|Year : 2017 | Volume
| Issue : 2 | Page : 58-63
Data collection methods in quantitative research
Professor, College of Nursing, CMC, Vellore, India
|Date of Web Publication||9-Jun-2020|
Source of Support: None, Conflict of Interest: None
The information provided by the study participants on specific area of research called the data are very important that enable accurate information on the research work done by nurse researchers. Data collection methods are used to collect data in a systematic way. The researchers choose and use various data collection methods. They are broadly classified as self -reports, observation, and biophysiologic measures. This article highlights on the sources of data and on the various data collection techniques which include interviews, questionnaires, scales, category system and check lists, rating scales, and biophysiologic measures. It also analyses the advantages and disadvantages of each of these methods. Emphasis should be given on choosing appropriate method to collect accurate information which will lead to good quality research outcomes.
Keywords: data, quantitative, interview, questionnaire, observation, biophysiologic
|How to cite this article:|
Sadan V. Data collection methods in quantitative research. Indian J Cont Nsg Edn 2017;18:58-63
| Introduction|| |
In quantitative research process, data collection is a very important step. Quality data collection methods improve the accuracy or validity of study outcomes or findings. Nurse researchers have emphasized on the use of valid and reliable instruments to measure a variety of phenomena of interest in nursing. We need to be aware of the various measurement methods which are of importance to generate evidences needed for nursing practice. Researchers must choose appropriate data collection methods and approaches. An ideal data collection procedure captures a construct that is accurate, truthful, and sensitive (Polit & Beck, 2017). Quantitative data are collected in a more structured manner as compared to the qualitative data which are unstructured or semi-structured.
| Data Collection and Data Resources|| |
Data collection is a real challenge for researchers and it requires much time and effort. Nurse researchers should first identify the type of data to be collected and the sources from where they can be collected. The data sources can be either the existing data or the new data. Existing data such as from the existing records and documents can be of great value in some of the research studies. For answering the research question, if existing data are unavailable, researchers need to collect new data. The type of data also can be classified as primary data and secondary data. Primary data collection involves data collected directly from the study participants by the researcher or a trained data collector which includes surveys, questionnaires, interviews, observations, or biophysiologic measures. Secondary data collection is the use of data that were collected for another purpose such as patient’s records, government data bases (Houser, 2011). After identifying what data need to be collected, the nurse researchers must choose the data collection methods and develop a data collection plan. The decision about choosing the data collection methods also should be based on the ethical guidelines, the cost, the time constraints, population appropriateness as well as the availability of research assistants to collect data.
| Data Collection Methods|| |
The most commonly used data collection approaches by nurse researchers include self-reports, observation, and bio- physiological measures. Whatever approaches the researcher uses, the data collection method differ along the four important dimensions: structure, quantifiability, researcher obtrusiveness, and objectivity. The data collected in quantitative studies are based on a structured plan which guides the researcher to what data to be collected, how long, and how to collect them. The information gathered must be quantified by doing statistical analysis. There is a possibility that the study participants can change their responses or behavior under some circumstances which need to be taken in to consideration. The data collected should be objective wherein similar observations are made even if two researchers observe the same concepts of interest. Data collection is an important component in creating relevant research evidence and hence need to be carried out rigorously. The most common methods of data collection are discussed below (see [Figure 1]).
1. Self - reports
Structured self- reports are the most commonly used data collection method among the nurse researchers (Polit & Beck, 2017). The self-report instruments are interview schedule, questionnaires, and scales. In an interview schedule, data are collected by asking questions orally either face-to face or through telephone. In a questionnaire or Self- Administered Questionnaire (SAQ), the study participants complete answering questions themselves, either on a paper or onto computer (Polit & Beck, 2017). Scales are also a form of self-report in which the phenomena of interest are measured (Grove, Burns, & Gray, 2013).
