.
In the worlds of science, health, and business, an accurate understanding of these differences is the foundation for informed decision-making, the development of new theories, and the improvement of existing practices and processes (Cvitanovic et al.
, 2.
Different test statistics have wide applications in various scientific disciplines, from social sciences, and economics, to health sciences and natural sciences (Mishra et al.
, 2.
Research and statistical analysis of difference tests can provide valuable insight into the phenomena we observe, such as the effect of medical intervention on a group of patients, a comparison of educational outcomes between two schools, or the impact of a marketing strategy on product sales (Vinje et al.
Difference test statistics are not just about numbers and formulas.
It also involves interpreting data, understanding context, and being able to identify the essence of statistical findings (Arican & Kuzu, 2.
That is why the use of different test statistics has become a hot topic in research .
ee Figure .
, confirmed by the bibliometric analysis from the Scopus database, reaching 50,675 documents between 1894 and 2023, taken on 27 October 2023 .
sing keywords Audifferent test statisticsAy.
Detailed information for further bibliometric analysis is reported in previous studies (Al Husaeni & Nandiyanto, 2.
To test the data, it is important to use appropriate analytical instruments.
The use and standards of research results will be more negatively impacted by errors in selecting analytical instruments, which can have a negative impact on drawing conclusions.
Data analysis errors can be avoided with the help of statistical software (Hayat, 2.
Figure 1.
The bibliometric analysis from the Scopus database, taken on 27 October 2023.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 47 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 IBM SPSS (Statistical Package for the Social Science.
is statistical software used for statistical data analysis.
Some of the main uses of SPSS software are statistical analysis, data manipulation, data visualization, big data processing, prediction and processing models, decisions and action taking, as well as research and publication (Nwogwugwu & Ovat, 2.
Apart from that, it is necessary to pay attention when carrying out statistical analysis, first carrying out the required tests, namely the normality test .
o determine whether the population variance is the same or no.
and the homogeneity test .
o determine whether the distribution of research data is normal or no.
The test results of the requirements are used to determine further analysis, namely parametric and non-parametric tests (Orcan, 2.
The difference between the two tests is based on the assumption that the data is normally distributed or not.
When choosing between parametric and nonparametric tests, it is important to understand the data characteristics and the research question (Delacre et al.
Several previous studies have illustrated the statistical effectiveness of difference tests in revealing significant differences in various contexts.
For example, in the field of education, research conducted by Nandiyanto et al.
Nandiyanto et al.
Ragadhita et al.
Afifah et al.
used different test statistics to compare student learning achievements between two different learning methods.
In the health sector, research conducted by Sun et al.
used different test statistics to evaluate the effects of various treatments on patients with certain medical conditions.
Additionally, in the context of the business run by Soni .
, different test statistics are used to compare the performance of different products or marketing strategies in measuring their impact on sales.
These studies demonstrate the extent to which difference test statistics can provide valuable insights to support informed decision-making.
Even though there have been many studies discussing the use and application of different test statistics, there is still no research discussing step-by-step statistical tests using SPSS.
Apart from that, there are still students and researchers who are not yet experts in calculating data using SPSS, especially significance difference tests.
Therefore, this article was created with the aim of providing a step-by-step guide in calculating data using SPSS software to test statistical requirements and test for significant differences.
Apart from that, this article discussed basic statistical concepts regarding test requirements .
Normality and Homogeneity test.
as well as parametric significant difference tests .
One Sample t-test.
Paired Samples t-test, and Independent Samples t-tes.
and non-parametric .
MannWhitney test and Wilcoxon tes.
In this article, real cases were also presented in the use of different test statistics in the educational sector using the SPSS application.
Differences in learning outcomes are caused by the use of experimental demonstration methods in teaching steam engines as a teaching aid, which were used to help students understand heat and renewable energy.
This article can provide a strong basic understanding of statistics and can be a useful reference source for anyone interested in understanding more about the role of statistics, especially difference tests in processing, analyzing, and interpreting data in various THEORETICAL FRAMEWORK Steam engine A steam engine is an external combustion engine that uses air or gas as its working fluid and works based on thermodynamic circulation principles.
Steam engines are commonly used as pump engines, power-generating engines, and solar-powered engines.
Steam engines utilize temperature differences, pressure changes, and changes in the volume of working fluid DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 48 in a closed system.
This engine could be an alternative driving engine that can use renewable energy sources as an answer to the fuel oil crisis that is currently hitting the world (Herrero et al.
, 2.
In theory, the combustion of a steam engine occurs outside, producing heat from the fuel which is not directly converted into motion, but first through a conducting medium and then converted into mechanical energy.
Mechanical energy can be converted into electrical energy if it is connected to an electric generator dynamo.
This machine was created because it can convert solar thermal energy, which is abundant in nature, into rotary motion energy to drive a generator and convert it into electrical energy (Malele & Ramaboka, 2.
Steam engine design is carried out by first determining the output power to be produced, expansion temperature, and compression temperature.
From these values, the method continues with calculations to determine the dimensions and materials used in the The basic principle of this machine is to get pressure by expanding gas when it is cooled.
In a Stirling engine, varying pressure is generated by the shifting piston which alternately works between the cold parts of the cylinder (Brit & Cowling, 2.
METHOD
Design Procedure The data collected consists of literature and field studies.
Data obtained from the literature review is used as an example, to demonstrate step-by-step carrying out test requirements (Normality and Homogeneity tes.
, significance difference tests parametric .
One Sample t-test.
Paired Samples t-test, and Independent Samples t-tes.
and non-parametric statistics .
Mann-Whitney test and Wilcoxon tes.
The flow diagram for calculating significant different test data is shown in Figure 2.
Figure 3 presents the different uses of statistically significant differences based on the data obtained Figure 2.
Statistical flow chart of significant difference test.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 49 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 3.
The different uses of statistically significant differences based on the data Meanwhile, data in the field was obtained directly.
The subjects used were Junior Islamic School class IX students, around Kebumen.
Central Java.
A total of 60 students were divided into two classes, namely the control class and the conventional class.
The design used is a pretest-posttest control group design.
The two classes are given different treatment.
In the control class, this is done by giving an initial test .
before being given treatment, after being given treatment .
learning vide.
, and a final test .
Whereas, the experimental class was carried out by giving an initial test .
before being given treatment, after being given treatment only in the form of material, and a final test .
Learning outcomes in the form of pretest and posttest are used to analyze differences in students' understanding of the concept of the steam engine as a power plant using conventional methods and experimental demonstration videos.
Calculation Steps Using SPSS Furthermore, the data collected is then calculated using SPSS version 26.
The selection of data processing features using SPSS is adjusted to the processing objectives and what data you want to conclude.
In this article, significant difference tests are calculated using SPSS which consists of test requirements .
Normality test and Homogeneity tes.
, parametric difference test (One Sample t-test.
