Data are the raw materials for research. Investigators generate data through some processes of measurement, counting, or observation. The tools and techniques employed to collect data depend largely on the study’s objectives.
What Data Collection Tool Do You Need?
Whenever we plan for data collection, we must ask ourselves the following questions:
- What information do we want to collect to answer our research questions implied in our research objectives? This refers to the selection of the variable;
- What approach will we follow to collect this information? This refers to the choice of the study design;
- What techniques and tools will we use to collect data? This refers to the data collection technique;
- Where do we want to collect data? This refers to the population to be sampled;
- What part of the population will be included in our investigation? This refers to the
- How many subjects will we include in our study? This refers to the sample size;
- How will we select our sample? This refers to the sample design.
This post is designed to concentrate on the third point, the data collection techniques.
Data collection may range from a simple observation at one location to a grandiose survey covering a wide area in any part of the universe. The method selected for data collection will largely determine how the data are collected.
Questionnaires, standardized tests, observational forms, tape records, or checklists are some devices used to record the data.
Data collection techniques enable us to systematically collect information about our objects of study (people, objects, phenomena) and the setting in which they occur.
In the collection of data, we have to be systematic. If data are collected haphazardly, it will be difficult to answer our research questions conclusively.
The same measuring instrument, the same operational definition of variables, the same unit of measurement, etc., must be maintained at all stages of data collection.
There are many different ways to collect data. The approach selected largely depends on
- The study objectives.
- The study design.
- The availability of time, money, and personnel.
An important consideration in deciding on the best way to collect data is whether the study is intended to
- Produce relatively precise quantitative findings, or
- Produce qualitative, descriptive, or narrative information.
Keeping these two points in view, the data collection techniques are broadly two types;
- Quantitative Data Collection Technique,
- Qualitative Data Collection Technique.
Quantitative Data Collection Technique
Most social science research is based on 3 survey methods of quantitative data collection techniques.
They are:
- Personal interview method;
- Self-administered questionnaires method, and
- Telephone interviewing method.
Qualitative Data Collection Technique
The data collection techniques most appropriate for studies whose objectives call for descriptive qualitative analysis tend to differ from those most appropriate for quantitative methods.
Quantitative methods are important to obtain data for making predictions, probabilistic statements, and generalizations.
Methods of qualitative data collection techniques are;
- Unstructured interviews,
- Focus group discussions,
- Observations,
- Document study, and
- Content analysis.
Some additional data collection techniques are;
- Nominal group technique.
- Delphi technique.
- Life histories.
- Essays.
- Case studies.
- Mapping.
- Rapid appraisal techniques or soundings.
- Panel study.
- Key informant approach.
Life Histories
A special application of the interview technique is the use of life histories. This technique allows people to tell stories, which provide insight into what they consider important in their views.
The life history technique is a special technique of interviewing, which employs a very limited sample, not exceeding 25. This technique fits well in traditional rural societies.
Issues that are especially suited for investigation using the life-history approach include, for example, the pattern of reproduction and women’s feelings about marriage, childbirth, and contraception.
Essays
Some anthropologists have used school children’s essays to explore their subjects’ (hidden) values and aspirations and the community they live in.
Essays may be analyzed to determine differences in beliefs concerning perceived causes of illness, the value of children in traditional societies, the benefits of having a small family, the rationale for health-related behavior, and the like.
Case Studies
A case study involves detailed investigations of a few people, a family, a community, or a particular situation. It falls under a qualitative study.
According to Young, a case study explores and analyzes a social unit’s life, be it a person, a family, an institution, a cultural group, or even an entire community.
Usually, several methods of data collecting information are used simultaneously. The subjects of study are often chosen using non-probability sampling.
For example, the cases may be selected to be typical or illustrative of a particular phenomenon or group. A researcher uses case studies for the following reasons:
- Explore new areas and issues where little theory is available, or measurement is unclear.
- To explain a complex phenomenon.
It is usual for researchers to combine case studies with quantitative analyses that use a larger data set.
Mapping
Mapping is a valuable technique for visually displaying relationships and resources. It is also useful and often indispensable as a pre-stage to sampling.
In installing CNG stations, for example, the mapping may be an invaluable guide.
Service Statistics
All national family planning organizations generate service statistics. Some organizations have established a management information system (MIS).
The quality of service statistics, however, varies from country to country and even within countries. Thus, they should be used with caution. Service statistics often help a researcher define the parameters of the problem he or she wants to study.
In some cases, they can be used to compare the results of a particular study with nationwide figures.
In operations research projects, it is often necessary to design supplementary forms to provide data unavailable from the regular service statistics.
Panel Study
Surveys are generally conducted at one point in time for obvious reasons.
However, there are several disadvantages to collecting data at only a single point in time.
Chance fluctuations may well distort the data.
Also, a cross-sectional survey, conducted at a single point in time, offers no way to study trends in the data or seasonal variations or tell whether a relationship between the two or more variables will remain the same and change with time.
A longitudinal study is a neat solution to the problem if the sample is not too large.
A common longitudinal survey is the panel study, in which the same respondents are re-interviewed at two or more points regarding the same problems.
A panel study is one of the best ways to measure changes among the same subjects over time. In such a study, we first sample a group at one point in time and then return at a later time to ask the same questions again.
