Chapter 9 Quan Data Collection and Sampling PDF

APPRAISING SAMPLING AND DATA COLLECTION IN QUANTITATIVE STUDIES SAMPLING IN QUANTITATIVE RESEARCH B. Sampling Designs in Quantitative Studies A. Basic Sampling Concepts Nonprobability Sampling Populations...

APPRAISING SAMPLING AND DATA COLLECTION IN QUANTITATIVE STUDIES SAMPLING IN QUANTITATIVE RESEARCH B. Sampling Designs in Quantitative Studies A. Basic Sampling Concepts Nonprobability Sampling Populations nonrandom methods in which every Population is the entire group of interest. element does not have a chance to be Researchers specify population included. characteristics through eligibility criteria. less likely than probability sampling to Researchers establish criteria to produce representative samples determine whether a person qualifies as most research samples in nursing and a member of the population (inclusion other disciplines are nonprobability criteria) or should be excluded samples (exclusion criteria). Quantitative researchers sample from 1. Convenience Sampling an accessible population in the hope of selecting the most conveniently generalizing to a target population. available people as participants The target population is the entire The problem with convenience population of interest. sampling is that people who are The accessible population is the portion readily available might be of the target population that is atypical of the population. accessible to the researcher. The price of convenience is the risk of bias. Samples and Sampling Convenience sampling is the Sampling involves selecting cases to weakest form of sampling, but it represent the population—a sample is a is also the most commonly used subset of population elements. method. In nursing research, the elements (basic units) are usually humans. 2. Quota Sampling A criterion for judging a sample is its researchers identify population representativeness. strata and figure out how many A representative sample is one whose people are needed from each characteristics closely approximate stratum. those of the population. By using information about the Sampling bias is the systematic population, researchers can overrepresentation or ensure that diverse segments underrepresentation of a population are represented in the sample. segment on a characteristic relevant to Quota sampling is similar to the research question. convenience sampling: Participants are a convenience Strata sample from each stratum. Strata are mutually exclusive segments of a population based on a specific 3. Consecutive Sampling characteristic. involves recruiting all people Strata can be designated in sample from an accessible population selection to enhance the sample’s over a specific time interval or representativeness—elements in each for a specified sample size. stratum can be sampled in the correct Consecutive sampling is often proportions. the best possible choice when there is “rolling enrollment” into an accessible population.  4. Purposive Sampling Evaluation of Nonprobability and Probability involves using researchers’ Sampling knowledge about the population Probability sampling is the only viable method of to handpick sample members. obtaining representative samples. If all elements Researchers might decide in a population have an equal chance of being purposely to select people selected, then the resulting sample is likely to do judged to be knowledgeable a good job of representing the population. about the issues under study. Probability sampling also allows researchers to This method can lead to bias estimate the magnitude of sampling error, which but can be a useful approach is the difference between population values and when researchers want a sample values. sample of experts. Nonprobability samples are rarely representative of the population— some segment of the Probability Sampling population is likely to be underrepresented involves random selection of elements The quality of the sampling plan is of particular from a population; each element in the importance when the focus of the research is to population has an equal, independent obtain descriptive information about prevalence chance of being selected. or average values for a population. The quality of the sampling plan is of particular 1. Simple Random Sampling importance when the focus of the research is to most basic probability sampling obtain descriptive information about prevalence Researchers using simple or average values for a population. random sampling often establish a sampling frame—a list of C. Sample Size in Quantitative Studies population elements Sample size—the number of study Elements in a sampling frame participants. are numbered and then a table When researchers calculate a of random numbers or an online percentage or an average using sample randomizer is used to draw a data, the purpose is to estimate a random sample of the desired population value, and larger samples size. have less sampling error. Samples selected randomly are Researchers can estimate how large unlikely to be biased. their samples should be for testing There is no guarantee of a hypotheses through power analysis. representative sample, but The risk of “getting it wrong” (i.e., failing random selection guarantees to achieve statistical conclusion validity) that differences between the increases when samples are too small: sample and the population are Researchers risk gathering data that will purely a function of chance. not support their hypotheses even when those hypotheses are correct. 2. Stratified Random Sampling Large samples are no assurance of The population is first divided accuracy, though: with nonprobability into two or more strata, from sampling, even a large sample can which elements are randomly harbor bias. selected. A large sample cannot correct for a The aim of stratified sampling is faulty sampling design; nevertheless, a to enhance representativeness. large nonprobability sample is better than a small one.  