Paired Sample t-test, and Independent Sample t-tes.
, and non-parametric test .
Wilcoxon test and Mann-Whitney tes.
The description of the basic steps in calculating and processing data using SPSS version 26 software is as follows:
Open the SPSS application.
Click Windows then search "IBM SPSS Statistics".
After that, click AuOpen IBM SPSS StatisticsAy (See Figure .
Furthermore, on the home page, click New Dataset to open a new worksheet (See Figure .
After clicking "New File", there are two worksheets, namely input data and output data .
ee Figures 6 and .
In data input, there are two worksheets, namely "Data View" and "Variable View".
Data View is used to display SPSS worksheets to display the contents of the input data.
Meanwhile, "Variable View" is used to edit and view data variables on the DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 50 Figure 4.
Open the IBM SPSS Statistics.
Figure 5.
The initial appearance of SPSS.
Figure 6.
Data input display in SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 51 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 7.
Data output display in SPSS.
After the SPSS software has been successfully opened.
First, create a variable.
On the data input worksheet, click Variable View in the lower left corner column, and then the Variable View worksheet will appear as shown in Figure 8.
When entering variables, there are several aspects you must pay attention to, namely as follows:
The variable name entered must begin with a letter and not be capitalized .
The data type selected must be adjusted to the type of data being analyzed, for example selecting .
then selecting "String" in the type column.
Apart from that, the data type when entering variable data must be adjusted to the data type, be it numeric, ordinal, or the scale contained in the AuMeasureAy column.
The number of decimals used must be adjusted to the data entered.
For example, the value of class IX students' understanding ability in the science of energy and heat material is presented in Figure 9.
Figure 8.
Variable view worksheet displays in SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 52 Figure 9.
Display of variable view data input results with SPSS.
The next step is to input data.
Click AuData ViewAy in the bottom left corner.
To enter data, it must be adjusted to the available variable columns (See Figure .
Figure 10.
Display of data input results with SPSS.
After entering all the data into the variable column, then, we use data processing features in SPSS such as the normality test, homogeneity test, parametric t-test, and nonparametric t-test.
RESULTS AND DISCUSSION
Example of processing data with requirements testing Before data calculation and analysis were carried out, a requirements test was first carried Prerequisite testing is a basic concept for determining which test statistics are required, and whether the test uses parametric or non-parametric statistics.
The prerequisite tests consist of a normality test and a homogeneity test.
Below is a detailed explanation of the normality test and homogeneity test:
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 53 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 1 Normality test The normality test is a test carried out to check whether our research data comes from a population with a normal distribution.
This test needs to be carried out because all parametric statistical calculations assume the normality of distribution.
Normal data has the characteristic that the mean, median, and mode have the same value (Mishra et al.
, 2.
In testing the normality test, data is said to be normally distributed if the significance value is more than 0.
> 0.
The basis for decision-making in the Kolmogorov-Smirnov Normality test is (Knief & Forstmeier, 2.
, .
If the significance value (Sig.
) is greater than 0.
05, then the research data is normally .
On the other hand, if the significance value (Sig.
) is smaller than 0.
05, then the research data is not normally distributed.
In this article, the normality tests that will be discussed are Kolmogorov-Smirnov and Shapiro-Wilk.
Kolmogorov-Smirnov The Kolmogorov-Smirnov Normality test is a test carried out to determine the distribution of random and specific data in a population.
Based on tests carried out by the National Institute of Standards and Technology (), the Kolmogorov-Smirnov test is suitable for data sizes of 20 - 1000.
However, in research in general, the Kolmogorov-Smirnov test is still used for data samples with a size of more than 1000 samples .
O N O 1.
Therefore, it is recommended to use the Kolmogorov-Smirnov test for data above 50 samples (Usman, 2.
For example, citing the data provided from research conducted by Zakaria et al.
, we have 60 respondents' data consisting of data on Motivation (X) and Mathematics Learning Achievement (Y) as shown in Table 1.
In this context, we carry out a normality test Kolmogorov-Smirnov on the Unstandardized Residual (RES_.
value for the regression equation for the influence of learning motivation on learning achievement.
We carried out the Kolmogorov-Smirnov normality test with the help of the SPSS version 26 statistical Table 1.
Example of Kolmogorov-Smirnov test research data .
esults are not norma.
Respondent Motivation (X) Achievement (Y) DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 54 Table 1.
(Continu.
Example of Kolmogorov-Smirnov test research data .
esults are not Respondent Motivation (X) Achievement (Y) The steps for carrying out the Kolmogorov-Smirnov normality test .
sing data in Table 1 with results that are not Norma.
with SPSS are as follows:
Prepare the data to be tested.
Open the SPSS application as shown in Figure 11.
Then, click Display Variables in the bottom left corner.
Next, in the "Name" column, write Motivation in row 1 and Achievement in row 2.
In the AuDecimalsAy column change it from DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 55 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 2 to 0 for Learning Motivation and Learning Achievement.
After that, change the "Measure" column for the motivation and achievement variables from unknown to scale (See Figure .
After that, click the "Data View" menu in the lower left corner.
After that, enter the data on the motivation and learning achievements of the respondents that have been prepared previously .
s can be seen in Figure .
The next step is to calculate or generate the unstandardized residual value (RES_.
to test its normality.
RES_1 is the difference between the observed value and the predicted value, and absolute is an absolute value.
The way to display the RES_1 value is by clicking the "Analyze" menu on the SPSS toolbar, then clicking "Regression" on the dropdown button then selecting Linear .
ee Figure .
Thereafter, the Linear Regression windows page will appear as shown in Figure 15.
In the Linear Regression dialog box, enter the Learning Achievement (Y) variable in the Dependent box and the Learning Motivation (X) variable in the Independent box.
Move it using the arrow buttons provided in the dialog box.
Once finished, click "Save".
Figure 11.
SPSS home page.
Figure 12.
Display the AuVariable ViewAy worksheet settings in Kolmogorov-Smirnov Test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 56 Figure 13.
Display the AuData ViewAy worksheet settings in a Kolmogorov-Smirnov Test with SPSS.
Figure 14.
The settings display the unstandardized residual (RES_.
value for the Kolmogorov-Smirnov Test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 57 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 15.
Linear regression settings for the Kolmogorov-Smirnov Test with SPSS.
After clicking "Save" in the Linear Regression dialog box, a new dialog box will appear with the title "Linear Regression: Save" as shown in Figure 16.
In the "Linear Regression:
Save" dialog box, the "Residuals" section check or select the option Unstandardised, then click the AuContinueAy button.
Figure 16.
Settings in the linear regression dialog box: Save Kolmogorov-Smirnov Test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 58 .
After that, you will be directed back to the "Linear Regression" dialog box, then click the "OK" button .
ee Figure .
Figure 17.
Click AuOKAy on the Linear Regression Dialog Box.
