Reading :

Populations, Sampling, and Sampling Strategies

This week we will explore the implementation (how) phase of the research process which includes choosing the population (sample) you are interested in studying, selecting a sampling strategy, determining the data you need to collect (type), and how to collect the data (instruments). When reflecting back about our PICO learning, a population (P) should be identified for investigation. It is seldom realistic, or even possible, to conduct research on theentire population(Houser, 2018; Polit & Beck, 2018).

• Population=an entire set of subjects of interest
• Sample=subset carefully selected characteristics from the population

Instead,samplinginvolves carefully selecting a group of people, events, objects, or other elements as a representative of thepopulation of interest(Houser, 2018); generally the termssubjectsorparticipantswill be used and include the definingsampling eligibilityorinclusion criteria-which is aset/listofkey objective attributes/characteristicsof the subjects to be included in the study sample.

There are many keyinclusion attributes/characteristicsused to define the population of interest (P) orsample/subject characteristics; the below represents a few:

• Age
• Gender
• Diagnosis/condition/issue
• Race
• Level education or years of experience
• Setting-unit/department, primary care clinic

Exclusion characteristicswould be attributes that cause a person to be excluded from the study. It is essential for the authors to include key demographics or descriptors of the sample/subjects.

In addition tosample characteristics, thesettingfor the study is the place where data collection occurs; this may be in a hospital, clinic, school, home, or other locations. The setting may indicate if the research was conducted in a rural or urban setting, a large or small setting, and so forth. As you are reading through articles, review how well the characteristics of sample subjects and the study setting match your own setting ---the closer the resemblance, the more applicable the evidence yielded from the research may be.

The sampling strategy is important because a significantsampling errorcould distort the findings and render them unreliable, invalid, and/or not able to be generalized by the entire population. The sampling strategy considers how subjects arerecruited, selectedand, when appropriate,assigned to groups. Forkey elementsof aneffective sampling strategyreview pp XXXXXXXXXXof our Houser textbook.

The sampling strategy considers how subjects arerecruited,selectedand, when appropriate,assigned to groups. There are two major categories of sampling: (1) probability, also known as random sampling, and (2) nonprobability sampling.

Probability or Random Sampling

Probability in sampling increases the chances that the sample accurately represents the population. Probability sampling is preferred in quantitative research because the goal is to generalize findings from the sample to the population. Probability sampling is not appropriate for qualitative studies because this type of research does not seek to generalize results. Two criteria must be met for achieving a probability sample -random selectionandindependent selection. Read Houser for more details.

Random selection

In probability sampling, ideally, every member of the population has the same chance to be selected as a subject, but sampling the entire population is rarely feasible, and therefore, seldom used.

• Random samplinginvolves choosing subjects at random from the population. This strategy should evenly distribute characteristics found in the population to the sample. These characteristics are called extraneous variables (EV). Extraneous variables are not the focus of the research study, but their presence may influence the dependent variable (DV), or outcome. Although random sampling is easier to achieve than sampling the entire population.
• Random assignmentis much more likely to be the sampling strategy chosen by the researcher. First, the researcher selects subjects who meet certain criteria found in the population. Next, the researcher randomly assigns the subjects to groups to increase the likelihood that any characteristics (EVs) that might affect the outcome (DV) will be equally distributed among the groups. Random assignment to groups is the major way to control for EVs in experimental studies.

Independent selection

The second criterion for achieving a probability sample is independent selection. This means that the selection of one subject is independent, or separate, from the selection of another. If independence is violated, the subjects may share characteristics that will show a correlation due to the sampling strategy instead of the effect of the independent variable (IV). Data results could be distorted.

Nonprobability sampling

Involves a strategy where subjects are not selected at random. Instead, they may be chosen due to convenience or for a particular purpose.

Convenience sample

Theconvenience sampleis themost common strategyin nursing studies. As the name implies, the subjects are chosen by the researcher because they are accessible. The weakness of convenience sampling is that the characteristics of the population Extraneous variables (EVs) are less likely to be distributed evenly among the sample. Bias may occur that will negatively influence the outcomes (DV) of the research. The researcher may decide to use a convenience sample to recruit and select subjects, and then make random assignments to groups to minimize the chance of bias by evening distribution of EVs over all groups.Convenience samplingis a common sampling approach for quantitative studies.

