Statistical analysis and data processing with software - which statistical tests do statisticians use?
At Statisticaia we provide a comprehensive answer to all the statistical analysis and data needs of the academic community. Our team of experts accompanies students and researchers at every level – from seminar papers, theses and doctoral dissertations to articles intended for publication in international journals. We help choose the right statistical test, check test assumptions, run the test and build statistical models.
Analyses are carried out in a variety of software packages according to the needs of the required analysis, including: SPSS, R, JASP, Stata.
When needed, we can help write up results in tables and graphs according to academic writing conventions.
Examples of statistical analyses we performed for previous clients
Descriptive statistics
Means, standard deviations, range, including distribution and normality tests.
Tests of differences between two groups
t-tests for independent samples, dependent/paired samples, and a single sample.
For variables on an ordinal scale, or when normality assumptions are not met, we use non-parametric tests such as Mann-Whitney and Wilcoxon, or Kruskal-Wallis when there are three or more groups.
Reliability tests
Various reliability tests, including Cronbach’s alpha, split-half reliability, parallel-forms reliability and test-retest reliability, according to the research needs and the type of data.
Effect size tests
Effect size tests (Effect Size) make it possible to assess the practical significance of research results beyond statistical significance alone. While statistical tests examine whether a significant effect exists, effect size examines how strong or important the effect is, and also the chance of reaching significance if the sample were enlarged. Effect size measures depend on the type of analysis, and there are several common measures such as: Cohen’s d, eta squared, R squared, and others. Combining effect size with statistical significance contributes to a deeper and more accurate understanding of the research findings.
Regressions
Regression analyses of various types, according to the research question and the type of variables: simple regression, multiple regression, hierarchical regression, binary logistic regression, multinomial logistic regression, non-linear regression, and regression with dummy variables. In addition, we check the assumption of homoscedasticity and multicollinearity.
We also run moderation and mediation model analyses using PROCESS.
Relationships between variables
Tests of association between variables, including Pearson’s correlation coefficient, Spearman’s correlation coefficient, partial correlations and multiple correlations. In addition, we run chi-square tests (including Cramer’s V and Phi) to examine relationships between nominal variables. For more complex relationships, Canonical Correlation Analysis can also be performed.
Analysis of variance
We run analysis of variance (ANOVA) of various types, according to the study design and the number of variables: one-way, two-way and multi-way ANOVA (Factorial ANOVA), as well as analyses with repeated measures.
When there are several dependent variables, we run multivariate analysis of variance (MANOVA), and analyses of variance controlling for covariates such as ANCOVA and MANCOVA.
In addition, we run post-hoc tests for multiple comparisons, including TUKEY, SCHEFFÉ, Bonferroni and others, according to the requirements of the study.
Advanced analyses
We offer advanced analyses for understanding data structure, including Confirmatory Factor Analysis, Exploratory Factor Analysis, Discriminant Analysis, and Cluster Analysis.
We also perform Survival Analysis, Structural Equation Modeling (SEM), and Principal Component Analysis (PCA), along with other methods suited to the type of study and its data.
What is needed for statistical analysis of a seminar paper?
Carrying out the statistical analysis for a seminar paper is an essential stage of academic work, since it makes it possible to test the research hypotheses and draw conclusions based on the sample data collected. To do so, several important stages must be completed, beginning with checking the reliability of the instruments used – mainly questionnaires. One common test of questionnaire reliability is Cronbach’s alpha, which measures the internal consistency of the items in each measure or scale in a given questionnaire. It is important to check reliability in order to assess how consistent respondents are in answering the research instrument that measures the variables, which should strengthen the validity of the research findings.
Next, descriptive statistics describing the characteristics of the sample should be presented. This stage includes presenting frequencies (how many participants gave a particular answer), percentages, means, standard deviations and score ranges. These data are usually presented in clear tables and sometimes also in figures or charts, to make it easier for the reader to understand the profile of the participants, which sheds some light on the generalizability of the research to the broader population and to different populations.
The central stage of the statistical analysis is testing the research hypotheses, carried out with suitable statistical tests, such as the t-test, analysis of variance (ANOVA), Pearson correlation, linear regression and more – depending on the type of hypotheses and variables. The purpose of these tests is to determine whether there are significant relationships or differences between the variables also at the population level and beyond the sample obtained, and thus to answer the central research question.
Finally, it is important to note that writing in most academic fields follows APA (American Psychological Association) rules, which dictate how statistical results should be reported, how to format tables, and how to word conclusions clearly and consistently. Following these rules gives the work a professional appearance and helps readers understand and evaluate the research findings.
What is needed for statistical analysis of a master's thesis?
When writing a master’s thesis, the statistical analysis is not merely a technical step in the structure of the work – it is a direct expression of the depth of research thinking and the understanding of advanced methodological principles. This stage is intended not only to describe the data, but to systematically point out patterns, relationships and trends, and to test the research claims on a valid and reliable empirical foundation.
One of the first tasks in the analysis is assessing the reliability of the measurement instrument, especially when the study is based on questionnaires or scales that measure attitudes and perceptions. Calculating Cronbach’s alpha makes it possible to examine the internal consistency of each measure, and sometimes a check of construct validity is also included as part of evaluating the instrument. This analysis helps determine whether the measures can be trusted to faithfully represent the theoretical variables at the focus of the study.
Next, a review of the sample characteristics is carried out in order to present the reader with the outline of the research group: demographic variables, attitudes, behaviors or other relevant characteristics. Descriptive statistics are applied here, including measures such as frequencies, percentages, means and standard deviations, as well as value ranges. Sometimes this analysis must also present early signs of sample bias, or detect outliers and distribution characteristics and test normality.
The central stage of the analysis is devoted to testing the research hypotheses, matching the analysis method to the type of research question. In simple cases correlations or tests of differences (such as t or ANOVA) will be used, but many theses also use more advanced tools – such as multiple linear regression models or hierarchical and stepwise models, logistic regression, moderation and mediation analyses, analysis of variance controlling for COVARIATES – ANCOVA, and measurement of indirect effects (mediation). When there are intervening or structural variables, complex models such as SEM or confirmatory factor analysis (CFA) can even be combined, according to the nature of the study and the questions investigated.
Finally, care must be taken to present at a high academic level, usually according to APA rules – including correct use of statistical terms, full reporting of p values, effect sizes and degrees of freedom, presenting tables in a standard format, and a clear and consistent explanation of the findings even for a reader who is not an expert in statistics. Precision in statistical writing reflects not only professionalism but also respect for the data and for the principles of academic research.