STATISTICAL METHOD
Cluster Analysis
Statistical Consulting for Grouping Cases
We choose the distance, scaling and algorithm to fit your variables, settle the number of clusters through several validation indices and stability checks, and report clusters as descriptive groupings, not proven types.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
METHOD OVERVIEW
What Is Cluster Analysis?
Cluster analysis is an unsupervised method that groups cases, such as respondents, patients or products, so that cases in the same group are more alike on the measured variables than cases in different groups. Hierarchical clustering builds nested groupings from a distance matrix and a linkage rule, while k-means partitions cases around a chosen number of centers. The result depends on the distance, the scaling, the algorithm and the number of clusters, and the groups describe the sample, not established types. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, provides cluster analysis within your study design, with the relevant checks, estimates, uncertainty measures, interpretation and publication-ready reporting.
WHEN TO ENGAGE
When Is Grouping Cases Appropriate?
Clustering fits a descriptive question about whether cases form recognizable groups, and these are the research situations in which researchers most often ask us to apply or confirm it.
Segmenting Survey Respondents Respondents or products are described by several attributes, and the team wants a small number of interpretable segments, each with a profile it can report.
Exploring Patient Subgroups Patients or cases may fall into subgroups on several measures, and an exploratory grouping is needed to describe and then to examine in new data.
Mixed Variable Types Variables mix continuous, ordinal and categorical measures, so a suitable dissimilarity, such as Gower's, must be chosen before any algorithm is run.
Reviewing a Completed Analysis A solution was chosen from one index on unscaled variables, and a reviewer has questioned it. Independent Statistical Review evaluates the stability and the labels.
ANALYTICAL APPROACH
How We Apply Clustering Methods
Within the family of Multivariate, Scale & Reliability, clustering is applied as a descriptive tool whose every choice is documented and tested.
Distance and Scaling We choose the distance to match the variable types, using Gower's dissimilarity for mixed data, and standardize deliberately, since different scales can change distance-based clusters completely.
Algorithm and Linkage K-means suits many cases with continuous variables, while hierarchical clustering shows nesting and allows different linkage rules, such as Ward's or average linkage. We compare them and handle outliers.
Number of Clusters We compare silhouette widths, the gap statistic and other indices with interpretability, because one index alone often gives an unstable solution, and we examine stability across resamples.
Meaning and Validation Clusters are descriptive groupings, not real types. We do not test cluster differences on the variables used to form them, and we advise confirmation in a separate sample or by external criteria.
DELIVERABLES
What You Receive
As part of our Quantitative Data Analysis service, the grouping is delivered with the checks and files needed to report and defend it.
Data & Variable Review Variable types, scaling, outliers, missing values and correlated variables that would give one dimension extra weight are reviewed, and the distance measure is documented before clustering.
Method-Specific Checks Candidate solutions are compared by silhouette widths, gap statistics and stability under resampling, and sensitivity to the algorithm, linkage rule and outliers is examined.
Statistical Results Cluster sizes, centers or medoids, profiles on the clustering variables and on external variables, silhouette summaries, and a dendrogram or cluster plot for the chosen solution.
Publication-Ready Tables & Figures Dendrograms, silhouette plots, cluster profile tables and ordination plots, formatted to APA 7 or the journal's requirements, ready to insert into the manuscript.
Interpretation Clusters are described through their profiles and stability, with any names kept tentative, and presented as a grouping of this sample, not as a discovered population type.
Reproducible Analysis Files R scripts, Python notebooks, SPSS syntax or SAS code are provided where applicable, so the solution and its random starts can be rerun if the data change.
ALTERNATIVE METHODS
When Another Method May Be More Appropriate
We choose the method by what the grouping must do. Clustering finds groups without a model, and these methods fit probabilistic groups, known groups or a variable-level summary.
01
Latent Class Analysis (LCA)
When categorical indicators are thought to reflect hidden classes, a model-based approach gives class probabilities and fit statistics that distance-based clustering does not.
02
Latent Profile Analysis (LPA)
When continuous indicators may define profiles, a mixture model handles unequal variances and assigns membership probabilities instead of forcing each case into one group.
03
Discriminant Analysis
When the groups are already known and the aim is to classify new cases or describe what separates them, a supervised method replaces unsupervised grouping.
04
Principal Component Analysis (PCA)
When summary scores of many correlated variables are the goal, components provide them, and any clustering of the scores is checked against the full variables.
CASE EXAMPLE
How We Work in Practice
An illustration of how a grouping question moves from a planned method to a reportable, defensible result.