After you click "OK" in the linear regression dialog box, a new variable will appear on the data view page with the name RES_1, as shown in Figure 18.
Figure 18.
Worksheet display "Data View": Occurrence of variable RES_1.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 59 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 .
If you open the output window in SPSS, the regression results will be displayed as shown in Figure 19.
At this stage, the Kolmogorov-Smirnov test has not been completed.
Figure 19.
Regression calculation results in SPSS windows output.
The next step is the Kolmogorov-Smirnov normality test.
The way to test normality using this method in SPSS is to select the "Analyze" menu on the top toolbar, then select "Nonparametric Tests" and click "Legacy Dialogs".
Then select Au1-Sample K-SAy in the Legacy Dialogs submenu .
ee Figure .
Figure 20.
Menu selection method for the Kolmogorov-Smirnov Test in SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 60 .
After that, the "One-Sample Kolmogorov-Smirnov Test" dialog box will appear on the monitor screen .
ee Figure .
Next, you enter the variable "Unstandardized Residuals" into the AuTest Variable ListAy box.
In the "Test Distribution" section, activate or check the "Normal" option, then click "OK" .
ee Figure .
Figure 21.
Kolmogorov-Smirnov Test one sample box dialogue.
Figure 22.
Settings in the One-Sample Kolmogorov-Smirnov Test dialog box.
After the calculation is complete, the SPSS window output appears.
Figure 23 shows the output results of the Kolmogorov-Smirnov Test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 61 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 23.
Windows output Kolmogorov-Smirnov Test results with SPSS Figure 24 shows the results of the Kolmogorov-Smirnov Test calculation using SPSS.
Based on these results, it is known that the significance value of Asymp.
-taile.
> 0.
Figure 24.
Kolmogorov-Smirnov Test calculation results: Example of abnormal data with SPSS Based on the SPSS output table, it is known that the Asymp.
Sig .
-taile.
significance value 011 < 0.
This follows the basic decision-making of the Kolmogorov-Smirnov normality test in Figure 24, the data is not normally distributed.
Thus, the assumptions or requirements for normality in the regression model are not met.
Examples of data with normal KolmogorovSmirnov test results are shown in Table 2.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 62 Table 2.
Example of Kolmogorov-Smirnov test research data .
esults are norma.
Respondent Motivation (X) Achievement (Y) RES_1 Figure 25 shows the results of data calculations in Table 2 using the Kolmogorov-Smirnov Based on the SPSS output table, it is known that the Asymp.
Sig .
-taile.
significance value of 0.
200 is greater than 0.
thus, following the basis for decision-making in the Kolmogorov-Smirnov normality test, it can be concluded that the data is normally distributed.
Thus, the normality assumptions or requirements in the regression model have been met.
Figure 25.
Kolmogorov-Smirnov Test calculation results: Example of normal data with SPSS DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 63 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Shapiro-Wilk The Shapiro-Wilk Normality Test is carried out to determine whether the distribution is normally distributed.
The Shapiro-Wilk test is used for limited data sizes, namely samples of less than 50 to produce precise and accurate decisions (Killic, 2.
To provide an understanding of the Shapiro-Wilk normality test, an example is presented by citing data from research conducted by Hairida .
, we have 30 respondents' data consisting of pretest and posttest data which is shown in Table 3.
In this context, we carry out the Shapiro normality test -Wilk to determine the differences in the influence of class VII junior high school student's knowledge of electricity.
Here, we carry out the Shapiro-Wilk normality test with the help of the SPSS version 26 statistical application.
Table 3.
Example of Shapiro-Wilk Test research data .
esults are norma.
Pretest Posttest The following are the steps for carrying out the Shapiro-Wilk normality test with SPSS software as follows:
Prepare the data to be tested.
Open the SPSS application and prepare a new SPSS program worksheet shown in Figure .
Next, click on the bottom left corner of "Variable View".
Then, in the "Name" column, enter the pretest and posttest variables in rows 1 and 2, respectively.
In the AuDecimalsAy DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 64 column, change it from 2 to 0.
After that, in the column, the "Measure" pretest and posttest variables were changed from unknown to scale (See Figure .
Figure 25.
SPSS worksheet home page.
Figure 27.
Display the AuVariable ViewAy worksheet settings in Shapiro-Wilk Test with SPSS.
After that, click "Data View" in the lower left corner and enter the pretest and posttest data as shown in Figure 28.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 65 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 28.
Display the AuVariable ViewAy worksheet settings in Shapiro-Wilk Test with SPSS.
The next step is to calculate the normality test using Shapiro-Wilk test.
Starting by clicking on the SPSS main menu, selecting the "AnalyzeAy then, selecting the AuDescriptive statisticsAy submenu then clicking "Explore" (See Figure .
Figure 29.
Settings for calculating the Shapiro-Wilk normality test with SPSS.
After that, the Shapiro-Wilk normality test worksheet page will appear as in Figure 30.
the "Explore" normality test dialog, move the pretest and posttest variables to the DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 66 "Dependent List" contact using the button provided.
After that, the Shapiro-Wilk normality test worksheet page will appear as in Figure 30.
Figure 30.
Explore settings on the Shapiro-Wilk normality test with SPSS.
Next, you click "Plots" and then a dialogue AuExplore: PlotsAy box appears.
Then, check "Normality plots with test", then click "Continue" (See Figure .
Figure 31.
Explore Plot settings in the Shapiro-Wilk normality test with SPSS.
After that, you will be directed back to the "Explore" dialog box then click the "OK" button (See Figure .
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 67 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 32.
Explore plot settings: Click AuOKAy on the Shapiro-Wilk normality test with SPSS.
After that, you can open the SPSS output window.
Figure 33 shows the output results from the Shapiro-Wilk Test with SPSS.
Figure 33.
Shapiro-Wilk Test Results Output Page with SPSS.
Figure 33 shows the results of the Shapiro-Wilk test calculations with SPSS.
Based on these results, it is known that the significance value (Sig.
) of the Shapiro-Wilk test for the pretest and posttest values is 0.
026 and 0.
160, respectively.
From the output results with a pretest significance value of 0.
026 > 0.
050 and a posttest of 0.
160 > 0.
050, it can be concluded that the data is normally distributed.
Homogeneity Test The homogeneity test is a statistical test procedure that aims to show that two or more groups of data samples taken from a population have the same variance.
A brief definition of the homogeneity test is testing whether the variances of two or more distributions are the same (Dogan et al.
, 2.
As with the normality test, the homogeneity test is used as a reference material for determining decisions on subsequent statistical tests.
The basis or guidelines for decisionmaking in the homogeneity test are as follows (Agus et al.
, 2.
If the significance value or Sig.
< 0.
05, then it is said that the variance of two or more population data groups is not the same or not homogeneous.
If the significance value or Sig.
> 0.
