Survey Blog
Official LimeSurvey Blog

We want our users and our community to feel safe and protected when working on and with our online survey software, which is why we will expound the essentials you need to know in terms of laws, measures, and security associated with data gathered by LimeSurvey. There are different legislative levels involved in the bigger picture of data security that LimeSurvey is subject to. Federal Data Protection Act (Bundesdatenschutzgesetz, BDSG) Starting at the organisational level, on which LimeSurvey receives data input from any potential user, data security and privacy have to be ensured by orders of the superior federal state level, i.e., the respective legislation of any of the 16 federal states in Germany. However, these multiple federal state data protection laws have to coexist with regulations on the country level, the Federal Data Protection Act, which was first introduced in 1978. You can read the details of the Federal Data Protection Act here.  Any German organisation such as LimeSurvey has to adhere to any data security regulations dictated by the Federal Data Protection Act and the federal state laws. The Federal Data Protection Act is further connected to European regulations. European Data Protection Directive 95/46/EC The Data Protection Directive 95/46/EC was adopted by the European Union in 1995 in order to regulate data processing and laws within its member states. It serves to specify and dictate minimal data security requirements that each member state is to incorporate into internal law. In fact, by 1998 all member states had created their own data protection laws according to the European Data Protection Directive, including Germany. You can find out more about the European Data Protection Directive 95/46/EC here. General Data Protection Regulation (GDPR) The General Data Protection Regulation is a new data protection regulation drafted by the European Union and will replace the previously existing European Data Protection Directive, as well as the German Federal Data Protection Act, and will, therefore, directly impact German laws on data security, including measures undertaken by LimeSurvey to keep users' data safe. The GDPR was implemented on 25 May, 2018 and aims at strengthening data protection throughout the EU by unifying regulations incorporated into national law of all member states. Thus, data security policies will assimilate in the future across countries in the European Union. The goal is to increase data security by standardising regulations within an established and dynamically cooperating inner-European market. The reason for this radical shift of EU regulations is founded in increasing data abuse possibilities due to expanding cloud computing and big data efforts prevalent among organisations in all areas nowadays. You can find out more about the European General Data Protection Regulation here. In Safe Hands with LimeSurvey Any data collected and processed by LimeSurvey is done strictly within the legal requirements. We are an Open Source organisation that thrives on our great community without forcing or restricting anyone's use of our online survey software. The community can only keep growing safely if data protection is ensured at all times, which is the most valuable user right in the eyes of the LimeSurvey company. LimeSurvey neither discloses any user data publicly, nor transfers any user data to any third party without an explicit consent. It has always been this way and will always remain this way.  

This segmentation which precedes any personalised addressing of customers requires defining the criteria to segment by, such as demographic data in order to better determine who your customers or prospects truly are. If you are running a survey with one of the aforementioned target audiences, it is vital to ask the right demographic questions in order to maximise your understanding of this data. In the following sections, we will take a look at the best demographic questions to ask and how to ask them in surveys. What Demographic Data to Collect in Surveys Naturally, not every survey should contain all conceivable demographic questions. The process of selecting and writing demographic questions is always strongly case-dependent and should be chosen based on research strategy and goals. Therefore, make sure to explicitly define the purpose of your data acquisition beforehand and you will be on the right track. Nevertheless, a certain basic set of demographic data will always be helpful in segmenting survey research data in order to get to know your main prospects or customers. What is your age? The art of asking about a respondent’s age is balancing the level of data detail as well as avoiding survey fatigue, i.e., grouping individual ages into ranges leads to fewer answer options, improves readability and possibly lowers response fatigue. How finely you choose to distinguish between age groups should be tested for comparison of response rates. It is important to work with a single choice question format such as a radio list question type, as these answer options need to be mutually exclusive. Also, make sure that your age ranges don’t overlap, otherwise your data evaluation will be difficult. Here’s an example of how you could structure the question: o Under 18 years old o 18 – 24 years old o 25 – 34 years old o 35 – 44 years old o 45 – 54 years old o 55 – 64 years old o 65 – 74 years old o 75+ years old What is your gender? Asking about a person's gender has become an increasingly sensitive topic, as apart from male and female gender attribution, multiple alternative gender descriptions have arisen recently that have been recognised as official gender statements. LimeSurvey offers a prebuilt button style or radio list gender question type. It, however, only includes male and female as answer options so far. If additional gender options are to be included, a simple radio list question type will do. Also, a simple open-ended question type can be used to give respondents the ultimate freedom in choosing their gender identification. Here’s a preview of possible questions regarding the sex/gender of your test person:   What is your sex? o Female o Male   What is your gender identification?   