You can subsequently break down the result of a survey, using background questions such as age, sex, income, education, geographic location etc., which enables you to find correlations, understand a result etc.
Dropout refers to the respondents who do not answer, because they do not want to, are not at home etc. Dropout is often specified as a percentage. The larger the dropout, the lower is the reliability of the survey.
Filter questions are used to leave the respondent forwards in the questionnaire, such as “if you answered yes, please go to question 10″. Filter questions are used to guide some respondents (depending on background questions or the way they have answered) to a question which only refers to that group of respondents. This means that you get better answers, since the question is relevant to the people who answer, and other people don’t have to bother answering it.
Cross tabulation means that you can analyse individual groups, depending on how they have answered – or not answered. For example, you could take a closer look at the results of people who live in a large town, are women or watch a certain TV channel. Cross tabulation makes it easier to see correlations and analyse things at a micro level.
A quantitative method includes surveys, usually via the web, post or telephone. In a quantitative survey, you can postulate hypotheses and do a wide-ranging survey to chart a target group, questions etc.
The difference between a quantitative survey and a qualitative survey is that the qualitative survey frequently uses in-depth interviews and case studies. This creates a deeper understanding of a phenomenon.
The arithmetic average, the sum of all numbers divided by the number of participants, observations etc.
The “middle value”, the number in an ordered series, which divides the series in two equal parts. If the series has an even number of observations, the median is the average of two values on either side of the centre.
A group of people have something in common, such as all the people in Sweden or all the people who have a certain profession. In the wider statistical definition, a population is a number of units or phenomena which is being investigated such as visitors to a shop, or manufactured products.
Randomization can mean two things:
- Random allocation of participants between groups in a survey. Randomisation is necessary, so you can use statistical methods to assess the probability that the result of a survey has been obtained through the action of chance. Randomisation can be done by allowing a computer to generate a random sequence (randomisation code) which decides the group into which each new participant will be allocated.
- Random order of questions, response alternatives etc. The order of questions and response alternatives in a questionnaire can sometimes guide the respondent to answer in a certain manner. Randomisation avoids this.
Reliability describes how trustworthy an investigation/survey is. If you measure the height of children in a school class, the measurement instrument affects the reliability. If you measure the distance between Gothenburg and Stockholm with a ruler, reliability is low.
A respondent is a participant in a survey. People who have finished the survey and answered all the questions are called “complete respondents”.
Raw data is data which has been acquired from a source and which has not been processed or been subjected to any manipulation, such as un-processed statistics from a survey. The data can then be processed by a computer program or by manual analysis. A distinction is made between data and information; information is the end product from data processing.
Significance is a concept which specifies that an observed value in a survey deviates from a hypothetical value, or other comparable value, so much that it is not caused by statistical uncertainty (chance).
Significance is classified in the form of confidence interval, where 95% (5%) indicating that the observed deviation in 19 causes of 20 is not due to chance and 99% (1%) in 99 causes out of 100.
By exporting your data to the statistical program SPSS, you can make a broader and deeper examination of the results. Are you using Survey Generator Medium or Large, you can get the data in this format.
Response rate refers to the number of completed respondents in relation to the number of possible answers. No response is called dropout.
Validity means that you have investigated what you were supposed to investigate and nothing else. It may seem obvious but it is not always easy to achieve high validity. If you measure the feet of a large number of individuals, to measure how fast they run, the survey may have an excellent reliability. If the instruments are perfect you have in the end a proper disposition of all lengths of the feet. The investigation however, completely lacks validity because there is no correlation between size of feet and how quickly we can run. An investigation can thus have a high reliability but still no validity.
About web surveys
How can I find representative addresses for various groups in society?
We collaborate with the CINT company, which has a panel system that contains a number of different types of panels, containing people who have said that they are willing to answer surveys. You can make a selection, using a number of different parameters such as where people live, age, sex and income, but also using softer background data such as interests, family membership or living conditions. If you want to create your own panel, you can use our Panel Module, where you can easily administer this.
Which is better – HTML or plain text distributions?
Briefly, you can say that the choice is between very secure text format or inspiring, attractive HTML format.
- Plain text is simpler (lacks images, colours, fonts etc.) and can be received by everybody. A plain text distribution runs a lower risk of ending up in the trash can.
- Using HTML, you can design your distribution to conform with the company’s profile, and make it more easy to read, which can increase the response frequency. There is a greater risk of ending up in the e-mail trash can, than if you use text format.
How many responses can I expect?
This is almost an impossible question to answer, somewhere between 10 – 90 percent. The response frequency depends on things like the type of survey, your target group, the relationship you have with the target group and any incentives (lotteries etc.). For example, you will get a higher response frequency from a survey of your colleagues than from a customer survey.
How can I increase the response frequency?
You can increase the response frequency by being clear about the purpose of the survey, sending some pleasant reminder mails and of course, mentioning if the respondents will get something to thank them for answering the survey.
How long should a survey be open?
The survey time depends on the number of responses you want to receive. You will commonly get most responses during the first day, but about three weeks is a good “open time”, which gives everybody time to answer. You can generate reports during the time that the survey is open, to see the tendencies in the responses.
How many times should you remind respondents to answer?
This depends on how long you want to have your survey open, but one or two reminders should normally be enough. It’s important that you don’t start nagging. This could get people to answer irresponsibly, just for the sake of answering, or feel so hunted that they actually refuse to answer.
About the Survey Generator tool
What can I use Survey Generator for?
You can more or less use Survey Generator for all types of email distributions or questions that you want to ask via the web. Among other things, we help our customers with: employee surveys, attitude and brand surveys, customer and member communication, panels, web competitions, course and project evaluations, applications and invitations, news shots and many other things.
Why should I use Survey Generator, in particular?
- User-friendly: Survey Generator is recognised as being intuitive.
- Functionality: It is suitable for both simple, internal surveys and for complex, professional surveys.
- Adaptable: Alstra is happy to do configurations, provide extra functionality or integration with your intranet etc.
- Cost-effective: You only pay for what you need.
How quickly can I start working in Survey Generator?
You can easily start at once, without special training. It is quick and easy to put in questions and respondents, and then distribute the survey. A wizard in the tool helps you to create your questionnaire, step-by-step. We also offer training courses, if you need extra support or want to learn more.
Some of the people who will answer the questionnaire don’t speak Swedish. How do I solve that?
You can set up questionnaires in almost 30 languages in Survey Generator, and the list of languages is extended all the time. Languages can be handled in two ways:
- Configured: You ask the respondent at the beginning of the questionnaire about the language he/she wants to answer in. The questionnaire language is then configured to suit his/her choice.
- Pre-selected: If you know in what language the respondents want to answer you can choose language in the survey.
Some of the people who will answer the questionnaire may have reading or writing disabilities or sight loss. How can I reach them better?
Survey Generator can be adapted to people with reading or writing disabilities or visual impairment so that they can answer online surveys. The optional ReadSpeaker allows the respondent with a simple click to get questions and answers read out loud.
I need help with my survey, what do I do?
If you need technical support, you can always send e-mail email@example.com (Survey Generator Small) or call +46 (0)8-22 22 65 (Survey Generator Medium and Large).
If you do not have time to create your survey or manage the entire investigation, we can help you. Please contact us for discussion of the approach and quote.
I do not have e-mail addresses to some of my respondents in the survey, how can I solve it?
- You create your survey in Survey Generator.
- We create a login page to the survey.
- You send a postcard to those without e-mail, with information about the link to the login page, or alternatively, a postal questionnaire containing information about the login page (you can use the Survey Generator and register the answers from the postal questionnaires that you receive).