05, then it is said that the variance of two or more population data groups is the same or homogeneous.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 68 In this article, the homogeneity tests discussed are ANOVA and Levene Test.
Below is a description of the steps for the ANOVA and Levene homogeneity tests.
ANOVA
Analysis of Variance (ANOVA) is a statistical analysis that determines the mean differences between groups.
The group in question can mean a group or type of treatment.
ANOVA itself is a statistical test procedure that is similar to the t-test, but ANOVA can test differences in more than two groups.
This is different from the independent sample t-test which can only test the difference in the means of two groups (Ustaoglu et al.
, 2.
ANOVA is used as an analytical tool to test research hypotheses to assess whether there are differences in means between groups.
The final result of the ANOVA analysis is the F test value or calculated F.
The basis for decision-making for the ANOVA homogeneity test carried out by comparing the F table and calculated F is as follows:
If the calculated F value > F table, then H1 is accepted or H0 is rejected, which means there is a difference in the averages for all groups.
On the other hand, if F count < F table then H1 is rejected or H0 is accepted, which means there is no difference in averages in all groups.
The data presented in this article is quoted from research conducted by Paul & Jefferson .
We present a homogeneity test on the learning outcomes of class A and B students (See Table .
In this context, we test homogeneity using ANOVA.
Table 4.
Example of ANOVA homogeneity test research data .
ot homogeneous result.
Science learning outcomes Class A Class B DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 69 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Table 4.
(Continu.
Example of ANOVA homogeneity test research data .
ot homogeneous Science learning outcomes Class A Class A Below we explain the steps for the ANOVA homogeneity test using the SPSS application:
Open the SPSS application, then click in the lower left corner "View Variables".
After clicking, the next step is to fill in the properties of the research variable as seen in Figure To create a variable, write Class A and Class B in rows 1 and 2.
In the AuDecimalAy column, change it from 2 to 0.
After that, the AuMeasureAy columns for Class variables A and Class B change from unknown to scale.
Figure 34.
Display the AuData ViewAy worksheet settings in ANOVA Test with SPSS.
To fill in the variables in the "Values" section, click the "None" column until the "Value Label" dialog box appears, in the "Value" box, enter number 1, in the "Label" box, enter Class A, then click Add.
Then, fill in the "Value" box again with the number 2, and in the "Label" box write Class B, then click Add.
If you have successfully filled in the variable properties correctly, the next step is to click the "OK" button.
If the process of filling in all variable properties is carried out correctly, the display in SPSS will look like Figure 35.
After that, you click "Data View" and then enter the Class data into the Class box, and the learning results data into the Learning box.
Then, you can carry out the analysis process by going to the "Analyze" menu, then clicking "Compare Means", and the next step is clicking "One Way ANOVA" as shown in Figure 36.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 70 Figure 35.
Variable view worksheet settings: Value labels ANOVA Homogeneity Test with SPSS.
Figure 36.
Settings for calculating the ANOVA homogeneity test with SPSS.
After you have completed step 3, a box will appear with the name "One-Way ANOVA".
Next, enter the variable "Science Learning Outcomes" into the "Dependent List" box and enter "Class" into the "Factor" box, then click options as shown in Figure 37.
After clicking options, you will be directed to the "One-Way ANOVA: Options" dialog box, in the "Statistics" section, put a tick for descriptive and Homogeneity of variance test, then click continue as shown in Figure 38.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 71 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 37.
One-Way ANOVA Setup.
Figure 38.
One-Way ANOVA Settings: Statics.
Click AuOKAy to end the command.
After carrying out stage 5, the SPSS output display will appear as shown in Figure 39.
Figure 39.
Output page for ANOVA homogeneity test results with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 72 Based on Figure 39, the significance value (Sig.
) of the data used is 0.
Meanwhile, the F table value is 7.
From the output results with a significance value (Sig.
) of 0.
007 > 0.
it can be concluded that the data used is not homogeneous.
Meanwhile, the output results of the F value and F table value are 7.
967 > 3.
130, thus it can be concluded that the pretest and posttest value data have different average values.
2 Levene Test Levene's test is a method of testing the homogeneity of almost equal variants.
Levene's test is carried out by looking for the difference between each data and the group average.
Levene's test is better used if the amount of data between groups is the same.
Levene's test means that data does not have to be normally distributed, but must be continuous.
As an example of research conducted by Ananda & Atmojo .
, we present a homogeneity test on chemistry learning outcomes data for students in classes A and B.
Details of student learning outcomes data are presented in Table 5.
Table 5.
Example of Levene homogeneity test data .
omogeneous result.
Chemistry learning outcomes Class A Class B More clearly, the steps for carrying out the Levene test using SPSS are as follows:
Open the SPSS application, then click in the lower left corner "Variable View".
Then, in the "Name" column.
Class and Learning Outcomes are written in rows 1 and 2.
Next, the AuDecimalsAy column is changed from 2 to 0.
After that, in the AuMeasureAy column for the DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 73 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Class variable, it is changed from unknown to nominal and the Learning Outcomes variable is changed to scale (See Figure .
Gambar 40.
Variable settings view homogeneity test: Levene Test with SPSS.
After that, to fill in the variables in the "Values" section, click the "None" column, and the "Value Label" dialog box will appear, in the "Value" box, fill in number 1, and in the "Label" box, fill in Class A then click "Add".
hen, fill in the "Value" box again with the number 2, and in the "Label" box write Class B, then click Add.
If filling in the variable properties has been successfully done correctly, the next step is to click the "OK" button.
If the process of filling in all variable properties is carried out correctly, the display in SPSS will appear as in Figure 41.
Figure 41.
Setting variable values in the Levene Test.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 74 .
The next step is to calculate the homogeneity test using the Levene test.
Start by clicking the SPSS main menu, then selecting the "Analyze" menu.
After that, click the Descriptive statistics submenu then click "Explore" (See Figure .
Figure 42.
Settings for calculating homogeneity tests with SPSS.
After you have completed step 3, you will be directed to a box called "Explore".
Next, move the "Chemistry Learning Outcomes" variable to the "Dependent List" box and move the "Class" variable to the "Factor" box, then click options as shown in Figure 43.
Figure 43.
Settings for calculating the Levene test of homogeneity with SPSS.
After you have completed step 4.
The next step is to click "Plots", a dialog box will appear then in the Boxplots box click "Factor level Together" and in the "Descriptive" box check "stem-and-leaf".
After that, in the Speard vs level with the Levene test box, click Aupower estimationAy and click AuContinueAy as shown in Figure 44.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 75 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 44.
Settings for calculating homogeneity tests.
Click "Continue" in the explore plots the Levene test box with SPSS.
To end the command, click "OK" and the SPSS output display will appear as shown in Figure Figure 45.
Output page for Levene test homogeneity test results with SPSS.
Based on Figure 45, the significance value (Sig.
) of the data obtained is 0.