Interviews can be used in descriptive studies and qualitative studies. They can be unstructured in which the content is controlled by the study participant or structured in which the content is similar to that of a questionnaire, with the possible responses to questions that are carefully designed by the researcher (Creswell, 2014).The questions can be closed, open ended, or probing. The interview questions are developed before the researcher begins data collection and are arranged in a logical sequence. The questions are asked orally to the participants, and explained if clarification is required by the participants. The order of questions arranged should be from broad to specific and topic wise. The sensitive nature of questions should be asked at the end. The vocabulary and sentence structuring of the question should be at the reading and understanding level of participants. After the interview questions are developed, the instrument is validated by experts for content. It has to be field tested through pilot study in order to identify problems in the design of questions, sequencing of questions, or procedure for recording responses. Piloting also provides an opportunity to establish the reliability and validity of the interview instrument (Grove et al., 2013).
The researcher or the data collector need to be skillful in obtaining interview data. They need to be familiar or trained with the content of the interview. The questions should be asked in a clear and unambiguous manner. The verbal and nonverbal communication must be unbiased. Sometimes the interviewer may have to repeat questions or explain questions, or probe to get more information. Probing has to be done carefully to avoid biased responses. The data can be written or recorded responses. It has to be done without distracting the interviewee. Prior permission should be obtained from Institutional Review Board (IRB) as well as from the study participants before the data collection procedure can be initiated.
In the present scenario, telephone interviews are used widely and it has been a convenient method to collect data. The interviewer has to be more tactful while collecting data through telephone interviews. They are faster and less expensive. However, the data can be collected only during a limited duration of time. The challenge is to ensure the true identification of the respondent, and can also cause bias and trigger unwanted responses since the researcher is not present physically. Focus group interviews, another method of data collection help to get aggregate perception of people, their feelings as well their thinking (Murugan, 2015). The researcher or the data collector who conducts the interview should get adequate training to ask questions in a logical sequence, and how to handle unanticipated questions or answers that might arise during the interview process.
Advantages and disadvantages
Through face to-face interview in-depth information can be obtained and it is more often a very flexible technique. In fact, the questions are restructured during the interview. It is an opportunity to obtain personal information and responses to all questions can be obtained. It provides an opportunity for probing and complex answers can be obtained. The response rate is higher and interviews provide a more representative sample. Data can also be collected from sick individuals, and from those who have problems with reading, writing, or difficulty in expressing. However, interview has its own disadvantages. It requires more time and it is expensive. As it needs more time, the sample size is usually minimized. Subject bias is always a threat to validity of the findings and its consistency in data collection from one subject to another (Grove et al., 2013). Interviewing children and lack of language skills are the other challenges faced while using interview schedule as a data collection technique.
The most common instrument used for data collection is questionnaires. The participants fill in their responses themselves on a paper pencil instrument or on computer directly. Questionnaires can be structured or unstructured. In structured questionnaires, both the questions and the responses/answers are provided and the study participants need to pick up the correct responses. In unstructured questionnaires, the participants are required to give their own responses to the predetermined questions. Structured questionnaires can consist of either open ended or closed ended questions. In open ended questions, the participants provide their own answers in narrative form whereas in closed ended questions, there are fixed answers to the questions and the participants need to choose the correct/best response (s).
Good closed ended questions are more difficult to develop than open ended questions but are easy to analyze. Open ended questions can yield rich information, provided the participants are expressive and cooperative. While constructing structured questionnaires, the researchers must be careful with the wording of questions for clarity, sensitive to participant’s psychological state, ensure absence of bias, and consider the reading level of participants (Polit & Beck, 2015).
Closed ended questions are of various types which are given below (Houser, 2011; Polit & Beck, 2017).
- Dichotomous questions: The participants decide between two choices of answers/responses such as yes/no and these type of questions are useful in collecting facts and provide only limited information.
- Multiple choice questions: There will be four to seven alternative responses to each of the questions, and the study participant chooses the responses as their answers. These type of questions are useful in collecting people’s opinion and views.
- Rank order questions: The participants choose their answers along a continuum such as most to least important. When we use these type of questions clear instructions should be provided.
- Forced-choice questions: In this type of questions, participants are required to choose between two related statements.
- Rating questions: Here, the participants need to evaluate something on a given ordered continuum of responses. The rating questions can be from 0 to 10
- Check lists: In this type of instrument, a series of questions are listed and arranged vertically and the responses are also listed along with the other.