ILLUSTRATIVE EXAMPLE · NONPROFIT SURVEY RESEARCH · UNITED STATES
A nonprofit survey center collected responses to eleven questions on attitudes and behaviors and wanted respondent segments to guide how it reported its community findings.
The plan, fixed before any solution was seen, used Gower dissimilarity because the questions mixed rating scales and yes-no items, with partitioning around medoids and solutions of two to eight clusters compared by silhouette width and resampling stability. One small cluster consisted of respondents who gave the same answer to every rating item, which we flagged as probable straight-lining and examined with and without those cases. The remaining solution was described through profiles, named tentatively and not tested on the variables that formed it.
Outcome: The center received silhouette widths and stability for two to eight clusters, profiles of the chosen solution, the analysis without flagged respondents, plots, the scripts and a methods paragraph.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
SOFTWARE
Statistical Software for Grouping Cases
Clustering is run in the statistical software researchers already use, with syntax or scripts provided where applicable.
RESEARCH DISCIPLINES
Grouping Questions Across Research Fields
Unsupervised segmentation arises wherever cases are measured on many attributes, and each field has its own variables, cluster validation standards and caution about labels.
A Defined Process for Your Analysis
We review your dataset, research question and design before any work begins, so the scope, deliverables, timeline and fixed project fee are clear from the outset.
01
Share Your Research Materials
Send your dataset, research question, study design and any existing analysis plan or output. We will let you know if anything else is needed to assess the project.
02
Receive Your Assessment
Your statistically trained Research Coordinator reviews the project and confirms the required analysis, deliverables, timeline and fixed project price before work begins.
03
Analysis & Delivery
Once you confirm the scope, an in-house statistician completes the agreed analysis and delivers the outputs, interpretation and reporting files.
RESEARCHER VOICE
What Researchers Say
Feedback from a researcher who found the structure of the data, not the software step, to be the difficult part, and valued an approach that caught the problem early.
“The difficult part was not pressing the button for Cluster Analysis; it was the clustered structure of the data rather than treating every row as independent. They caught the problem early, showed us the defensible approach, and kept the reporting focused on the estimates that mattered. The final output was concise, clearly interpreted, and ready to use in our research reporting.”
RESEARCH RESPONSIBILITY
Statistical Support Without Taking Over Your Research
We provide the statistical work for your grouping of cases, and the study, its meaning and its authorship stay entirely with you.
Agreed Statistical Support We provide the agreed methodological, analytical and statistical reporting support for your grouping of cases, within the scope confirmed before work begins.
No Intellectual Contribution OnlineSPSS claims no intellectual contribution to your research. Our work is statistical support, and authorship remains yours.
Your Research Decisions The research question, any name given to a cluster, interpretation and conclusions remain with you, and our statistical interpretation stays within what the data and the design support.
Confidential Handling Research data and project details are handled confidentially and are not published. An NDA is available where required, and can be arranged before sensitive materials are shared.
FREQUENTLY ASKED QUESTIONS
Questions About Clusters and Segments
Short answers to what researchers ask before grouping cases, from the number of clusters and the algorithm to confidentiality and the assessment process.
How many clusters should I choose?
No single index decides the number of clusters. We compare silhouette widths, the gap statistic and other criteria across a range of solutions, check how stable each one is under resampling, and weigh whether the groups are interpretable and large enough to use. When indices disagree, which is common, we report the alternatives and the reasoning, not one number.
Should I use k-means or hierarchical clustering?
K-means is efficient with many cases and continuous variables, but it needs the number of clusters in advance, is sensitive to outliers and favors compact, similar-sized groups. Hierarchical clustering shows how cases merge and works from any distance, including Gower’s for mixed data, and suits smaller samples. We often compare both and report how much the solution depends on the method.
Can I test whether my clusters differ significantly?
Not in the usual way on the variables used to form them. The algorithm has already maximized the differences, so the p-values are optimistic and are not valid as ordinary tests. Differences on external variables not used in the clustering can be described, with intervals, and confirmation in a separate sample is the stronger check.
Can you review an analysis that has already been completed?
Yes, we can review a completed clustering before submission or a response to reviewers. We check the distance and scaling, the algorithm and linkage, how the number of clusters was chosen, the stability evidence, the treatment of outliers and the wording of the cluster labels, and confirm what stands or needs to change.
How do you protect my dataset and research confidentiality?
Your dataset, output files and any protocol or analysis plan are handled confidentially, and project information is not published. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Need Support With
Grouping Your Cases?
Share your research question, variables and data. We assess whether clustering fits your aim and confirm the method before any work begins.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014