In this post
How to conduct surveys
How a survey is developed and used is of huge importance to make sure that we can collect the data that we want. If a survey takes too long for a person to carry out or the questions are too confusing, then these problems will reflect in the data that we end up with. For the most part it is best to assume that people are very busy and make sure that a survey does not take very long to carry out: therefore, sticking to simple questions that require a simple yes or no answer where possible is the best way to conduct a survey.
At times you may need to ask questions that will give a longer answer; this is usually saved for last in a survey to make sure that people will not see this and then give up all together before any information is gained! A question that requires an answer in the form of a paragraph will normally be asking an opinion on something for which people will have an endless range of answers.
For the most part, it is best to stick to questions that will only need somebody to tick a box or give an easy one word/number answer. For example:
This question is straight to the point and easy to answer. As well as thinking about making the task easy for the participant of a survey, it should also be remembered that someone will need to organise the data once it has been collected. This is much easier if the questions are short, and it is possible to quickly scan and see an answer like a tick or cross instead of a long paragraph that needs to be read.
In certain situations it is necessary to have a question that will require a longer answer but any question should always be as easy as possible for the situation and the data that we are trying to find.
Things to avoid in survey questions:
- Incomplete questions or spelling mistakes
- Vague questions
- Leading questions
- Questions that are too intrusive and personal
- Ambiguous questions
Avoiding these pitfalls will make for an easier survey for both the person being interviewed and the survey maker. Here are a few examples of the above which should be avoided:
Leading questions in surveys
This question shows the opinions of whoever wrote it and tries to persuade the subject that their view is correct. You will find that a higher percentage of people would simply agree with what was said and choose Cats than if the question was unbiased. Therefore, a question which takes a clear side and asks for a person to agree with this side will end up giving you data that is not a fair reflection of what the person truly believes. A better way to write this question that would not lead the survey taker into one answer would be:
This question clearly does not lean to one particular answer and will therefore not persuade people to side with the opinion that is given.
Ambiguous questions in surveys
The problem with this question is that if you have visited the cinema a certain number of times (say 2) that you could actually tick more than one box. If I had visited the cinema twice I could tick both the first box (0 – 2) and the second (2 – 5) as my answer is contained in both of these. Therefore, the question is ambiguous and needs to be changed to make sure that this does not happen. Below is an example of how this question should be asked:
Looking at these categories now it is clear that, no matter how many times a person has visited the cinema, they can only tick one box correctly. This will give much more accurate data from the survey question.
Intrusive or personal questions
It is best to avoid intrusive and personal questions when making a survey as you do not want to be irritating! By avoiding rude questions an individual will be much more likely to help you with the study and will feel happier about taking the survey.
Asking a person’s age directly is considered by many to be quite rude and should therefore be avoided. However, it is possible to gather a rough estimate with a question that is a little nicer by grouping some ages and asking a person to tick which is appropriate for them. People will be much happier answering a question like this instead:
These things are how you want survey questions to be so that the person filling out the survey feels at ease and not pressured into a certain answer and can finish giving information as quickly as possible and get on with what they were doing. Obviously, it is sometimes needed to ask a longer question but these are usually saved for the end of a survey and ask for an opinion on how they think the survey went. This can give very valuable information on how to conduct future surveys.
The experimental method
A basic way to describe the experimental method is to say that we are carrying out experiments! By doing experiments we look to alter certain things while keeping others the same throughout. From this we can draw conclusions on how these changes have affected an outcome. This is called controlling variables, a variable simply being something which is not fixed and can be altered.
To successfully measure the effect that something has on an outcome we need to change only this one thing from one experiment to another and keep other things exactly the same. For example, if we wanted to measure the strength of two people we should make them carry out the same test with things like the temperature, method etc. being kept the same for both people. This will then mean that we have only changed the part of an experiment which we are looking at – the participant.
The observational method has much less input and is basically a way for people to just look at something in its natural environment without altering many things. This allows us to see what happens when there is essentially no meddling from the person observing.
In most situations (especially mathematics) we look to collect and work with quantitative data as opposed to qualitative. Quantitative data means something that can be expressed in numbers and qualitative data is expressed in words and can be much harder to graph and draw statistical conclusions from.