- Visual Analogue Scale (VAS): The subjective experiences such as pain are measured using VAS. It is a straight line and the end of the line are labelled as the extreme limits of the participant’s experiences or feelings.
When developing a questionnaire, the researcher has to first identify the information that need to be collected. A blue print has to be made including the different aspects of the topic. A literature review on the pertinent topic will guide the researcher in the development of questionnaires. Each question should be carefully designed and clearly expressed according to the level of the participants. They should not be ambiguous or vague. Long question can threaten the validity of the instrument (Grove et al., 2013). The instrument should have proper instructions on how to fill the responses. The validity and the reliability of the developed questionnaires should be established and the instrument must be pilot tested before it is used in the study.
Advantages and disadvantages
Self-reported questionnaires are administered either in person or through e-mails. Questionnaires usually appear easy to develop, but it requires much time and effort. They are less expensive, involves less time and less energy to administer as compared to interview schedule. Electronically mailed questionnaire is faster and cheaper too. A larger number of samples can be included in the study, and it provides an opportunity for complete anonymity in data collection. There is less possibilities for interviewer bias. However, information may be incomplete leading to missing data. The response rate especially when data are collected through posted mails and e mails is less.
Scales are a form of self-report and are a more precise form of measuring a phenomena than questionnaires (Grove et al., 2013). Scales measure the characteristics or traits of human beings in which more emphasis is placed on verification of reliability and validity. They are also called as psychometric instrumentation (Houser, 2014). There are existing scales available. If located, the researcher should get the psychometric properties of theses scales and document them. Psychosocial variables such as pain, nausea etc., are commonly measured using scales. Scores are given to each of ? the items to be measured in the scale. The types of scales commonly used in nursing studies are rating scale, Likert scale, Sementic Differential scale (SD) and Visual Analogue Scale (VAS) (Grove, Gray & Burns, 2015), which are discussed below (see [Figure 3]).
- Rating scale: In a rating scale, an ordered series of categories of a phenomena being studied is listed. A numerical value is assigned to each of the categories and the distinction between the categories vary. On this scale, the participants choose the best catergory that fits their experience. They are easy to construct but should be careful with extreme statements. e.g., Faces Pain Scale
- Likert scale: It is the most commonly used scale which contains many declarative statements that express the view point on a topic. The study participants are expected to indicate the degree to which they agree or disagree with the view point expressed in the statement. Likert scale usually consists of 10 to 20 items, each addressing an element of the concept which is being measured. The scale has both positively and negatively worded statements. The responses provided by the participants are scored and the total scores are summated
- Semantic Differential Scales: Psychological characteristics of people are measured by semantic differential scales. It is used to measure the variation in the views of a phenomena of interest. It is a 7 point scale and one end is the most positive and the other margin is the most negative. Each line is considered as one scale and the scores are summed up.
- Visual Analogue Scale (VAS): VAS is used to measure magnitude, strength, and intensity of people’s feelings, sensations, or attitude about symptoms or situations. In VAS, there is a vertical or horizontal line with descriptors at both ends, as well as range of possible feelings of participants along the line. The participants need to place a mark on the line to indicate the intensity of their sensation or feeling, e.g., Visual Analogue Pain Scale. It is commonly used in health research.
2. Observational Methods
Observation technique is used to record the specific behaviors, actions of people and events. Observational measurement can be unstructured or structured. Unstructured observations are done spontaneously and recorded as what is seen in words. Whereas, in structured observations, the researcher should carefully decide what to observe, how to observe, how long, and how to record the observed data. Observational measurement are usually more subjective than other methods of data collection. However, in some situations, observation may be the only way to collect information. In structured observations the specific behaviors of study subjects or the events to be observed or studied should be carefully defined.
In observation measurement, the observer plays an important role. A participant observer plays an active role and take part in the activity or event being observed. The nonparticipant observer adopts a passive role while observing the phenomena of interest. If observations are done by more than one data collector, establishing interrater reliability is vital. The most commonly used observation methods are discussed below (Polit & Beck, 2017):
2.1 Category systems and check lists
Behaviors, events or attributes of the subjects to be studied are grouped into categories and the categories are observed and recorded. The categories must by explained clearly. The maximum number of categories for effective observation is 15 to 20. The observer makes inference from the recorded observation from the category (Grove et al., 2013).