The output results with a significance value (Sig.
) of 0.
933 > 0.
05, meaning that the data used is homogeneous which indicates that the average value of class A and class B is the same.
Example of How to Calculate Normally Distributed Data (Parametric Statistic.
to Test the Difference in Significance One Sample t-test The one-sample t-test is generally used to compare the average of the sample being studied with the average of the existing population.
The one-sample t-test can be used to test hypotheses in the form of descriptive statistics if the research data is on an interval or ratio The One Sample t-test is a comparison test to assess the difference between a certain value and the average of a population group.
The one-sample t-test is also called the student t-test or one-sample t-test because the t-test here uses one sample.
One-sample t-test is a part of parametric statistics.
Therefore, the basic assumption that must be met is that the research data is normally distributed (McGovern et al.
, 2.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 76 For example, citing data from De Winter .
, the data sample was taken with the assumption that the "Average science ability score of vocational school students in Bandung City is more than 75Ay.
To prove this, we chose 30 students in the city of Bandung with their National Science National Examination scores.
Table 6 shows an example of one sample t-test data for 30 students.
Table 6.
Example of One Sample t-test data .
he results are differen.
Average learning outcomes The steps for the one-sample t-test with SPSS based on the data in Table 6 are as follows:
Open the SPSS program, in this study.
SPSS version 26 was used.
Click "Variable View", then enter the variable name such as in Figure 46.
To create a variable, enter the average_ outcomes in the "Name" column.
In the decimal column, change it from to 0.
After that, change the "Measure" column from the unknown to scale.
Next, click "Data View" in the lower left corner.
After the data view page is visible, enter the data on the average value of learning outcomes of 30 respondents which has been prepared previously (See Figure .
When carrying out a one-sample t-test using SPSS version 26, the first point that needs to be done is to test the normality of the data.
The method for testing data normality is the same as the previous demonstration in the previous discussion regarding normality The normality test results from the average learning outcome data for 30 students are shown in Figure 48.
Based on the results in Figure 48, the Shapiro-Wilk Sig value is 0.
064 > 0.
Thus, it can be concluded that the average data on learning DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 77 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 outcomes for 30 students is normally distributed.
Thus, the assumption of normality in the one-sample t-test has been fulfilled.
After that, you will carry out the One Sample t-test.
The way to do this is to click the "Analyze" menu on the toolbar, then select AuCompare MeansAy in the dropdown menu and click AuOne Sample T-Test.
Ay (See Figure .
The next step is the "One-Sample T-Test" dialog box (See Figure .
Figure 46.
Display the AuData ViewAy worksheet settings in the one sample t-test with SPSS.
Figure 47.
Display the AuData ViewAy worksheet settings in the one sample t-test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 78 Figure 48.
Normality test results in the independent sample t-test Figure 49.
Settings for calculating the One Sample t-test with SPSS.
Figure 50.
One-Sample T-test dialog box.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 79 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 .
In the "One-Sample T Test" dialog box, you must enter the learning outcome variable [Result.
into the Test Variable.
box, in the "Test Value" box type 75.
The value 75 is entered in the test value field because in research We want to compare the average score of student learning outcomes with a score of 75.
Then in the final step, click the "OK" (See Figure .
Figure 51.
Settings for the "One-Sample T-Test" dialog box in the one sample t-test with SPSS.
After that, the SPSS Output windows page will appear as shown in Figure 52.
Figure 52 shows the descriptive statistical values and the One-Sample test results table.
Based on the results, it is known that N = 30 means the number of samples used was 30 students.
Mean = 68.
67, meaning the calculated average value is 68.
Std.
Deviation or standard deviation has a Std value.
Deviation of 6.
042 and Std.
The Mean Error is 1.
Thus, the results of the One Sample Test show that the calculated t-value is -0.
The df .
egree of freedo.
value is 29.
The Sig.
-taile.
or significance value with a two-sided test Figure 52.
Output results of the One Sample t-test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 80 If we formulate the research hypothesis in the one sample t-test as follows:
H0 = The average science ability score of vocational school students in Bandung City is equal H1 = The average value of science ability of vocational school students in Bandung City is more The basis for decision-making for the One Sample T Test is .
If the Sig.
-taile.
< 0.
05, then H0 is rejected .
If the Sig.
value, .
-taile.
> 0.
05, then H0 is accepted.
Based on the "One-Sample Test" output table in Figure 52, it is known that the Sig value, .
-taile.
00 > 0.
05, so following the basis for decision-making above it can be concluded that H0 is accepted.
Thus, it can be interpreted that the average science ability score for vocational school students in Bandung City is greater than 75.
2 Independent sample t-test The independent sample t-test is a comparative test or difference test to find out whether there is a significant difference in the mean between 2 independent groups with interval/ratio The two independent groups referred to here are two unpaired groups, meaning that the data sources come from different subjects.
Some assumptions must be met in the independent t-test namely (Obafemi, 2.
, .
Interval or ratio data scale .
Independent or unpaired data groups, data per group is normally distributed .
The data per group does not have outliers, and the variance between groups is the same or homogeneous.
To help with comprehension, we've included a sample set of data below that includes 60 respondents split into two groups: the control class and the experimental class.
According to research by Susanti and Rustam .
, the information offered refers to the mathematics learning results of class Vi students .
ee Table .
Table 7.
Example of Independent t-test Data .
esults are not differen.
Mathematics learning outcomes Class Pretest Posttest Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 81 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Table 7 (Continu.
Example of Independent t-test Data .
esults are not differen.
Mathematics learning outcomes Class Class Class Control Control Control Control Control Control Control Control Control Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment Experiment The steps for carrying out an independent t-test using SPSS:
Open the SPSS application.
Once the SPSS worksheet is open, click the bottom left corner "Variable view".
Then in the "Name" column enter the Class.
Pretest, and Posttest Then, in the "Measure" column the pretest and posttest variables from the unknown to scale, the Class variable from the unknown to nominal.
In the "Decimals" column it is changed from 2 to 0 (See Figure .
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 82 Figure 53.
Display the AuVariable ViewAy worksheet settings in an independent samples t-test with SPSS.
Next, fill in the variables in the Values section by clicking the AuNoneAy column until the Value label dialog appears.
Then, in the "Value" box, enter number 1.
In the "Label" box, enter Class Control, then click Add.
Then, fill in the "Value" box again with the number 2, and in the "Label" box write Class Experiment, then click Add.
If you have successfully filled in the variable properties correctly, the next step is to click the "OK" button.
If the process of filling in all variable properties was carried out correctly, the display in SPSS will appear as in Figure 54.
If the value label has been filled in, results will appear as in Figure 55.
Figure 54.
Display the AuVariable ViewAy worksheet settings in an independent samples t-test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 83 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 55.
Data input results in the independent sample t-test with SPSS.
