RESEARCH SUPPORT
SPSS Secondary Data Analysis
Merged Files, Weights and Survey Designs
Existing datasets prepared and analyzed in SPSS: files joined on a verified key, survey weights and design variables applied as the documentation requires, and estimates reported with confidence intervals and a record of every step.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
What Does SPSS Secondary Data Analysis Involve?
SPSS secondary data analysis answers a new research question with data that someone else collected, such as a national survey, an administrative file or a registry extract. The work begins before any model: files are joined on a verified key, user-defined missing codes are declared, and the documentation is read to learn how the sample was drawn. If the design was complex, the analysis then needs the weights, strata and clusters, which SPSS handles in its Complex Samples procedures, available in some editions or as an add-on module. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, provides this support: an in-house PhD-level statistician prepares and analyzes the agreed files in SPSS and explains each decision. Which dataset answers your question, and what its estimates mean for your field, are decisions you make.
WHEN TO ENGAGE
Moments When an Existing Dataset Needs a Closer Look
Secondary data rarely arrive ready to analyze: files fail to line up, the documentation is opaque, or a reviewer wants to know how the sampling design entered the estimates.
Files That Will Not Line Up Two or more files describe the same people or households, the identifiers differ in format or repeat, and no one has yet checked how many cases the joined file should hold.
Documentation Still Unread A public-use file has arrived with a codebook, a user guide and several weight variables, and it is unclear which weight, stratum and cluster variables the planned analysis needs.
Estimates for One Subgroup The question concerns one part of a surveyed population, such as older adults, and the subgroup must be analyzed without distorting the variance estimates the design requires.
A Reviewer Asks About Weighting A supervisor or journal reviewer wants to know whether weights were used, why the estimates differ from published figures and how refused or skipped answers were treated.
WHAT THE WORK COVERS
From the Codebook to the Reported Estimate
Existing data are analyzed well only after they are understood, which is why the work starts with the codebook and ends with a table whose source a reader can verify.
Reading the Documentation
Sampling design, weight variables, skip patterns and coding conventions are read from the codebook first, because they decide which procedures and which variables the analysis may use.
Merging Files
SPSS merged data files are checked on a key variable for duplicates and unmatched cases, with case counts compared before and after, whether files add variables or add cases.
Missing and Special Codes
Codes such as 99 or -1 are declared missing, and incomplete records are judged as in Missing Data Analysis, so that refusals never enter a mean as real values.
Design-Based Estimation
SPSS complex samples analysis uses the plan variables for weights, strata and clusters, so standard errors and intervals reflect the design.
Models on Existing Data
The agreed tables and models are produced through Quantitative Data Analysis, fitted to the survey design and not to a simple random sample.
Reporting the Limits
Reporting states the source, the year, the unit of analysis and what the variables only approximate, in line with the conventions of Public Health research.
COMMON DATASET ISSUES
Errors That Enter Before the First Table
Secondary analyses usually go wrong in preparation, where a quiet merge error or an unused weight shows only when a figure fails to match the published source.
Weights Applied as Counts A weight is switched on as a frequency, so the sample size is inflated when the weights sum to a population total, and every p-value is then too optimistic for a survey.
One-to-Many Merge Repeats A household file joined to a person file without checking the key repeats household values for every member, so household facts are counted several times in later averages.
Subgroups Filtered Out Cases outside the subgroup are deleted before a survey analysis, which can distort the variance estimates, so a subpopulation is specified in the plan and the other cases stay in.
Special Codes Read as Values Refused or not applicable answers coded 97 to 99 stay in the data as real numbers, pulling means upward and creating outliers that vanish once the codes are declared missing.
Categories That Changed Between Years Variable names or categories differ between survey years, and the files are stacked as if the coding matched, so one category quietly stands for two different things.
A Measure That Only Approximates The closest available variable stands in for the construct in the research question, and the gap between the two is never stated as a limitation in the report.
DELIVERABLES
What Comes Back With the Analysis
A reader of a secondary analysis wants provenance, a record of every join and the reason behind each weight, and the six documents supply them.
Dataset Provenance Note The source, year, population, sampling design and variable definitions in one note, so a reader knows what the data can and cannot represent in your study.
Merge Audit Table Case counts for every input file and for the joined file, with the key checked for duplicates and each unmatched case listed with the decision taken on it.
Design Specification The survey design settings with the weight, stratum and cluster variables named, and a note on why each was chosen from the documentation provided with the data.
Commented SPSS Syntax The syntax from the raw files to the final tables, with each recode and merge step commented, so the analysis can be rerun when a newer release of the data appears.
Weighted Tables With Intervals Estimates with design-based confidence intervals and the unweighted case counts behind them, laid out in the format your paper or report requires.