o Male   o Female   o Other:   What is your marital status? Learning about respondents’ marital status is often one of the less important demographic questions, as this segmentation does not usaually deliver significant value to the majority of most businesses. Nevertheless, there are plenty of benefits of gaining information from this data, especially since it is a fairly straightforward question. It makes sense to enable multiple choice question types in this case, as a combination of answer options is possible, e.g., someone who is single could also be recently separated from a meaningful relationship.Here is an example of how to ask for marital status:  Single  Married or domestic partnership  Widowed  Divorced  Separated  Other How would you specify your ethnicity? Questions about ethnicity, race or origin are highly sensitive in nature due to conflicts and crises that have shaped society through past events and prejudices. Nevertheless, the data derived from this question is popular to segment survey responses in order to derive trends and patterns in regard to cultural impact. As this is not a mere aspect of documentation, but also based on belief systems and emotional identification, you should use a multiple choice question type to help ensure that respondents have multiple options to choose from depending on how they identify. This example could be used as a template for the question:  African American  White  Hispanic or Latino  Native American  Asian  Other What is the highest educational level you have acheived? Receiving data on respondent’s qualifications is a great way to analyse educational impact on answers given in your survey.Make sure to provide a complete and distinct list of qualification levels here to cover any possible situation. Usually, answer options involving degrees can be thought of as mutually exclusive, as they build on each other in a hierarchical structure when asking about specifying the highest qualification, so you should use a radio list question type here. o Less than high school or secondary school degree o High school or secondary school degree o Bachelor's degree o Master's degree or diploma o Doctorate o Other What is your annual gross household income? Similar to the age question presented above, income questions can face the problem of survey fatigue depending on how detailed, and thus how many answer options are presented at once. Categorising income in ranges will again help reduce the available options and make it more transparent. Make sure to use a currency that is reccognized globally to improve comprehensibility and comparability of answer options, e.g., U.S. dollars or EURO. As money is always a sensitive topic, it is recommended that if you are not in desperate need of this data to make the income question a voluntary one, as many people do not like giving information on their personal finances. o Less than €30.000 o €30.000 to €49.999 o €50.000 to €69.999 o €70.000 to €89.999 o €90.000 to €99.999 o €100.000 or more What is your current employment status? If you are questioning a wide variety of people, getting information on their employment status can be very interesting and valuable. The difficulty with this question, however, is to provide all possible occupational statuses. Listing a few common ones and offering an Other option, however, should solve the problem. Make sure to use a multiple choice question here since people can have different occupational statuses at the same time, e.g., a student can also be employed part time or self-employed.  Employed full time (40 or more hours per week)  Employed part time (up to 39 hours per week)  Unemployed  Self-employed  Unable to work  Retired  Student  Other Which industry do you work in? This is an optional question and obviously will only be relevant if the aforementioned question was answered with a full time, part time or self-employment response. This is also a tricky one, as you may use a single or multiple choice question type and have a very long list of industries which may lead to survey fatigue, or you could offer an open-text question which will significantly increase the amount of time it takes to complete the survey. A great approach here would be to create a question that works with search recognition, a hybrid of the aforementioned options in a manner that allows the respondent to start typing in an open-text field, and based on this, offers predefined answer options that will standardise the results to be evaluated later on. This, however, will require coding expertise. Leverage the Value That Demographic Data Can Deliver The key to obtaining the maximum benefit from demographic data is contingent upon two aspects: Firstly, you need to align your questions with your research goals and strategy by choosing the necessary demographic questions and phrasing them in a way that they are appropriate and meaningful. Secondly, you need to make sure not to overload your respondents with too many demographic questions at the same time.  So you have to find a good balance between getting enough data for your research requirements and not having too many questions, which might cause your respondents to abandon your survey before completion.

