What Is Student Scientific Research and Where Do You Start?
What Is Student Scientific Research
Student scientific research is the process of asking a testable research question, building a model or hypotheses, collecting data, and using an appropriate method to reach a conclusion. For a quantitative project, the final package usually includes a proposal, questionnaire, .sav or .csv dataset, SPSS or SmartPLS output, discussion, and implications.
You can read What is scientific research if you need the broader concept first. Once your data file is open, the immediate task is to identify the question your study answers and the variables used to answer it.
A project usually includes independent variables, dependent variables, and sometimes a mediator or moderator. Suppose you are studying students' intention to use e-wallets. Perceived usefulness and perceived ease of use may be the independent variables, while usage intention is the dependent variable. Each construct is measured with multiple items in a Likert questionnaire, such as a scale from 1, strongly disagree, to 5, strongly agree, following the logic of the scale developed by (Likert, 1932).
The minimum structure of a quantitative study can be written as a sequence: research problem, objectives, questions, model, hypotheses, scales, sample, data, tests, and conclusions. If one link in that sequence is still unclear, do not rush into EFA or regression simply because an SPSS file is already available.
Why Student Scientific Research Matters in Quantitative Research
Research helps you turn a general observation into a question that can be measured. The statement “students may enjoy online learning” is broad. The question “does perceived ease of use affect students' intention to continue using an online learning platform” identifies the analytical direction and the type of relationship to test.
The first step is to identify the research subject, scope, context, and unit of observation. If your population is first-year students at one university, you should not write your conclusion as if the result represents all students in Vietnam. A clearly defined scope makes both sampling and the discussion of limitations easier to defend.
The next step is to identify the research field and find an appropriate scale source. Two projects may both study “satisfaction” while using different contexts, such as banking services, online classes, or shopping on an e-commerce platform. Adjust the wording to fit your context, then assess the scale's reliability and validity using real data.
Quantitative research helps answer specific decisions. You may need to know which factor has the stronger effect, whether a relationship is statistically significant, how much variance in the dependent variable the model explains, or whether the scale distinguishes between constructs. SPSS is suitable for Cronbach's Alpha, EFA, correlation, and regression. SmartPLS is suitable when you build a PLS-SEM model with latent variables and path relationships.
If your question requires understanding experiences or motivations through personal accounts, that is a different direction from the quantitative process covered here. You can read more about what qualitative research is to distinguish the scope, but do not mix the two processes when your proposal requires a questionnaire and numerical data.
What Counts as an Acceptable Student Scientific Research Project
No single number determines whether an entire study is acceptable. A study can have a high Cronbach's Alpha while still using a poorly specified model. Compare each statistic with the purpose of the test, the type of analysis, and the cited source. The table below contains commonly used thresholds in quantitative research, but it does not replace the methodological argument for your project.
| Check | Reference level | Canonical source |
|---|---|---|
| Cronbach's Alpha | 0.7 or above | (Nunnally, 1978) |
| Alpha for a new or exploratory scale | May be 0.6 or above | (Hair et al., 2010) |
| Corrected Item-Total Correlation | 0.3 or above | (Nunnally and Bernstein, 1994) |
| KMO | 0.5 or above | (Kaiser, 1974) |
| Bartlett's Test | Sig. below 0.05 | (Kaiser, 1974) |
| Eigenvalue for factor extraction | Above 1 | (Kaiser, 1960) |
| Total Variance Explained | 50% or above | (Hair et al., 2010) |
| Factor loading in EFA | 0.5 or above | (Hair et al., 2010) |
| Outer loading in PLS-SEM | 0.7 or above | (Chin, 1998) |
| Composite Reliability | 0.7 or above | (Fornell and Larcker, 1981) |
| AVE | 0.5 or above | (Fornell and Larcker, 1981) |
| HTMT | Below 0.85, or 0.90 for closely related constructs | (Henseler et al., 2015) |
| VIF in PLS-SEM | Below 5 | (Hair et al., 2019) |
| R² in PLS-SEM | 0.25 weak, 0.50 moderate, 0.75 substantial | (Hair et al., 2011) |
| Q² | Above 0 | (Stone, 1974) |
Sample size also needs an explanation. For a scale, the rule of 5 to 10 observations per item is commonly cited from (Hair et al., 2010). For regression, the formula n at or above 50 + 8m, where m is the number of independent variables, is stated in (Tabachnick and Fidell, 2013). These are reference points. You still need to consider your sampling method, the number of groups to compare, and your supervisor's requirements.
How to Read Student Scientific Research Output
Suppose you run Cronbach's Alpha for a scale called “Usage Intention” with four items, YD1 through YD4. The table below is illustrative output, not the result of a real study. The column names follow the format SPSS commonly produces, so you can compare them with your own file.
| Reliability Statistics | Cronbach's Alpha | N of Items |
|---|---|---|
| Usage Intention | 0.842 | 4 |
| Item-Total Statistics | Corrected Item-Total Correlation | Cronbach's Alpha if Item Deleted |
|---|---|---|
| YD1 | 0.631 | 0.810 |
| YD2 | 0.704 | 0.778 |
| YD3 | 0.589 | 0.827 |
| YD4 | 0.672 | 0.796 |
In this example, Alpha is 0.842 and the corrected item-total correlation for all four items is above 0.3. Do not delete an item merely because the “Cronbach's Alpha if Item Deleted” column could increase slightly. Check the item's wording, theoretical meaning, and consistency with the remaining items.