Purposive sample

The purposive sample is one where selection of subjects is intentional, or done on purpose. This is a common strategy in qualitative studies where probability sampling is not needed because there is no need for generalization to a larger population. Subjects are carefully chosen for their ability to inform and enlighten the researcher about a phenomenon or other aspect of the research question under consideration.

Sample Size and Power

Sample sizeis the number of subjects that are studied. Qualitative studies may have rules and criteria for what is an appropriate sample size, but they are not as strict as those for quantitative studies. A case study may have only one subject. Therefore, samples in qualitative studies can be small in size. Typically, the researcher will recruit and collect data on subjects untilsaturationis reached, which is the point at which the researcher determines no new results are emerging. Saturation is one way to build confidence in the results of a qualitative study.

In quantitative research, apower analysisdetermines how large a sample should be to yield statistically significant results. In general, the sample size for a quantitative study must be a minimum of 30 subjects per group. Generally, larger sample sizes increase the likelihood that the researcher will notice subtle differences among groups.

The researcher may report the number of subjects who were recruited, the number that consented to participate, and the number that dropped out during the study.Biasmay be detected if a certain segment of the population was not represented in the final results because they were not recruited, did not consent, or dropped out. When the potential for bias exists, the trustworthiness of the findings may come into question.

• Look for a description of the sampling strategy and how it was applied.
• Does the strategy fit with the research design?
• If this was a quantitative study, were there enough subjects to generalize the findings to the population?
• If this was a qualitative study, was saturation achieved?

These factors and more should be considered when you are selecting research-based evidence to apply to your practice.

Data and Measurement

Measurementis "assigning numbers or some other classification by determining the quantity of a characteristic that is present" (Houser, 2018, p XXXXXXXXXXNumbers are used to collect data. According to Houser, (2018, p. 190), numbers are:

• "objective
• standardized
• consistent
• precise
• statistically testable
• undefined"

The researcher may work with numbers that can be calculated and have numerical value; for example, body temperature. At other times, the researcher assigns numbers to traits simply to classify them.

Data and Levels of Measurement

When thinking about quantitative research, variables must be expressed as numbers in order to use statistics for analysis. Variables are measurable characteristics identified from the elements of the research question (PICOT) that can be measured in a detectable way (Houser, XXXXXXXXXXDifferent types of numbers have different levels of measurement. See the table below for examples of differing types of measurable data and categories of levels of measurement.

(Information adapted from Houser, 2018; Polit and Beck, 2018)

Errors

Errors may occur when collecting data. Ameasurement erroris the difference between the true number and the number that the instrument reads. Read more aboutrandomandsystematicerrors in your textbook. Errors in measurement can ripple throughout the rest of the research process and lead to faulty findings.

Data Collection and Instruments

Reliable, valid, clear, consistent, and unbiased data should be collected by trustworthy, credible, reliable, valid instruments. Data can be collected by primary data collection methods and secondary data collection methods. Refer to Houser for more details.

The computer industry has adopted the acronym GIGO, orgarbage-in garbage-out, to explain that no matter how well information is processed, the quality of the information that comes out can be no better than the quality of the information that goes in (Kilkenny, & Robinson, XXXXXXXXXXIf data are improperly collected, the findings based on that data are worthless, orgarbage. How does the researcher collect reliable, valid, clear, consistent, and unbiased data?

The following areprimary data collectionmethods, where the researcher or data collector directly measures the subjects (Houser, 2018):

• Physiologic measuresinvolve instruments that measure physiologic dimensions, such as blood gases, pulmonary function, and other diagnostic tests.
• Psychometric instrumentsinclude scales or surveys.
• Interviews and questionnairescollect subjective data, such as attitudes, or descriptions of experiences provided by individuals or focus groups.
• Observationobtains objective data related to behavior that the subject may not be able to describe, such as activity during sleep or agitation in the person withdrawing from alcohol.

Secondary datamay be collected from sources that were not created for the current research study. Typically, the researcher looks through records, or "mines," the relevant data that pertains to the variables. Includes data that has already been collected and examples include but are not limited to:

• patient charts;
• census reports;
• public health records.