Essentially, this means you can either show your entire survey on one page, which is called scrolling survey design in the research world.  You can show your survey on multiple pages, i.e., one question per page, which is called paging survey design.  You can alternatively use a hybrid format offered exclusively by LimeSurvey that allows you to show one question group per page consisting of several questions, which we will call categorising survey design. What seems like an easy decision at first will require some further consideration when looking closer at this design topic.  In light of the availability of different technical devices to take surveys on nowadays, such as desktop computers, tablets or smartphones, formats may appear more advantageous on one device, yet more disadvantageous on another device. This adds extra complexity to survey design. Researchers have been testing and evaluating this design question for years now by examining key metrics such as bounce rates, completion times and rates, as well as item nonresponses. Insights indicate that surveys taken on desktop computers do not differ significantly for scrolling and paging designs with regard to completion and bounce rates. Merely, a tendency to skip certain questions was discovered for scrolling survey designs which could have had any number of reasons such as a mental overload, boredom or other problems connected to survey fatigue. Tips For Different Survey Design Formats One thing has to be stated clearly: there is no general best strategy that will apply to any type of survey. Ultimately, it comes down to the survey characteristics with regard to survey length in terms of time and number of questions, as well as technical equipment and connectivity. Nevertheless, scenarios can be created within which a certain design format may be more suitable than others due to the underlying survey conditions. This leads to the following tips for the usage of different survey design formats: In general, if your survey is short in terms of the time it requires to be completed, as well as a low number of questions, the recommendation will always be to use a scrolling design to keep the survey short, regardless of the device used. A paging or categorising design will spread questions over several pages and require respondents to click on the next button as well as to adjust to the subsequent page(s), which prolongs a survey. This will almost always exhaust or lower respondents’ interest potentially resulting in lower completion or higher bounce rates. A paging design usually only makes sense if you have a long survey that is to be taken on either a desktop computer or a laptop, as these devices possess larger screens, as well as allow for faster cursor movements and clicks. This way, surveys can be completed in a more convenient manner when shifting from page to page. Obviously, the aforementioned paging design can also be replaced by a categorising design which is the more elegant way as it groups question subtopics to facilitate respondents recognising logical connections between questions. Because tablet and smartphone devices have smaller touchscreens, a categorising design should be used with long surveys, as it would minimise the negative effects of both pure paging and scrolling designs. In the categorising design, fewer clicks and new page loads are required than with the paging design, as well as less scrolling and information overload are present than with the scrolling design. The advantage of paging and categorising design is perfectly clear: every time a respondent completes one or a few questions on a page and clicks the next button, this data is submitted and saved, even if the respondent drops out on the subsequent page. If a respondent completes the same amount of questions in a scrolling design and drops out without submitting, all that data is lost. In this scenario, paging and categorising designs are more valuable as partial data may still deliver important insights.     ⇨ All in all, we do, however, recommend that you try to avoid creating long surveys with many questions, as these will oftentimes lower completion rates and increase bounce rates due to survey fatigue. Try and split longer surveys by subtopics into multiple shorter surveys. This way, you will make sure to increase survey relevance to your target audience and earn more valuable data to analyse. In the end, it all comes down to how well you know your target audience. The more you find out about devices used and survey perseverance, the better you can suit your survey design to achieve the best results.  

Take a look at the following assortment of the most popular and widely used chart types to get a good idea for evaluating and reporting on your own survey data. When to Use a Pie Chart single choice questions mask questions (e.g. Gender, Yes/No) Pie charts are best used when you are presenting a share of something, ideally items explaining a single variable to benefit readibility and assessment. Logically, all shares in a pie chart should thereby add up to 100%. For example, when you have binary survey responses, e.g., yes/no question types, or questions with only a few answer options/categories. Overloading a pie chart with too many categories, i.e., 6 or more, makes it hard to identify labels as well as shares and should be avoided. When it comes to data types, nominal data such as demographics, i.e., data without an order or ranking, as well as ordinal data such as likert scale ratings, i.e., data with an order or ranking, can be visualised using pie charts. Nominal data can be arranged clockwise from largest to smallest share and ordinal data can be arranged clockwise according to ranking or rating scale. In order to increase distinction between shares in a pie chart, strong colour contrasts or colour shades can help to achieve this. Pie charts are a nice and visually appealing solution if you are trying to show general trends rather than specific accurate data. General Tip: Avoid including "Other" as an option choice, as an answer as different responses will not be displayed in a pie chart, but will be shown as "Other". When to Use a Donut Chart single choice questions mast questions (e.g. Gender, Yes/No)  Donut charts work similarly to pie charts. Donut charts have a circular cutout in the middle whereas the shares within pie charts completely fill out the middle.  Despite these seemingly small differences, the way these chart types are viewed and interpreted from a psychological standpoint vary substantially. Shares of a pie chart all run to the middle and focus a viewer's attention to the centre where all pieces are processed as by the size of their areas. Shares of a donut chart however focus the viewer's attention to the length of each piece, thereby demonstrating a more linear approach of processing results. The latter way of interpreting data is naturally easier for people. Apart from that, donut charts can be used in the same way as pie charts, i.e., assessing shares of a whole. When to Use a Bar Chart single choice questions multiple choice questions arrays mask questions (e.g. Gender, Yes/No) There are several different ways to visualise data with bar charts. Most commonly used are the vertical and the horizontal bar charts. Both vertical and horizontal bar charts are composed of columns which are either aligned vertically or horizontally to the x-axis respectively. Furthermore, they are suitable to visualise nominal and ordinal data.  The vertical bar chart is helpful when displaying percentages or means of only a few categories. Otherwise, too many categories would increase the width of the entire chart across a screen and make it hard to read or even display. Rule of thumb here is to reduce categories included in a vertical bar chart to only about a handful. Compared to pie charts, bar charts facilitate direct comparison between categories, especially when ordinal data is arranged from largest to smallest column or vice versa. Furthermore, bar charts can easily display negative values which is impossible with pie charts and makes the former superior and more flexible in most cases. The horizontal bar chart comes in handy where the vertical bar chart falls short. If you have more than a handful of categories, select the horizontal bar chart as it extends vertically and improves readibility. Stacked and grouped bar charts are great solutions when displaying subcategories, i.e., break down categories for a more detailed view of your data. The only difference between stacked and grouped bar charts is the way subcategories are displayed, namely subcategories in the same bar versus in different bars. General Tip: Make sure to display your bar labels horizontally, otherwise viewers have to tilt their heads to read them. Also, don't include the answer option "Other" for the same reason disclosed for pie charts. When to Use a Line Chart Single choice questions Multiple choice questions Arrays Mask questions (e.g. Gender, Yes/No) Line charts consist of data values that correspond to individual answer options and are best used to display trends, comparisons or peaks over time, therefore are great when running surveys repeatedly. General Tip: Reduce the number of labels on the x-axis to a maximum such as 5 to ensure readibility; however, make sure to have enough data points to enable proper extraction of trends. Also, adjust the relative height of the y-axis to match your highest value point within the line chart to improve the visual appeal of the chart. Furthermore, do not include the answer option "Other" for the same reason disclosed for pie charts. Chart Types in LimeSurvey In LimeSurvey, for each question that is created in a survey a chart type can be predefined in the statistics options of the respective question. If not changed, it will be set to bar chart by default. Not to worry though, in the statistics overview after collecting survey responses, you can simply and flexibly switch back and forth between all available chart types. Besides bar, pie, donut and line chart, LimeSurvey offers a few more chart types that are not as mainstream: radar (spider) chart and polar chart. Radar charts are a great way of visualising multivariate data, i.e., data with multiple variables, on two dimensions while the variables are represented by axes starting from the same point. Thus, radar charts are beneficial when other chart types such as bar charts become rather cluttered with an increasing number of variables. Polar charts are essentially radar charts since both display data as circular graphs and can be used nicely for visualising multivariate data. However, while radar charts connect the variable results to a "web", the polar charts display results in adverse bars within the circular diagram. General Tip: Keep the number of variables in your radar or polar chart to a maximum number of 8 to avoid chart overload and confusion. Advanced Tip: If you are dealing with large amounts of survey data in general, we recommend you consider creating pivot tables to get more efficient and structured data filtering options when generating flexible and powerful pivot charts. Conclusion Learning how to use these different chart types effectively will be of great benefit not just when presenting survey results, but also and more importantly when deriving actionable measures from these results. It is always important for you to consider that your audience may not be familiar with the survey setup, i.e., structure, variables and items in detail, which makes it essential to facilitate the readibility and interpretation of your survey results by choosing the most suitable chart to get your point across. Always remember: There is no one right chart to visualise your survey data. It ultimately comes down to weighing the pros and cons of each chart.  

