Collection of data is an important part of identifying the needs of your clinic community, but it can also be a sensitive process! Who is asked, what questions, why, when, where, and how can all impact participants’ responses. This is especially true in the workplace where there are unequal distributions of power.

SNPR toolkit-10.png
SNPR toolkit-19.png

A Note About Power:
In community organizing, the word “power” is often discussed. Its meaning is most clear from the Spanish word “poder.” In Spanish, this single word serves both as the noun power and the verb to be able to do something. It is important when doing any work with communities to recognize how power is distributed. Thinking of power as the ability to do something allows us to examine how power is distributed across members of a clinic community. In many clinics, as in other places of work, power is concentrated among managers, leaders, or providers.



Make it clear that the decision to take a survey or express criticisms will not affect participants’ ability to get care or retain their employment (if this can be guaranteed).

SNPR toolkit-16.png

It is important that participants feel free to decline participation.

  • Would you feel like you could decline if your doctor asked you to take a survey? What about your boss?

  • How might you answer a survey differently about your work environment if your boss is collecting and processing the data?

If you want to gain the most accurate insights from the survey results, it is vital to understand the perspectives of many different groups stakeholders. For example:

  • The perspective of those with private insurance, Medicaid, and those without insurance likely differ.

  • Those who speak English may have a different experience in your hospital from those who need an interpreter to communicate.

  • Patients of color may have different experiences than White patients.

  • Female patients likely use different services at your hospital than male patients, and their attitudes about providers and staff may differ.

Getting a diverse sample of participants provides unique challenges. It’s not just who’s taking the survey, but also who’s administering it.

  • Considering power dynamics is very important when designing processes to collect data and implement and monitor improvement work. What are the power dynamics between who is administering the survey and who is responding to it?

  • For examples, will staff fear that their responses might be identifiable to clinic leadership?

  • Will patients feel comfortable answering questions if surveys are administered by their providers?

SNPR toolkit-13.png


A Note About Language:

  • Think carefully about how including/excluding certain language groups might impact your results.

  • It can be costly to conduct surveys in more than one language. Depending on the clientele at your institution, you will have to balance the tradeoffs.

  • If there is a language beside English that is particularly prevalent at your clinic, consider have an interpreter available to read the questions aloud to patients



SNPR toolkit-15.png

Being clear about what questions are being asked and what will be done with the data is important. There are trade-offs between using a full survey created by an outside group (such as the one in this toolkit), choosing some questions from different established surveys, or completely making up your own.

In making this decision, you may want to consider:

  • What do you hope to do with the data you collect?

  • How important is comparison to other clinics or health systems?

  • Do you need or want to publish survey data?

    • NOTE: Plans to share data with external partners often require review from a Human Subjects Protection or Institutional Review Board.

  • Do you want to ask the same questions of patients, staff, providers, and other community members?

    • On the one hand, if you ask the same questions of all clinic community members, it can be easier to compare across groups. On the other hand, you may want to know certain topics that are specific to certain community members. Oftentimes, the answer is a mix of both!

Whether you use the survey instrument provided in this toolkit, or if you make up your own, here are some survey design guidelines you can follow: 

SNPR toolkit-08.png

If you make your own survey items:

  • When possible, use the same scale for Likert-type items (e.g., 1-5, “strongly agree” to “strongly disagree”)

    Considering keeping questions in the same direction. For example, 5 is always more positive (e.g., strongly agree to positive items) and 1 is always more negative. If you decide to reverse code some items (i.e., flip the scale), make sure to reverse the score when you are analyzing your data.

  • Reassure participants they can skip questions they do not feel comfortable answering

  • If you have the capacity to process open-ended data, consider asking participants to share their questions or comments with data collection leaders, in an open-ended question at the end of the survey, or even written into the margins of paper surveys. Sometimes this input can be invaluable!

  • Whether to survey on paper or using the internet or both is a complex decision. You may consider:

    • How does your target audience use technology?

    • What mechanism will be most feasible?

    • What equipment do you have to assist in gathering surveys?

    • Who will be included or left out depending on what mechanism you choose?

If you would like to know more about survey design, here are some helpful resources:

From the US Department of Health and Human Services’ Agency for Healthcare Research and Quality: CAHPS Survey Administration

The Pew Research Center:Questionnaire Design


Being clear about why a survey is being done is important. This may help people feel comfortable participating. Understanding the “why” of surveying also may encourage people to actually take the time participate

SNPR toolkit-17.png

A Note On Incentives: Members of the clinic community are busy and their time is valuable. You may want to consider a small token of appreciation to those who participate. This can range from coffee and snacks in the surveying area, to small gift cards, to entry in a raffle for survey participants. Though not necessary, these small tokens are a great way to let members of the clinic community know that you value their time

SNPR toolkit-08.png

HOW to analyze survey results

Now that you’ve conducted your survey, how should you analyze the data to help you identify areas in need of improvement? The way you analyze and present your data can greatly impact how you interpret your results. There are thousands of ways you can “slice and dice” your data, and each decision you make will affect your interpretations. Below are two decisions that frequently need to be made when analyzing survey data.