If you run EFA, read the KMO and Bartlett's Test, Total Variance Explained, and Rotated Component Matrix tables in sequence. KMO indicates whether the data are suitable for factor analysis. Bartlett's Test checks whether the correlations between items are sufficient to continue. The Rotated Component Matrix then shows which factor each item loads on most strongly.
For regression, the Model Summary table gives R Square and Adjusted R Square, the ANOVA table tests the overall significance of the model, and the Coefficients table gives the coefficients, t, and Sig. In SmartPLS, check Outer Loadings, Construct Reliability and Validity, Discriminant Validity, Path Coefficients, and the Bootstrapping results. Do not include CFI, TLI, or RMSEA in a SmartPLS report because these indices belong to CB-SEM and are presented in the context of (Hu and Bentler, 1999).
What to Do When Student Scientific Research Does Not Meet the Thresholds
Work from the data toward the model instead of deleting groups of items to make the output look better.
Check the Coding and Original Data
Open Variable View and Data View in SPSS. Check whether reverse-coded items have been recoded, whether missing values were entered as 0, and whether identifier columns were accidentally included in the analysis. A reverse-coded item can pull down Alpha and the corrected item-total correlation even when the scale content is appropriate.
Check Each Item
Review Item-Total Statistics, factor loadings, and the item wording. If an item has a low corrected item-total correlation or loads on several factors, record the reason for considering its removal. Do not delete an item only because it lowers a statistic, especially when it represents an essential part of the construct.
Rerun the Analysis in a Controlled Way
If there is a methodological basis, remove one item at a time and run the analysis again. Save the output from each round and record which item was removed and how the statistics changed. It is rare for EFA to give a clean matrix on the first pass. Removing an item and rerunning the analysis can be appropriate, but you must explain the decision using the scale content and the criteria you selected.
Reconsider the Model and Sample
If several scales fail at the same time, the issue may be in the questionnaire, how respondents understood the items, or the sample quality. Software cannot repair the original data. Discuss with your supervisor whether to retain the data, adjust the instrument, or collect additional responses instead of creating extra cases or changing values to meet a threshold.
Student Scientific Research and Research Topics
A “research topic” is the specific subject or problem you choose, such as factors affecting students' intention to buy cosmetics online. “Student scientific research” refers to the full process of carrying out that project, from defining the problem to reporting the results.
A “questionnaire” is the instrument used to collect data. A “scale” measures a construct with one or more items. A “research model” describes the constructs and the expected relationships between them. A “hypothesis” is a directional statement about a relationship to be tested, such as perceived usefulness having a positive effect on usage intention.
You should also distinguish research from a descriptive report. A table showing the percentage of students who use an application only describes the sample or current situation. Once you build hypotheses, measure constructs, and test relationships, the work has a clearer quantitative research structure.
Common Mistakes
The first mistake is choosing a topic that is too broad. “Research on consumer behavior” does not tell the reader who will be surveyed, in what context, or which variables will be measured. Narrow the topic by population, location, time, and research variables. The Scientific Research Topics page can help you review possible directions before finalizing the question.
The second mistake is taking a scale from an arbitrary paper without reading the construct definition. You need to know which construct each item measures, what the original context was, and whether the wording fits your respondents. Review the questionnaire before distributing it widely.
The third mistake is running too many tests without connecting them to the hypotheses. A long output does not make a study stronger. Each table should answer a specific methodological question, such as whether the scale is reliable, whether the model is appropriate, or whether a hypothesis is supported.
The fourth mistake is presenting illustrative output as real results. You can use simulated examples such as the table above when learning how to read an output table. In your thesis, every Alpha, Sig., R², loading, and path coefficient must come from the data file you actually analyzed.
Frequently asked questions
Does student scientific research have to use SPSS?
No. SPSS is suitable for many projects using Cronbach's Alpha, EFA, correlation, and regression. SmartPLS is suitable for PLS-SEM, while AMOS is suitable for CB-SEM. Choose the tool according to your model, scale type, and department requirements, rather than choosing it only because the software is widely used.
How many observations are needed for student scientific research?
The answer depends on the number of items, the number of independent variables, the model, and the sampling method. The rule of 5 to 10 observations per item is cited from (Hair et al., 2010), while regression can refer to n at or above 50 + 8m under (Tabachnick and Fidell, 2013). State the basis for your sample-size decision in the methodology chapter.
Will a Cronbach's Alpha below 0.7 automatically invalidate the scale?
You cannot decide immediately. For a new or exploratory scale, Alpha from 0.6 may be considered under (Hair et al., 2010). Review the Corrected Item-Total Correlation, the wording of each item, the number of items, and the research context before making a decision.
Does student scientific research need hypotheses?
A quantitative study that tests relationships between variables generally needs hypotheses to establish the analytical direction. Each hypothesis should connect to the model, theoretical basis, scale, and results table. If the project only describes the current situation, the hypothesis structure may differ, but the objectives and measured variables still need to be clear.
Can I change the data so the output meets the threshold?
You can clean data according to criteria defined in advance, such as removing a response with too much missing information or identifying an invalid response pattern. You cannot change answers, add fabricated cases, or delete items only to meet a threshold. If the data do not meet the criteria, preserve the processing record and explain the limitation to your supervisor.
Open your .sav or .csv file now. Create a table with the variable name, role in the model, scale, and planned test, then compare each output with the sources listed above. If you need to run the analysis on your own SPSS or SmartPLS data, DoThesis M4 analysis is the module for this step.