D. Critical Appraisal of Sampling Plans Data for quantitative studies tend to be In appraising the description of a quantifiable and structured, with the same sampling plan, you should consider information gathered from all participants in a whether the researcher has adequately comparable, prespecified way. explained the sampling strategy. Quantitative researchers generally strive for Ideally, research reports should describe methods that are as objective as possible. the: type of sampling approach used, population and eligibility criteria for Self-Reports/Patient-Reported Outcomes sample selection, sample size, with a Structured self-report methods are used when rationale, a description of the sample’s researchers know in advance what they need to main characteristics. know and can frame appropriate questions to A key criterion for assessing a sampling obtain the desired information. plan in quantitative research is whether Structured self-report data are collected with a the sample is representative of the formal, written document known as an population. instrument: If the sampling strategy is weak or if the An interview schedule when questions sample size is small, there is reason to are asked orally face-to-face or by suspect some bias. telephone. Research reports ideally should provide A questionnaire when respondents information about response rates (i.e., complete the instrument themselves. the number of people actually participating in a study relative to the A. Question Form and Wording number of people sampled) and about In a totally structured instrument, participants possible nonresponse bias—differences respond to the same questions in the same between participants and those who order. declined to participate (also sometimes Closed-ended questions include prespecified referred to as response bias). response options that may range from a simple yes or no to complex expressions of opinion, DATA COLLECTION IN RESEARCH ensuring comparability of responses and Data collection methods vary along several facilitating analysis. dimensions, including whether the researcher Examples of closed-ended questions include: collects original data or uses existing data. 1. Dichotomous Question: "Have you Existing records serve as an important data ever been pregnant?" (Yes/No) source for nurse researchers, as clinical data 2. Multiple-choice Question: "How gathered for nonresearch purposes can be important is it to you to avoid a analyzed to answer research questions. pregnancy at this time?" (Options range Researchers usually collect new data and must from "Extremely important" to "Not decide on the type of data to gather. Three types important.") frequently used by nurse researchers include: 3. Forced-choice Question: "Which a. Self-reports (Patient-Reported statement most closely represents your Outcome data): Participants’ responses point of view?" (Options on control over to researchers’ questions, commonly life.) collected in nursing studies. 4. Rating Question: "On a scale from 0 to b. Observations: Direct observation of 10, how satisfied were you with the people’s behaviors and characteristics nursing care you received during your for certain questions. hospitalization?" c. Biomarkers: Biophysiological measures Some structured instruments may also include used to assess important clinical open-ended questions, allowing participants to variables. respond in their own words (e.g., "Why did you stop smoking?"). Open-ended questions can provide richer information, but responses to closed-ended questions are easier to analyze.  There is also a risk that researchers may fail to Alternative Self-Report Approaches include important responses in closed-ended Vignettes present brief situational descriptions formats. for participant reactions. Visual analog scales measure subjective Advantages of Questionnaires: experiences on a bipolar continuum. Less costly and advantageous for Q-sorts involve participants sorting statements geographically dispersed samples. along specified dimensions. Internet questionnaires are economical, though response rates tend to be low. Evaluation of Self-Report Methods Offer anonymity, which can be crucial for Self-reports are a direct way to gather sensitive topics. information about how people feel or believe. However, they also have weaknesses regarding Advantages of Interviews: validity and accuracy; researchers must assume Higher response rates in face-to-face interviews; respondents are truthful. respondents are less likely to refuse to talk to an It is essential to be aware of potential biases in interviewer. self-reported data when interpreting research More feasible for populations unable to fill out a findings. questionnaire (e.g., young children). Telephone interviews combine relatively low Observational Methods in Research costs with high response rates for brief For some research questions, direct observation instruments. of people’s behavior is an alternative to self-reports, especially in clinical settings. B. Summated Rating Scales Observational methods can be used to gather Psychosocial scales often incorporated into information such as: self-report instruments assign a numeric score Patients’ conditions (e.g., their to individuals along a continuum, measuring sleep–wake state) attitudes, perceptions, and psychological traits Verbal communication (e.g., exchange such as anxiety or depression. of information at discharge) One common technique is the Likert scale, Nonverbal communication (e.g., body which consists of several declarative statements language) expressing a viewpoint. Respondents indicate Activities (e.g., geriatric patients’ how much they agree or disagree. self-grooming activities) Such scales enable fine discrimination among Environmental conditions (e.g., noise individuals with different opinions and can levels) measure a wide array of attributes. In studies that use observation, researchers have flexibility on several dimensions, such as Common Issues with Scales the focus of the observation (e.g., broadly Scales are susceptible to response set biases: defined events like patient mood swings vs. Social Desirability Bias: Tendency to small, specific behaviors like facial expressions). misrepresent oneself to align with social Observations can be made through the human norms. senses and then recorded manually, but they Extreme Response Bias: Tendency to can also be done with equipment such as video select extreme alternatives consistently. recorders. Acquiescence Bias: Tendency to agree with statements regardless of their Structured Observations content. Researchers often use structured observations Researchers can mitigate biases by developing when participants cannot be asked questions or sensitively worded questions, fostering a cannot be expected to provide reliable answers. nonjudgmental atmosphere, and ensuring Structured observation involves the use of confidentiality. formal instruments and protocols that dictate what to observe, how long to observe it, and how to record the data.  It is not intended to capture a broad slice of life Evaluation of Observational Methods but rather to document specific behaviors, Certain research questions are better suited to actions, and events. observation than to self-reports, particularly when people cannot describe their own Methods of Structured Observation behaviors (e.g., stress-induced behavior, The most common approach to making grieving). structured observations is to use a category Observational methods have an intrinsic appeal system for classifying observed phenomena. for directly capturing behaviors, and nurses may A category system represents a method of be especially sensitive observers. recording in a systematic fashion the behaviors Shortcomings of Observational Methods: and events of interest that transpire within a Potential reactivity (behavioral distortions from setting. being observed) and bias due to the observer’s Some category systems require that all values and prejudices. observed behaviors in a specified domain (e.g., Observational biases probably cannot be body positions) be classified, while others eliminated but can be minimized by training and categorize only particular types of behavior. monitoring observers. Category systems must have careful, explicit definitions of the behaviors and characteristics Biomarkers to be observed, and each category must be Nurse researchers have utilized biomarkers explained. (biophysiological measures) for a diverse range Category systems serve as the basis for of objectives, which include: constructing a checklist—the instrument Conducting studies to explore basic observers use to record observations. biophysiological processes. Investigating how nursing actions and Exhaustive vs. Nonexhaustive Category Systems interventions can affect physiological Observers using an exhaustive category outcomes. system must place all observed behaviors in Performing product assessments to one category for each “unit” of behavior, while evaluate new medical products or nonexhaustive systems allow observers to interventions. watch for instances of listed behaviors that may Assessing the accuracy of or may not occur. biophysiological information that nurses gather during their practice. Rating Scales in Observational Methods Examining the correlates of Another approach to structured observations is physiological functioning in patients who to use a rating scale, which requires observers are experiencing various health to rate phenomena along a continuum. problems. The observer may be required to make ratings In nursing research, both in vivo and in vitro at intervals throughout the observation or to measurements are employed: summarize an entire episode after observation is In vivo measurements are taken completed. directly from living organisms, which can include measurements like blood Observational Sampling Techniques pressure and body temperature. Researchers must decide when, and for how In vitro measures involve collecting long, structured observations will be undertaken. biophysiological material from Time sampling involves selecting time periods participants and analyzing it in a during which observations will occur, while laboratory setting. Examples include: event sampling requires researchers to select Chemical measures such as integral events to observe. hormone levels. Microbiological measures that involve counting and identifying bacteria.  Cytological or histological Measurement properties of primary interest measures, which might involve include reliability and validity. tissue biopsies. There is an increasing interest in the study of Reliability microbiomes, particularly gut microbiomes, Reliability refers broadly to the degree to which among nurse researchers. scores are free from measurement error and the Additionally, anthropomorphic measures, such consistency of those scores across different as body mass index (BMI) and waist instances of measurement. circumference, are frequently used to assess Test–retest reliability is assessed by physical health. administering the same measure to the same Biomarkers are generally considered to be group of individuals at two different points in relatively accurate and precise. They offer time, under the assumption that any observed advantages over psychological measures, such differences in scores are due to measurement as self-reports of anxiety or pain, which can be error rather than actual changes. subjective. Interrater reliability is crucial when multiple Biophysiological measures are regarded as observers are involved in scoring, requiring objective, meaning that two nurses measuring a assessments to ensure consistency among physiological variable, such as using a raters. spirometer, are likely to obtain identical Internal consistency is evaluated by checking readings. Furthermore, patients have limited if multiple items within a measure assess the ability to distort measurements related to their same trait. Higher internal consistency indicates biophysiological functioning. that the items are reliably measuring the same The overarching goal in developing a data construct. collection plan is to yield data of the highest possible quality. Validity A critical component of data quality is the Validity pertains to the degree to which an training of individuals who collect and record instrument measures what it claims to measure. data, ensuring that established procedures are For example, when developing a scale for meticulously followed. resilience, it is crucial that the scores accurately Researchers must also create conditions that reflect that construct rather than related ensure privacy for participants and foster an concepts like self-efficacy or perseverance. atmosphere conducive to open and honest Different aspects of validity include face validity, responses. content validity, criterion validity, and construct The adequacy of the measures used to validity. operationalize constructs is of paramount Cross-cultural validity refers to the effectiveness importance. of translated or culturally adapted measures in Measurement involves the assignment of performing equivalently to the original instrument numbers to reflect the magnitude of a particular across different cultural contexts. attribute present in an individual or object. Various factors can contribute to measurement Critical Appraisal of Data Collection Methods error, including personal states (e.g., fatigue), The ultimate aim of a data collection plan is to biases in how questions are answered generate data of exceptional quality. (response set biases), and situational factors Researchers’ choices regarding their data (e.g., ambient temperature). collection methods can significantly impact the In the case of self-report measures, the wording quality of the resulting data and the overall of questions can significantly impact the findings of the study. reliability of the data collected. However, critically appraising data collection methods in published studies can be Psychometrics challenging, as researchers often provide This field focuses on the theory and insufficient detail in their descriptions. methodologies involved in the measurement of psychosocial phenomena, such as depression, pain, or anxiety.  It is essential for researchers to transparently communicate their data collection methods, enabling readers to evaluate the robustness of the evidence. Reports should include details about the reliability and validity of measures used, with internal consistency coefficients derived from the current study when possible. This information is vital for establishing confidence in the study's findings.