POR – The Process
Sampling Methods
Your sample is simple the actual people who answer your questions. Sampling is the procedure through which you select those individuals.
There are two basics sampling categories: Probability and non-probability sampling.
Probability (Random) Sampling
Each element of the target population must have the same probability to be chosen. Any subjective moment must be eliminated from the sampling procedure. This is the preferred method by mathematicians and statisticians.
Unfortunately, reality is much more complex than the mathematicians would like, and in many occasions, we have to interfere with the pure concept of randomness.
Non-probability sampling
There is a variety of non-probability sampling methods. They all have in common that randomness is not required in the sampling process. We will study here the most relevant for our project.
Quota Sampling
This method has the same goal: to get a representative sample.
Quota sampling starts with a statistical analysis of your target population. On the basis of these data, you design the structure of your sample, making sure that it will reflect the structure of your target population.
Convenience sampling
This is the most common sampling method for projects in a learning environment. Since the main goal of the assignment is to learn the method, and the outcomes are not supposed to be published, we just recruit for the study whoever is available.
But even in studies intended to be published, convenience sample is commonly used. A high percentage of scholarly articles use college students as participants, even though it is clear that the outcomes cannot be applied to the entire population.
We will use this type of sampling in our projects.
Snowball Sampling
In this case, you start with small group who you know fulfills the criteria of your target population. Then, you ask your participants to identify other people who may belong to the same group.
You may also use the snowball strategy for your sample.
If you want to explore this subject, you can find in this Web-site more detailed information about the different sampling strategies.
Interview Types
The interview is the key moment of the process of gathering data.
Again, there is a variety of very well stablished procedures and strategies to collect the information. The classic ones are: Personal (face to face) interview, telephone Interview, mail interview and online interview.
Each one of the interview methods has its advantages and disadvantages. A brief summary follows:
Personal (face to face) Interview
Advantages:
Interviewer can show products, visuals, etc.
In some cases, sampling point helps you identify the target population.
Interviews can be longer.
Interviewee focuses stronger on the questions.
Best use of different types of graphic scales.
Disadvantages:
The most expensive interview format
Sometimes, the sampling point may not be representative of the target population.
Telephone Interview
Advantages:
Low cost.
Interviewee can be easily contacted.
Ideal for random sampling.
When using CATI software, the outcomes are available immediately.
Disadvantages:
You cannot interview an answering machine.
Not everybody is available during the normal working time.
You cannot use visuals, products, graphic scales, …
Telemarketer syndrome.
People tend to lose concentration on the phone.
Mail Interview
This type of interview actually belongs to the past.
Advantages:
Cost effective.
The questionnaire can include pictures and graphic material.
Interviewees can answer them at their convenience.
Disadvantages:
Time consuming (you will need several weeks before you get the responses back).
Non-response rates are very high.
Non-response is always significant.
Online Surveys
Online surveys constitute the present and the future of survey research. Precisely for this reason, we need to be aware of the weaknesses of this sampling method.
Advantages:
You can use audiovisual material.
Very fast.
Lowest cost.
People feel safer and give more honest answers to delicate questions (drugs use, sexual behavior, …)
You can easily integrate open-ended questions.
Disadvantages:
Internet users do not represent the whole population.
No control over the sample.
Huge non-response rates due to the communication overload.
More information about the different types of interviews in this link.
Translation Process
In most cases, it is necessary to translate the research question into several questionnaire questions. This translation process is necessary because of:
- Language related issues – People may not understand the language. Words can also have emotional or moral load.
- Level of abstraction – Research questions are sometimes too abstract. Interviewees’ answers are better the more accurate and concrete the questions are formulated.
- Private sphere – In some cases, the issue may inhibit the interviewees. In other cases, people do not want to give information about certain areas of their lives.
- Interviews are boring – and people get tired. Concentration weakens then.
- Social desirability – People tend to express opinions that they believe are in consonance with the general view.
- Need for recognition – We use every chance we have to show off.
- Self-illusions – people often have a distorted image of themselves. The have an “ideal I”, and answer according to this ideal – and not the actual I.
- Cognitive dissonance – In many cases, the impulse to avoid logical contradictions may also distort the outcomes of the survey.
- Subconscious depth – In many cases, we are investigating issues our target population may not be aware of.
- Sponsorship Effect – People tend to express opinions that they think will please the interviewer.