Keep in mind, the "surveyis the most common method to collect research data" (Houser, 2018, p. 195).

Instruments for Data Collection

Instruments measure variables (captures the data). Instruments must bereliableandvalidin order to yield useful data results/analysis.

Reliableinstruments measure a variable with precision and consistency.

Validinstruments measure in a manner that isaccurateandtruthful. A valid instrument measures the correct measure/data.

An instrument may be reliable but not valid in that it may consistently measure something that is not accurate. Instruments must be reliable in order to be valid and both attributes must exist in order to measure data in a way that inspires confidence in the research findings. Each method of data collection has its own strengths and weaknesses with respect to reliability and validity.

Themethods sectionof articles you read should describe the measurements and the instruments that the researcher used to collect data. The researcher should describe thereliabilityandvalidityof the instrument. The existence of both attributes lends credibility to the claims that the researcher makes in the findings.

Answer the following question about one popular type of data collection instrument.

1. A popular data collection instrument is the Likert scale. What is a Likert scale? What type of number does it measure? Type your answer below.

   Click to view answer

As you are reading the methods section of articles, take notices of the samples/subject attributes. Read the author's description of data and measurements.

• Where the attributes defined?
• Do the data collection methods accurately measure the attributes?
• Consider the reliability and validity of the data gathering tools/instruments.

The accuracy of the results from analysis of the data depend on these considerations. These are issues that the researcher must consider when collecting, analyzing, and interpreting data.

Reading Research Literature

When reading research studies, the details of the implementation are typically found in thesampleandmethodsorprocedures section. The authors should describethe detailsof how the study was carried out in such a way that the reader (you) should be able to actuallyreplicatetheir study. These details should include information about:

• population or sample
• sample characteristics/attributes, selection and techniques
• data:
  • what data measures were collected
  • when was data collected
  • how was data was collected
  • where was data collected
• data collection tools/instruments—validity and reliability

Remember, it is all about thedetails!

Gaining insight into how to read various sections of a research article/study can help you determine whether the research is relevant to your practice, and provides evidence upon which to base your decisions.

• Samples

  Information about the samples is usually located in the methods section; it may also be labeledsamples,subjects,participants. Some information may be in theabstract. Characteristics of the sample can be found in the analysis or results section.

• Data Collection

  A thorough description should be provided in themethodssection; sometimes it may be calledproceduresorprotocols. This should be a major part of the write-up of the study as details are needed for a replication study.

• When reading research, themethods sectionshould describe the measurements and the instruments that the researcher used to collect data. The researcher may describe thereliabilityandvalidityof the instrument. The existence of both attributes lendscredibilityto the claims that the researcher makes about the findings.

Read over Houser'sWhere to Looksection in each chapter to learn how to locate information when reading articles.

Summary

The implementation of the research study is a crucial step of the research process. The way the researcher defines characteristics in the population leads to the selection of instruments to measure and collect data on the subjects (sample) that represent the researcher's population of interest (which would be the (P) in the PICOT). The key terms that you choose to search for, research-based evidence, should try to match the way in which the researcher operationalizes the definitions of the variables.

It is important to remember that numbers can represent various attributes. Some are merely for classification, while others can be quantified and analyzed statistically. Be sure that instruments are reliable and valid, and that they accurately collect data. Next week, we will explore data results and analysis.

Week 5 Discussion: Samples and Data Collection (graded)

Purpose

This week's graded topics relate to the following Course Outcomes (COs).

• CO 2: Apply research principles to the interpretation of the content of published research studies. (PO 4 & 8)
• CO 4: Evaluate published nursing research for credibility and significance related to evidence-based practice. (PO 4 & 8)
• CO 5: Recognize the role of research findings in evidence-based practice. (PO 7 & 8)

Due Date

• Answer post due by Wednesday 11:59 PM MT in Week5
• Two replies to classmates and/or instructor due by Sunday 11:59 PM MT at the end of Week5

Directions

• Discussions are designed to promote dialogue between faculty and students, and students and their peers. In discussions students:
  • Demonstrate understanding of concepts for the week
  • Integrate scholarly resources
  • