What is secondary analysis? A comprehensive overview

There’s no sense in recreating the wheel in research or development—or business generally, for that matter. If someone has already innovated in a new approach or elevated a deliverable, you don’t always need to go back to square one to reinvent it. Instead, you “stand on the shoulders of others” and look to innovate from there with your research. This concept harnesses the value of secondary analysis.

Within your projects and initiatives, you can leverage secondary analysis, like case studies, census data, or past clinical trials, to accelerate growth and innovation. But there’s more to understand when peeling back this onion.

Building on the idea of leveraging existing work, we’ll provide a comprehensive overview of what secondary analysis really is, including its applications in various fields and its advantages and disadvantages. You’ll also discover the most effective ways to incorporate it into your projects.

Secondary analysis is any form of research that relies on or uses previously conducted research for the purposes of a new study. If existing data is cited or previously conducted studies help to achieve a new outcome, it’s secondary analysis.

It happens quite often, especially when researchers use quantitative or qualitative data that has been gathered previously and analyze it in a new way. The secondary data in these instances is often published or made available publicly with permission to cite and use it

Secondary data analysis is important for innovation. It can serve as a historical reference to studies of yesterday.

This kind of previously summarized data, usually written by other parties, can be great for reaffirming a similar result or finding you may have. It’s also where you can find commentary or analysis of the steps taken before you’re on the scene.

Secondary analysis is important because it can accelerate your research by providing reliable springboards from previous research. This allows you to pick up where someone else left off in the field.

Imagine you’re studying a local population or community demographic. In this case, turning to the latest census data will be more useful than conducting your own headcount. Even simpler: if you’re making a cake from scratch, it’s a good thing you don’t have to churn your own butter to complete the recipe. You get the idea. But there’s more to secondary analysis than just census and cake.

Imagine all the libraries of studies and the internet’s access to peer reviews, case studies, statistics, and data. Think about how much data is out there. A lot! However, it won’t all be useful in your studies or projects. This is why it’s best to understand the categories and types of secondary research. From there, you can narrow your focus to the data and research that best applies to your work.

Statistical analysis

The SAS Institute says statistical analysis is the science of collecting, analyzing, and presenting giant batches of data in an effort to identify any underlying patterns and trends. This is where you might find census data, for example.

A collection of thousands of data points is combined with the intention of discovering new insights. And those statistics are applied to every corner of business, government, health, and science.

This type of secondary analysis is great for drawing new insights, testing new hypotheses, and validating new findings. We rely on statistics to help us identify areas of improvement and avoid mistakes every day.

Literature reviews

Think of these as published pieces of information within a particular segment or subject area.

The primary purpose of any literature review is to provide a multifaceted and comprehensive overview of current knowledge, identify gaps, and establish a theoretical framework for further research.

Sometimes, literature reviews are part of a topic within a specific time period. Researchers can present them as an in-depth analysis or just a simple source summary.

Case studies

Case studies are more in-depth analyses of a person, group, or specific event.

In secondary research, you’ll look for existing case study reports, published papers, and documented instances to gather and analyze data. These studies can provide comprehensive insights into specific phenomena, processes, or practices.

Nearly every aspect of the topic is explored, highlighting challenges, solutions, how those solutions are applied, and final outcomes. They can prove or disprove a theory but typically serve as a demonstrative piece of evidence or analysis.

Content analysis

Content analysis is, as it sounds, the study of certain phrases, words, or themes within a body of qualitative data.

This type of secondary analysis provides context so you can analyze particular relationships between words, meanings, or concepts. It’s a type of secondary research that involves examining and interpreting pre-existing material to uncover patterns, trends, and insights.

Starting your research from an advanced position of knowledge gives you an advantage. Secondary data analysis presents many benefits, regardless of the topic or information you’re exploring.

Cost-effectiveness

Secondary analytics are cost-effective. Since studies and data already exist, you don’t have to repeat certain tests or steps, alleviating costs without compromising your findings. You’ll gain access to high-level data that might be too cost-prohibitive to perform independently.

It can also be cost-saving in exploratory research. It’s helpful to gain valuable preliminary or exploratory research first to refine your hypothesis before committing to primary data collection.

Time-saving

Someone else may have already invested in the study, producing findings and data you can readily use within your research and saving you time.

Being able to access large datasets, which are typically extensive and robust, provides researchers with a wealth of information that would otherwise be too time-consuming or difficult to gather independently.