It is said that the NPS value is heavily linked with revenue growth and, therefore, a key metric to establish customer loyalty by a provider of a product or service. The Net Promoter Score consists of a single powerful question that simply asks “How likely is it that you would recommend company/product/service XYZ to a friend or a colleague?”. It offers an array of answer options ranging from values 0 (“Not at all likely”) to 10 (“Extremely likely”). It was developed by Fred Reichheld and introduced in 2003.* How to calculate the Net Promoter Score First, your survey respondents are divided into three groups: Those that chose answer options 0 to 6 are categorized as detractors, i.e., people who are not particularly satisfied with your products or services. Respondents who chose 7 or 8 are labelled passives who are neutrally motivated users and later ignored when calculating the Net Promoter Score. Lastly, those who chose 9 or 10 are your promoters, i.e., your loyal and satisfied customers who would gladly recommend your company and its products and services to other people. The actual NPS calculation is done in two steps. First, you need to determine the percentage of promoters and detractors: Number of promoters / Number of total respondents * 100 Number of detractors / Number of total respondents * 100 Example: You have received a total of 100 responses on your survey. 30 respondents (detractors) chose 0 to 6, 30 respondents (passives) chose 7 and 8 and 40 respondents (promoters) chose 9 and 10. 40 promoters / 100 total respondents * 100 = 40% promoters 30 detractors / 100 total respondents * 100 = 30% detractors The second calculation step is the respective NPS equation: % promoters - % detractors = Net Promoter Score The Net Promoter Score can reach a maximum value of +100(%) if you receive 100% promoters and 0% detractors and passives or a minimum value of -100(%) if you receive 100% detractors and 0% promoters and passives. Let’s take a look at our previous example and calculate the Net Promoter Score. Example: We had 40% promoters and 30% detractors.   40% promoters – 30% detractors = 10(%) Net Promoter Score Note: There is broad uncertainty about whether the Net Promoter Score is displayed as a percentage or not. The fact is, there is no specific evidence so far to determine which is better. However, there are publications by the inventor of the NPS himself, Fred Reichheld, who specifies the NPS in percent, for example, see here which is why I decided to stick to percent. What is Considered a Good Net Promoter Score? This question is not easily answered or assessed in a standardised way despite being purely numeric in determination. As mentioned before, the NPS can range from -100 to +100. If you have an equal number of promoters and detractors, you will always end up with an NPS of exactly 0. Therefore, it makes sense to treat 0 as the mean NPS, making it an average score, neither very good nor bad. Anything below 0 should sound an alarm, since it indicates customer satisfaction levels are tending towards negative values, and an organisation’s products, services or processes require considerable improvement to avoid substantial customer churn. Obviously, strategy and efforts should be aimed at achieving the highest possible positive NPS score. According to global NPS standards, a score from 0 to 49 is considered “good”, a score from 50 to 69 is considered “excellent”, while a score of 70 and higher is considered “world class”. For benchmarking reasons, pages such as Customergauge.com offer great insights into well-known brands’ NPS results. For example, Tesla Motors currently has an NPS of 97, while Apple has an NPS of 89, Samsung, however, only manages an NPS of 70. Interestingly, the Apple iPhone has an NPS of only 55 and the Samsung Galaxy only 45, hence, making the overall brands more satisfactory from a consumer perspective than the companies’ products. Tip: If you are able to retrieve information on competitors’ or industries’ Net Promoter Scores, this may give you a good starting point to work towards gaining a competitive edge. Advantages and Disadvantages of the Net Promoter Score There are plenty of advantages of conducting a Net Promoter Score survey to gain knowledge on how satisfied your customers are: Setup and processing are easy and fast due to user-friendliness. NPS question is intuitive for most respondents. Shortness of survey reduces survey fatigue and results in better response rate. Studies by Satmetrix and Bain & Company have shown that NPS and revenue growth are correlated. There can be good benchmarking possibilities. Customer classification can be undertaken easily.   There are also disadvantages to the Net Promoter Score: It is criticized as being too general and unspecific, as it doesn’t reflect the reasons for low satisfaction of the detractors. It’s a snapshot of the general satisfaction level, but lacks natural motivation and plan what to do with the attained data. It fails to predict future loyalty trend. Tip: if you feel you are not getting enough information out of your NPS question data, simply extend your survey by questions such as ”Would you elaborate on why you chose {0,1,2,3,4,5,6,7,8,9,10}” and ”Which changes would have to be implemented regarding company/product/service XYZ in order for your recommendation to increase by just one point?”, each with an open text box. These additional questions extend your NPS status quo by asking customers specifically for improvement areas and enabling concrete action to spur growth.  LimeSurvey offers a Net Promoter Score Survey Template which includes the aforementioned questions. Since it’s a template, it is fully editable and serves as a general survey framework. If you want to give us feedback, comments, or suggestions regarding our Net Promoter Score survey template, feel free to send us an email. *Net Promoter Score is a trademark of Satmetrix Systems, Inc., Bain & Company and Fred Reichheld