Then by bringing together the responses of subjects, we see whether a characteristic or attitude continues or whether it is taken up and dropped over time.
The disadvantages of a panel study, among others, are as follows:
- Since the interview is conducted several times, the cost involved is much greater than a cross-sectional study.
- The respondents may be reluctant to participate in repeated interviews.
- Some respondents may not be available for successive interviews.
Another longitudinal survey is the trend study, in which the same number of respondents from the same population is surveyed each time, but not necessarily the same.
This design compares similar data collected in different years (and on different subjects).
This year’s contraceptive usage data could be set against data from the last 5, 10, or more years. In this way, the comparison could be made across time.
The disadvantage of this design is that differences in the data from one survey to the next may not be the result of a trend but merely the reflection of differences in the persons surveyed.
Key-informant Approach to Data Collection
The key-informant Interview (KII) approach to needs assessment assumes that certain individuals know the community and target population.
It further assumes that these individuals are in a position to articulate the community needs to assist in program planning accurately. Key informants are often interviewed utilizing a semi-structured format.
The person or persons selected to be key informants must have a broad knowledge of the community, its social structure, the services to be provided, and its people.
It is an excellent way of recovering information about past events or ways of life that are no longer observable.
The needs assessment objectives can help identify the most appropriate kind of person to act as a key informant.
The researcher might consider public officials, long-time residents, business managers, administrators, church leaders, and persons representing a variety of lifestyles, ages, viewpoints, or ethnic backgrounds.
Few people in a community can speak about everything; therefore, the problem should be focused on before selecting the informant.
A variety of methods can be derived from working with key-informant. Questions can be developed in advance; they may be outlined or unstructured.
The key-informant method is especially useful for
- Obtaining a deeper knowledge of minority viewpoints
- Raising citizens’ consciousness about a community problem.
- Involving citizens in public problem-solving who would be less inclined to answer a questionnaire, etc.
- Data Quality Check
There are several ways to check the quality of the interview data.
- Sometimes a researcher will deliberately ask two or more questions that give the same type of information. The first question might be asked at the beginning of the interview and the second at the end. The two questions are then examined for consistency of response. This is one way to check the reliability of the data.
- For difficult questions, sensitive questions, or questions where the researchers want to be sure the information is correct, the interviewers can be instructed to probe. That is to say; the interviewer can repeat the question in a slightly different form or repeat the respondent’s answer and then ask if the information is correct. For example, a woman might report having two sons and three daughters. The interviewer might say, ‘You have five children, two males and three females. Is that correct? Are there other children you may have forgotten to tell me about?”
- Field supervisors should help the interviewers with difficult situations and check that they are doing their work. Some studies use a ratio of one supervisor to every five interviewers.
- Most studies that use an interview procedure attempt to reinterview a certain percentage of the respondents. Depending upon the sample size, a general rule is to re-interview between 5 and 10 percent of the sample. The data from the first interview are then checked against the data from the second interview for consistency. If there are major inconsistencies, particularly on such questions as age, marital status, and parity, then clearly, there is a problem somewhere. The problem might be with the questionnaire, with the interviewers, with the tabulation procedures, or in some other areas.
- Once the data have been collected and tabulated, it is possible to do statistical checks for errors or consistency. For example, a frequency distribution of the parity of women may reveal that several women claim to have 18 or 19 living children. Since this is highly unlikely, the investigator is faced with the choice of discarding these questionnaires, eliminating the information from the question on parity, or going back and re-interviewing the women who claim to have 18 or 19 children.
The following table summarizes the advantages and disadvantages of various data collection techniques:
| Techniques | Advantages | Disadvantages |
|---|---|---|
| (a) Observation (in particular of behavior) | Gives more detailed and context-related information. Permits collection of information on facts not mentioned in the questionnaire. Permits tests on the reliability of responses to questionnaires | Observer bias (researcher may only notice what interests him. Ethical issues concerning confidentiality or privacy may arise. The Presence of data collectors can influence the situation observed. Thorough training of research assistants is required. |
| (b) Interview | Suitable for illiterates. Classification of responses is possible. Higher response rate than a written questionnaire. | The presence of a researcher can influence the responses of respondents. Reports of events may be less complete than information gained through observation. |
| (c) Small-scale, flexible interview | Extra information can be gathered through observation. | Observer bias if the information is not scheduled. More difficult to analyze in a standard way. |
| (d) Large-scale fixed interview | Easy to analyze | Easy to miss important information |
| (e) Written or mailed questionnaire (selfadministered) | No need for research assistants to administer it. No bias in asking the questions differently. Anonymity: more honest responses. | Not possible to use with illiterate respondents Low response rate Questions may be misunderstood. |
| (f) Content analysis | Inexpensive. Permits examining past trends. | Data is not always easily accessible. Ethical issues concerning confidentiality may arise. Information may be imprecise or incomplete. |
The distinction between data collection techniques and data collection tools:
| Data collection techniques | Data collection tools |
|---|---|
| (1) Content analysis | (1) Checklist, data compilation form |
| (2) Observation | (2) Eyes and other senses, pen, and paper watch scales, microscope, etc. |
| (3) Interview | (3) The interview schedule, checklist, questionnaire, tape recorder |
| (4) Administering written questionnaires | (4) Questionnaire |