Before carrying out an independent sample t-test using SPSS version 26, the first point that needs to be done is to test the normality of the data.
The method for testing data normality is the same as previously explained in the previous chapter regarding normality testing.
After that, the output results will appear as shown in Figure 56.
Based on Figure 56, the output results of the pretest and posttest Shapiro-Wilk normality tests each have a significance value above 0.
This shows that the data is normally distributed.
Figure 56.
Normality test results in the independent sample t-test.
After the data is normally distributed, an independent sample t-test can be carried out.
The way to do this is to click the "Analyze" menu on the toolbar, then select Compare Means in the dropdown menu and click "Independent-Samples T-test" as shown in Figure DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 84 Figure 57.
Menu access steps for independent sample t-test with SPSS.
After clicking "Independent-Samples T Test" a display will appear as shown in Figure 58.
Enter the pretest and posttest results in the Variable test column then in the variable grouping column select "Class (? ?).
If so, click Define Group .
ee Figure .
In "define group" select "use specified values" with Group 1 filled with the number 1 and group 2 filled with the number 2.
After that click "Continue" and to to give the final command-click AuOKAy.
Figure 58.
Independent-Samples T Test in Boxes.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 85 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 59.
Settings for the "Independent-Samples T-test: Define Groups" dialog box in the independent sample t-test with SPSS.
After clicking "OK" in step 5, the SPSS calculation results display will appear in the form of a t-test calculation table .
ee Figure .
Based on Figure 60, there significance value .
of the pretest data .
-taile.
640, while the significance value .
-taile.
of the posttest data is 0.
Figure 60.
Output results of independent sample t-test test data calculations.
Here, we formulate the research hypothesis in an independent sample test as follows:
H0 = There is no difference in the average value of mathematics ability of class Vi students between control and experiment H1 = There is a difference in the mathematics abilities of class Vi students between control and experiment.
The basis for decision-making for the Independent t-test is If the Sig.
-taile.
< 0.
05, then H0 is rejected On the contrary.
Sig.
-taile.
> 0.
05, then H0 is accepted.
Based on the output of the "Independent t-test" table in Figure 60, it is known that the Sig .
-taile.
640 > 0.
05 and posttest values 0.
596 > 0.
050, respectively, by the basis for decision making above it can be concluded that H0 is accepted.
Thus, it can be interpreted DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 86 that There is no difference in the average value of mathematics ability of class Vi students between control and experiment.
3 Paired sample t-test A paired t-test is a parametric test that can be used on two different statistics from two paired samples.
The paired sample t-test generally takes the form of interval or ratio scale data .
uantitative dat.
Paired samples are the same subjects, but experience different This different test model is used to analyze the research model before and after.
The conditions for the paired sample t-test are (Alfajri et al.
, 2.
The data owned by the subject is interval or ratio data Both groups of paired data are normally distributed As an example, we present a research demonstration of the steps of a paired samples ttest from a study conducted by Alvi & Yerimadesi .
The data described is the result of class XI students learning acid-base material through the use project-based learning method .
ee Table .
Table 8.
Example of research data Paired Sample t-test .
he results are differen.
Respondent Pretest Posttest The steps for the paired sample t-test using SPSS are explained as follows:
Start the SPSS program.
Open the SPSS worksheet and select "Variable view" in the lower left-hand corner.
Then, insert the Class, pretest, and posttest variables in the "Name" DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 87 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Then, the class variable from the unknown to nominal, and the pretest and posttest variables from the unknown to scale, are listed in the "Measure" column.
In the "Decimals" column it is changed from 2 to 0 .
ee Figure .
Figure 61.
Display the AuVariable ViewAy worksheet settings in a paired samples t-test with SPSS.
After that, click "Data View" and then enter data in the respondent, pretest, and posttest After the data is filled in, the pretest and posttest scores from the 30 respondents as shown in Figure 62.
Figure 62.
Data input results Paired sample t-test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 88 .
The data must first be checked for normality before doing an independent sample t-test using SPSS version 26.
The procedure for determining if data is normal is the same as that described in the preceding chapter's section on normality testing.
The output data will then appear as illustrated in Figure 63.
According to Figure 63, the output findings of the pretest and posttest Shapiro-Wilk normality tests each have a significance value above 0.
This demonstrates that the data are normality distributed.
Gambar 63.
Output results of the paired sample t-test normality test with SPSS.
A paired sample t-test can be performed once the data has been established to be normally This can be accomplished by selecting "Analyze" from the toolbar, choosing "Compare Means" from the dropdown menu, and then selecting "Paired-Samples T-test" as shown in Figure 64.
Figure 64.
Settings for calculating the Paired Sample t-test test with SPSS.
After clicking "Paired-Samples T Test" a box will appear as shown in Figure 65.
Enter the pretest and posttest variables in the Paired Variables box.
The pretest is entered into AuVariable 1Ay .
eft sid.
and the posttest is entered into AuVariable 2Ay .
ight sid.
, then next click "OK".
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 89 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 65.
Paired-samples t test box.
The SPSS calculation results display will show up as a t-test calculation table once you select "OK" in step 5 .
ee Figure .
According to Figure 66, the pretest-posttest data's significance value .
-taile.
Figure 66.
Paired Sample t-test Results Output Page with SPSS.
The formulation of the research hypothesis in the peered sample t-test is as follows:
Ho = There is no difference in the pretest and posttest of class XI students on the acid-base material project-based learning method.
H1 = There is a difference in the pretest and posttest of class XI students on acid-base material using a project-based learning method.
The basis for decision-making in the One Sample T Test is .
If the Sig.
-taile.
< 0.
05 then H0 is rejected .
Conversely, if the Sig value.
-taile.
> 0.
05, then H0 is accepted.
Based on the output of the "Paired-Samples T Test" table in Figure 66, it is known that the Sig value, .
-taile.
000 < 0.
050, thus by the basis for decision-making above it can be concluded that H0 is rejected.
Thus, there are differences in the pretest and posttest of class IX students on acid-base material using a project-based learning method.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 90 4 Example of How to Process Data not Normally Distributed (Non-Parametric Statistic.
to Test Significance Differences 1 Mann-whitney test The Mann-Whitney Test is a non-parametric statistical method that is used to compare groups of independent sample t-test data with the assumption of normality not being met.
This test is a test used to test two small independent samples (Two Independent Sample Tes.
The Mann-Whitney test requirements are as follows (Happ et al.
, 2.
Sample data is not normally distributed .
Two samples that know each other are not related or have no influence on each other .
ample members of two different group.
Ordinal or interval data scale samples .
The number of samples in both groups is the same Here, we present a demonstration of how to perform the Mann-Whitney test using SPSS.
Data obtained from research conducted by Usman .
The data are the results of science learning for class IX students as many as 60 students who were divided into two classes, namely classes A and B .
ee Table .
Table 9.
Example of Mann Whitney test research data (The results are differen.