Make sure to examine data as a whole, but also think of groups that might present different patterns.

For example, the following survey item: “Providers in the clinic are open to ideas from staff about how to improve care processes.”

Averaging answers from both staff and providers, could demonstrate a majority answer neither agree nor disagree. However, when examined separately, staff and provider responses differ in important ways.

survey bar 2.png

Similarly, there are likely groups of patients whose data you want to examine separately.

Groups to consider: male and female patients, members of different racial/ethnic groups, patients with different types of insurance

For example, the following survey item: My provider(s) include me in making decisions about treatment. Examining patients as a whole may reveal moderate scores, but may mask differences if you compare English speaking patients to those who use an interpreter during their clinic visits.

survey bar 1.png


Data visualizations are a powerful tool for helping you interpret your data and for telling a compelling story when reporting back your findings to members of the clinic community.

Different types of visualizations are ideal for different types of data. Here are some guidelines for when to use visualizations that are common in survey research:

  • Pie charts are great for comparing proportions or percentages within a group

  • Bar charts are helpful for comparing groups to each other

  • Line graphs are helpful for looking at relationships between variables or looking at change over time.

SNPR toolkit-18.png

A Note On Anonymity: When you share findings with anyone, and especially within your own clinic community, it is essential to maintain anonymity. It is important to only share results that are general enough that clinic members can’t figure out who they come from. For example, if your clinic only has three social workers, reporting findings specific to social workers would make it easy to others to figure out who those results came from.

  • Current guidance from the National Institute of Health is that you should avoid presenting data from groups small than 10.

  • Group survey respondents in rationale ways to protect small samples. For example, instead of presenting physicians and nurse practitioners separately, you can group them as providers

If you are interested in learning more about visualizing health care data, please visit Visualizing Health.

pie graph.png

Pie ChartS

For example, imagine that you are interested in how easily patients at the clinic are able to access mental health services. In the survey instrument for providers and staff included in this toolkit, we asked employees how strongly they agreed with the following statement: If a patient needs treatment for a mental health concern, the team can refer them to services easily.

Imagine that 65% of employees agreed with that statement, 30% disagreed, and 5% weren’t sure. Because were are taking about percentage of a group of people, a pie graph would be an appropriate way to visualize these data.


But what if you want to compare employees’ perceptions to those of patients (survey item If I need treatment for a mental health or substance use concern, I can get the care I need)? Rather than making two pie charts, a better way to compare these two groups of members of the clinic community would be a bar graph.

double bar.png

line graph.png


Last, imagine that your clinic begins consistently surveying community members for several years. You may want to be able to compare across years to see if the clinic is improving. A line chart would be a useful tool for showing these changes over time.


SNPR toolkit-17.png
SNPR toolkit-18.png

Surveying by mail is a common occurrence. You may consider:

  • How will mailed surveys expand your possible participants? Who will be included?

  • Do you have a budget to cover the costs of printing, preparing, and mailing out surveys?

  • What response rate is realistic for your population?

  • Can the survey be included in other patient or employee mailed materials?

Surveying in clinic is also common. You may consider:

Consider the tradeoffs of when to survey and who hands out surveys.

  • Providers: While you may get higher response rates, patients may feel pressure to respond. Providers may be unwilling to add this to their workflow.

  • Front desk staff: How will surveys be distinguished from other clinic paperwork? Does the waiting room offer enough privacy? Do you have clipboards/pens?

  • Nurses: Are they willing/able to add surveys to their workload? How will surveys be collected?

  • Administrative leaders: Staff meetings can be convenient times to survey providers/staff. How can privacy be maintained and coercion avoided for those who may not want to participate?

When you administer your survey can also affect who decides to participate.

  • Consider what season, day of the week, and time of day. Different patient populations may be reached depending on these options.

  • Are there other surveys the clinic, health department, researchers, etc. are conducting? Try to avoid multiple surveys at the same time in the same clinic if at all possible.

  • Is there a staff meeting when you can get many employees to take the survey at the same time? As you are surely aware, clinic staff are busy, so administering surveys during meeting times can help you avoid burdening the staff.