Limits of the Data Note A note on the limits of reusing these data, covering coverage, measurement and what the variables approximate, so that the scope of the findings is stated accurately.
ILLUSTRATIVE EXAMPLE
A Survey Subgroup Analysis With Linked Files
An illustration of how file preparation and survey design change what a secondary analysis is able to report.
ILLUSTRATIVE EXAMPLE · PUBLIC-USE HEALTH SURVEY RESEARCH · UNITED STATES
A postdoctoral researcher in gerontology wanted to estimate how often older adults had skipped medication because of cost, using a public-use national health survey released as separate person and household files.
The files were joined on the household key after checking for duplicates, and the case count of the joined file was compared with the person file. The complication was the design: the survey used stratified cluster sampling with unequal selection probabilities, so a weighted frequency table that ignored the strata and clusters would most likely have produced intervals that were too narrow. The analysis used the supplied weight, stratum and cluster variables, specified older adults as a subpopulation instead of deleting everyone else, and treated refused answers as missing. The comparison with published tables was documented, and the findings were described as estimates for the surveyed population.
Outcome: The researcher received the weighted estimates with their 95% confidence intervals, the merge audit table, the design specification, the commented SPSS syntax and a note on the limits of the data.
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
Choosing the Program for Existing Data
Existing data often arrive in another program's format, so work moves between tools only where it adds something and the final tables come from one traceable run.
HOW WE ASSESS YOUR PROJECT
We Review Your Actual Research
Your assessment is based on a review of your actual research materials and analytical requirements, not a generic project category.
You Submit Your Project Details
Share whatever is available at the current stage, such as the research aim, protocol or proposal, dataset, variable list, manuscript or statistical requirements.
We Review Your Research and Define the Scope
Your statistically trained Research Coordinator reviews the materials to understand the study design, research questions, data structure, analytical requirements and expected deliverables. If clarification is needed, additional documents or information can be requested before the assessment is finalized.
Your Analysis Is Completed and Delivered
Once the scope and fixed project fee are agreed, the in-house statistician completes the work and the Research Coordinator remains the communication point through delivery, with the analysis, deliverables and timeline as agreed.
RESEARCH RESPONSIBILITY
What Stays With You in a Secondary Analysis
OnlineSPSS prepares and analyzes the files you choose, while the dataset, the framing of the question and the conclusions drawn from the estimates remain yours.
Agreed Statistical Support OnlineSPSS provides the agreed methodological, analytical and statistical reporting support, defined in a scope that is confirmed before any work begins.
No Intellectual Contribution OnlineSPSS does not claim any intellectual contribution to your paper or report. Authorship remains the researcher's.
Your Research Decisions The research question, the choice of dataset, the substantive interpretation of the estimates and the conclusions remain yours, with overall responsibility.
Confidential Handling Research data and project details are handled confidentially, project information is not published, and an NDA is available where required.
FREQUENTLY ASKED QUESTIONS
Questions About Analyzing Existing Datasets in SPSS
Existing data raise questions that data you collected yourself do not: how to join files, who must approve the use and how categories can be compared across years.
What does analyzing an existing dataset in SPSS include?
SPSS secondary data analysis at OnlineSPSS covers the preparation of an existing dataset and its analysis in SPSS: verified file merges, declared missing codes, weights and design variables, and estimates reported with confidence intervals. Which files, weights and estimates fall inside the work is decided at the assessment.
How can I check that a file merge in SPSS worked?
Compare case counts before and after, confirm that the key variable is unique in the file where it should be, and look at the cases that did not match. The Identify Duplicate Cases command helps with the first check, and the Merge Files dialogs can add a variable recording which file each case came from. A merge is only as reliable as the check that follows it.
Do I need ethics approval to analyze data someone else collected?
Requirements differ by institution, by country and by whether the data are public-use and de-identified, so the ethics office and the data provider’s terms of use decide, not the analyst. Some institutions exempt public-use files, and others still ask for a determination. Check before analysis begins, and keep the documentation with the project.
Can categories be compared across different survey years?
Only after the coding has been aligned. Question wording, response options and category definitions can change between releases, so a harmonization table is built first, mapping each year’s codes to a common scheme and noting variables that cannot be matched. Where a definition changed, a trend across the break is reported with that caveat.
How do you protect my research data and confidentiality?
Your data and project details are handled confidentially, and project information is not published. Researcher names and institutions are not used publicly without permission. 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, such as the data files, the codebook and any weight documentation, 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
Which Dataset Are
You Working With?
The survey or registry, the question it should answer and the files already in hand are what the merging and weighting plan is built around.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014