Science learning outcomes Class A Class B Code Class B Class A DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 91 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 The detailed steps for the Mann-Whitney test are presented below:
Open a new SPSS worksheet, then click "View Variables" in the bottom left corner.
In the "Name" column, write Class and Learning_outcomes.
Then, in "Measure" the column for the Class variable is changed from unknown to nominal and the Learning_oucomes variable is changed from unknown to scale.
After that, the AuDecimalsAy column is changed from 2 to 0 .
ee Figure .
Figure 67.
Display the AuVariable ViewAy worksheet settings in a Mann-Whitney test with SPSS.
After that, to fill in the variables in the "Values" section, click the "None" column, and then the "Value Label" dialog box will open, in the "Value" box, enter number 1, in the "Label" box, enter Class A, and then click "Add".
then, fill in the "Value" box again with the number 2, and in the "Label" box write Class B, then click Add.
If the variable properties have been appropriately filled in, the next step is to click the "OK" button.
If the process of filling up all variable characteristics is followed successfully, the display in SPSS will appear like Figure 68.
Figure 68.
View variable settings: Values dialog in the Mann-Whitney test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 92 .
After completing step 2.
Next, enter the data.
Click "Data View" then enter data in the class and learning outcomes columns.
After the data is filled in, a display will appear as in Figure 69.
Figure 69.
Data input results from Mann-Whitney test with SPSS.
Previously, to carry out the Mann-Whitney test, the data was checked for normality using SPSS version 26.
However, the data used here was confirmed to be normally distributed.
The method for testing data normality is the same as that explained in the previous chapter regarding normality testing.
The output results of the normality test are presented in Figure 70.
Based on Figure 70, the output findings of the Shapiro-Wilk normality test have a significance value below 0.
This shows that the data is not normally distributed.
Because the data is not normally distributed, statistical analysis of different tests uses the Mann-Whitney test.
Gambar 70.
Output results of the Mann-Whitney normality test with SPSS.
The next stage is to carry out Mann-Whitney calculations using SPSS.
On the toolbar, click the "Analyze" menu then click "Non-Parametric Tests" then click "Legacy Dialogs" then click "2 Independent Samples.
" as shown in Figure 71.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 93 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 71.
Settings for calculating the Mann-Whitney test with SPSS.
After clicking "2 Independent Samples" a box will appear as in Figure 72.
Then, move the Learning_oucomes variable to the "Test Variable List" box and move the Class variable (??) variable to the Grouping Variables box.
After that, click "Define Groups.
" then the "TwoIndependent-Samples Tests" dialog box will appear as in Figure 73, and click "Continue".
Then, to end the command, click "OK".
Gambar 72.
Mann-Whitney test Box.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 94 Gambar 73.
Setting the "Two Independent Samples" dialog box in the Mann-Whitney test with SPSS.
After clicking "OK" the output of the Mann-Whitney test table calculation results will appear which is shown in Figure 74.
Based on the results in Figure 74, the value of Asymp.
Sig.
-taile.
Figure 74.
Mann-Whitney t-test results output page with SPSS.
The following is how the study hypothesis is expressed in the peered Mann-Whitney test:
H0 = There is no difference in learning outcomes for class A students and class B students.
H1 = There is a difference in learning outcomes for class A students and class B students.
Mann-Whitney's decision is made based on the following criteria:
H0 is not accepted if the Sig.
-taile.
< 0.
On the other hand.
H0 is accepted if the Sig value .
-taile.
> 0.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 95 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 The Asymp.
Sig.
-taile.
value is 0.
00 < 0.
05 based on the output of the "Mann-Whitney" table in Figure 74.
Therefore, using the criteria described above, it can be said that H0 is As a result, there are differences in learning outcomes for classes A and B.
2 Wilcoxon test The Wilcoxon test is an alternative test to the paired sample t-test if it does not meet the normality assumption.
This test is also known as the Wilcoxon Match Pair Test.
The Wilcoxon test functions to test differences between paired data, test comparisons between observations before and after treatment, and determine the effectiveness of treatment.
The conditional assumptions of the Wilcoxon test are (Liu et al.
, 2.
The dependent variable is ordinal or interval or ratio data scale but is not normally .
The independent variable consists of two categories which are paired .
The shape and distribution of data between the two paired groups is symmetrical Here we provide a demonstration of Wilcoxon action research from the Knief & Forstmeier .
Data information in the form of pretest and posttest scores for class X biology for 30 students can be seen in Table 10.
Table 10.
Example of Wilcoxon Test research data .
here are difference.
Biology learning outcomes Pretest Posttest DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 96 Below are the detailed steps for carrying out the Wilcoxon test:
Open a new SPSS worksheet, then select "View Variables" at the bottom left.
Write Pretest and Posttest in the "Name" column.
Next, the Pretest and Posttest variables in the "Measure" column are changed from unknown to scale.
The decimal column is then adjusted from 2 to 0 .
ee Figure .
After you have finished setting up the Variable View worksheet.
Next, click "Data View" and then enter the prepared data.
The Data View display is presented in Figure 76.
Before, the data was examined for normality using SPSS version 26 to perform the Wilcoxon test.
The regularly distributed nature of the data utilized here was nonetheless As was discussed in the last chapter on normality testing, the procedure for determining if data is normal is the same here.
Figure 77 displays the normality test output According to Figure 77, the Shapiro-Wilk normality test's output pretest and posttest findings have a significance level below 0.
As a result, it can be shown that the data is not normally distributed.
Due to the non-normal distribution of the data, the Wilcoxon test is used in the statistical analysis of various tests.
Figure 75.
Display the AuVariable ViewAy worksheet settings in the Wilcoxon test with SPSS.
Figure 76.
Display the AuData ViewAy worksheet settings in the Wilcoxon test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 97 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 77.
Output results of the Wilcoxon normality test using SPSS.
The next step is to use SPSS to do Wilcoxon computations.
As illustrated in Figure 78, select "Analyze" from the toolbar, then "Non-Parametric Tests," "Legacy Dialogs," and "2 Related SamplesA" Figure 78.
Settings for calculating the Wilcoxon test with SPSS.
After clicking "2 Related Samples" a box will appear as in Figure 79.
Then move the variables .
to the "Test Pairs" box.
Then, move the variable .
to column Auvariable 1Ay and .
to column Auvariable 2Ay.
After that, in "Test Type" checklist Wilcoxon and click "OK".
(See Figure .
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 98 Figure 79.
Wilcoxon test dialog box.
The output of the Mann-Whitney test table calculation results will appear after clicking "OK," as illustrated in Figure 80.
The value of Asymp.
Sig.
-taile.
00 according to the findings in Figure 80.
Figure 80.
Wilcoxon t-test Results Output Page with SPSS.
The peered Wilcoxon test for the study hypothesis is presented as follows:
H0 = There is no difference between the pretest and posttest biology scores H1 = There is a difference between the pretest and posttest biology scores.