Behavior of participants are observed to see whether the behavior occurred or not and is recorded as tally marks in various categories. In observational check list, single category is selected for observation. Check lists are designed using category system. For example, while measuring the behavioral indicator related to pain, it can be cry and facial expressions which are called as categories. The facial expressions are measured by checking whether brow bulge, eye squeeze, and naso- labial furrow occurred or not
2.2 Rating scales
Rating scales permit the observers to rate the behavior of the participant or the event on a scale at specified time intervals and then it is quantified. However, rating scales provide more information than check lists. If they are combined with category system and check lists, the data collected will be much useful in studying the phenomena. Rating scales can be used for observation as well as for self - reporting.
3. Physiologic measurement
Many of the nursing studies have included physiological measures to assess the outcomes of nursing care. Today, nurse researchers use different kinds of bio-physiologic measures in research. For example, Yeo (2009) examined the effects of a walking versus stretching exercise on preeclampsia risk factors such as heart rate, and blood pressure in sedentary pregnant women as cited in Polit and Beck (2017). Biophysical and bio-chemical measures are the two categories included in the physiological measurement. Use of stethoscope and sphygmomanometer to check the blood pressure is an example of bio-physical measurement and laboratory value of blood sugar is an example of bio-chemical measure. Physiological measurements are either direct (body temperature) or indirect (blood sugar). Physiological measures can be obtained through self- reports, observation, laboratory tests, and electronic monitoring. Example: irregular heartbeats can be self reported by subjects, observed by the nurse as well as monitored electronically (Grove et al., 2015).
In this method of data collection, use of specialized equipment are needed to measure the study variables. The two types of biophysiological methods used for data collection include in vivo and in vitro. In vivo measurements are done directly in or on living organism, whereas, in vitro measurements are performed outside the body as in case of checking the blood sugar level (Polit & Beck, 2017). In many studies, nurse researchers link physiological variables with psychological and social variables such as linking stress with blood pressure measures over a period of time.
Advantages and disadvantages
The data obtained through laboratory tests and electronic monitoring provide precise and accurate data and are direct measures of many physiological variables. Bio-physiologic measures are more objective. Since the data collection setting is hospitals, the cost involved in collecting bio-physiologic information may be low. However, there are few disadvantages of using these measurements. The measuring instrument itself can affect the study variables. There are possibilities of related risks while applying energy and instruments. Special care must be taken in selecting appropriate instruments in relation to practical, ethical, medical, and technical considerations.
| Conclusion|| |
Data collection methods play a vital role in generating evidence through research. Each of the measurement approaches have their own advantages and disadvantages. Researchers must identify the type of data that need to be collected. Importance should be given to make sure that the data collection techniques are carefully chosen, applied, and properly managed to provide accurate information which can support the quality of research work done by the nurse researchers. Nurse researchers should be aware of the different data collection approaches and need to get familiarized with them.
Conflicts of Interest: The author has declared no conflicts of interest.
| References|| |
Creswell, J.W. (2014). Research design: Qualitative, quantitative and mixed methods approaches (4th ed.)
. Thousand Oaks, CA: Sage.
Grove, S.K., Burns, N., & Gray, J. (2013). The practice of nursing research: Appraisal, synthesis, and generation of evidence
(7th ed.). St. Louis: Elsevier Saunders.
Grove, S.K., Gray, J., & Burns, N. (2015). Understanding nursing research: Building an evidence based practice
. Kundli: Reed Elsevier India.
Houser, J. (2011). Nursing research: Reading, using, and creating evidence
. Kundli : Sanat Printers.
Murugan, M.G. (2015). A critical analysis of the techniques for data gathering in legal research. Journal of Social Sciences and Humanities, 1
Polit, D.F., & Beck, C.T. (2015). Essentials of nursing research: Appraising evidencefor nursing practice
. New Delhi: Wolters Kluwer.
Polit, D.F., & Beck, C.T. (2017). Nursing research: Generating and assessing evidence for nursing
. Philadelphia: Lippincott Williams & Wilkins.
[Figure 1], [Figure 2], [Figure 3]