Ability to answer additional research questions

Within the scope of your project or research, there will be questions you can’t answer first-hand. Using secondary analysis allows you to answer those additional research questions using pre-existing findings or results.

You can also analyze old longitudinal data and still find new trends, theories, or applications. Enabling longitudinal studies can also help you track changes and trends over time.

You can’t run your business and projects entirely by piggybacking on secondary data alone. There are some disadvantages to relying solely on analysis that has already been published.

Data quality concerns

You weren’t “there.” You weren’t part of the study or model behind the secondary analysis, so there’s a risk that mistakes were present or findings weren’t entirely accurate. You’ll need to verify that the sources aren’t citing outdated information or presenting original data with bias.

Relevance is another potential issue. Some data might be outdated or not reflective of current conditions, especially in rapidly changing fields or industries. These concerns about data quality could be problematic for your new research project or objective.

Data accessibility

Access to certain secondary data sources may be restricted, require payment, or come with usage limitations. Statista and Deloitte, for example, provide high-quality reports, but some are only accessible after payment.

Also, some academic articles may require a subscription or an account affiliated with a higher education institution.

Need to de-identify information

When you incorporate secondary analysis into your project, you’ll need to remove identifiers from the original source. For example, de-identified patient data won’t have specifics about the patient’s personal information.

Follow this simple roadmap for carrying out your own secondary analysis study.

1. Identify and define the research topic

The first step is recognizing the goal of the project or question you’re attempting to answer. Define your research topic in a way that provides clarity for the datasets you’ll need to collect.

2. Find research and existing data sources

Consider which data sources exist that might present the findings you need to help answer your question.

Start digging through reputable sources within relevant timelines to explore what data already exists. These might include academic databases, government records, organizational reports, and online repositories.

3. Begin searching and collecting the existing data

With an idea of what types of secondary analysis will offer the best data for your project, start searching for relevant studies and collect those that might help. Collect several sets of metrics and analyses to inform your analysis.

4. Combine the data and compare the results

When you feel you’ve collected the right studies, you can begin combining your findings and comparing results.

Before you get started, you might have to “clean” the data, handle missing values, or merge datasets from different sources.

Look for trends and common themes or results. Verify if outlier statistics are anomalies or valuable to your study.

5. Analyze your data and explore further

Analyze your collected data through the lens of market research. Look to determine what the data shows, ensure it makes sense, and connect the dots to reach your original goal.

Use appropriate statistical or qualitative analysis methods to examine your data. For you, that might involve descriptive statistics, inferential statistics, or thematic analysis, depending on the nature of the data and your research questions.

With literally thousands of websites and research at your fingertips, you can find relevant secondary research sources practically anywhere. You might also locate research from both internal and external sources.

Internal data

As a company, you can look for internal secondary data to help you reach the answers you’re looking for. Here are some examples:

• Historical sales reporting

• Website analytics from previous years

• Past employee training and testing results

• News articles

• Reports

• Internal conversations

• Customer databases

• Internal communications records

• Financial reports

External data

Look outside your company for similar reports other companies may have already executed. Trust industry-specific sources on the web as well as municipal or non-profit studies that may also lend credibility to your work.

Academic journals, public databases, industry reports, trade publications, and government agencies can all be valuable resources for secondary data. Here are some examples:

• Elsevier

• Springer Nature

• Census.gov/data

• Statista

• Demographics Now

• Nielsen

• Deloitte

Most of the best innovations and research initiatives are served by both primary and secondary research analysis.

Don’t look at one or the other. Instead, harness both for the most impactful research. Springboard from the work of others and use secondary research to bridge gaps in your efforts. Then, conduct your primary research to deep dive into those arenas, combining both for the most thorough investigation.

Secondary analysis is more common than you might realize. You may have used it before without realizing it.

Here are some use case examples of secondary analysis used across business applications, research and development, learning, healthcare, and more.

• A grad student expands on an advisor’s research to contribute to a thesis.

• A data analyst uses their own data to run additional reports.

• A researcher uses new software to further explore historical reporting.

• An entrepreneur studies demographic information to create more effective marketing personas.

• A school principal uses nationwide studies to inform curriculum development.

• A digital marketing specialist uses site metrics to outline areas of improvement for user experiences.

As a researcher, be sure to evaluate the data you source to make sure it’s accurate, timely, and reputable. Before applying any secondary analysis, remove bias by answering the following questions about the source:

• What was the secondary study’s original purpose?

• Who collected the original data? (credentials)

• What data was collected and when?

• What were the methods used and dataset limitations at the time?