The following factors are used to determine whether to use the Wilcoxon test:
H0 is not accepted, if the Sig.
-taile.
< 0.
On the other hand.
H0 is accepted if the Sig value .
-taile.
> 0.
According to Figure 80, the output result of the Mann-Whitney test calculation is 0.
05, meaning that H0 is rejected thus it can be concluded that there is a difference in the biology pretest and posttest scores.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 99 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 5 An Example of Processing Difference Test Data Based on Real Applications in the Field of Education To understand further how to carry out a statistical difference test .
-tes.
, we provide a step-by-step example of a real case of class IX Islamic Middle School students which is presented in determining understanding of the concept of a steam engine as an electricity energy using conventional methods and experimental demonstrations 60 students were divided into two classes, namely control and experimental classes.
Data obtained in real cases in the form of pretest and posttest results are presented in Table 11.
Here, the different analysis test that we use is the independent sample t-test statistical test because we want to examine the difference between the use of conventional learning methods and experimental demonstrations on student learning outcomes.
All analyses were carried out using SPSS.
Table 11.
Experimental and control data from class IX students.
Score (%) Control Class
Respondent Pre-test
X10
X11
X12
X13
X14
X15
X16
X17
X19
X20
X21
X22
X23
X24
X25
X26
X27
X28
X29
X30
Post-test
Experiment Class
Respondent Pre-test
Post-test
Y10
Y11
Y12
Y13
Y14
Y15
Y16
Y17
Y18
Y19
Y20
Y21
Y22
Y23
Y24
Y25
Y26
Y28
Y29
Y30
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 100 The steps for carrying out data analysis are in Table 11 using SPSS:
Open the SPSS worksheet.
Then click "Variable View", in the "Name" column write pretest, posttest, and class.
The decimal column was changed from 2 to 0 and the "Measure" column for the pretest and posttest variables was changed from unknown to scale.
Meanwhile, the class variable in the "Measure" column is changed from unknown to nominal (See Figure .
Figure 81.
Display the AuVariable ViewAy worksheet settings.
After that, fill in the variables in the Values section by clicking the None column until the Value label dialog appears.
Then, in the "Value" box, enter number 1.
In the "Label" box, enter Class Control, then click Add.
Then, fill in the "Value" box again with the number 2, and in the "Label" box write Class Experiment, then click Add.
If you have successfully filled in the variable properties correctly, the next step is to click the "OK" button.
If the process of filling in all variable properties is carried out correctly, the display in SPSS will appear as in Figure 82.
If the value label has been filled in, results will appear as in Figure 83.
Figure 82.
Display the AuVariable ViewAy: Dialog value labels worksheet settings.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 101 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 83.
Display the AuData ViewAy worksheet settings.
Before carrying out the difference test analysis, first carry out a normality test to see the distribution of the data.
To perform normality test calculations with SPSS, click the "Analyzer" menu on the toolbar then click "Descriptive Statistics" on the submenu and click "Explore" (See Figure .
Figure 84.
Settings for calculating the normality test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 102 .
After you have finished setting up the normality test calculation method, the "Explore" box will appear as in Figure 85.
Then, move the pretest and posttest variables to the "Dependent List" box, while the class variables are moved to the "Factor List" box.
If thus, then click on the right side of "Plots" and a box will appear as in Figure 86.
After that, check "Normality plots with test" and click "Power estimation" then click "Continue".
To end the command, click "OK".
To test normality, the Shapiro-Wilk test was used because the sample was less than 50, where each class, namely the control class and the experimental class consisted of 30 students.
After that, the normality test windows output results appear.
Figure 87 shows the output results of the normality test with SPSS.
Based on Figure 87, the Sig.
Shapiro Wilk > 0.
which indicates that the data in this study is normally distributed.
Figure 85.
Explore settings on the normality test with SPSS.
Figure 86.
Explore: Plots settings in the normality test with SPSS.
DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 103 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Figure 87.
Normality test results output page with SPSS.
Because the research data is normally distributed, independent sample t-test analysis can be carried out.
Calculation of the independent sample t-test with SPSS is done by clicking "Analyze" then clicking "Compare Mean" and clicking "Independent-Samples T-test".
(See Figure .
Figure 88.
Settings for calculating the Independent sample t-test with SPSS .
After clicking "Independent-Samples T Test" a display will appear as shown in Figure 89.
Enter the pretest and posttest results in the "Test Variable.
" column then in the variable grouping column select "Class (? ?).
If thus, click Define Group .
ee Figure .
In "Define Group" select "Use specified values" with Group 1 filled with the number 1 and Group 2 filled with the number 2.
After that click "Continue" and click "OK" to end the command.
The output of the results of the Independent Sample t-test table calculation will be displayed after pressing "OK," as illustrated in Figure 91.
Based on the findings in Figure 91, it show that Asymp.
Sig.
-taile.
pretest and posttest each have values of 0.
009 and DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 Fiandini et al.
How to Calculate Statistics for Significant Difference Test Using SPSS A | 104 Figure 89.
Independent-Samples T test dialog box Figure 90.
Setting the dialog box "Independent samples t-test: Define Groups" in the independent sample t-test with SPSS Figure 91.
Output results of Independent Sample t-test test data calculations.
The study hypothesis is stated as follows in the standard Independent sample t-test:
H0 = There is no difference in learning outcomes using demonstration and conventional experimental methods .
H1 = There are differences in learning outcomes using demonstration and conventional experimental methods DOI: https://doi.
org/10.
17509/ijost.
p- ISSN 2528-1410 e- ISSN 2527-8045 105 | Indonesian Journal of Science & Technology.
Volume 9 Issue 1.
April 2024 Hal 45-108 Based on Figure 91, the output results of the independent sample t-test calculation of the pretest and posttest values are 0.
009 and 0.
From these results, it is known that both values are below 0.
050, which means H0 is rejected and H1 is accepted.
This shows that there are differences in learning outcomes using forced and conventional experimental methods.
Based on these findings, it can be concluded that the experimental demonstration method is the best compared to conventional methods.
CONCLUSION
This article discusses the importance and ways of carrying out mean difference tests in statistical analysis.
The mean difference test is a useful tool to determine whether there are significant differences between two or more groups of data.
Apart from that, this article also discusses practical steps on how to carry out a mean difference test using various statistical software, such as IBM SPSS.
This includes the steps to design an experiment, collect data, select appropriate statistical methods, and interpret the results of a mean difference test.
Overall, this article provides important insights into the role of the mean difference test statistic in data analysis and how to carry it out correctly, thereby helping researchers and data analysts make stronger and more relevant conclusions in various research contexts.
AUTHORSAo NOTE The authors declare that there is no conflict of interest regarding the publication of this The authors confirmed that the paper was free of plagiarism.
REFERENCES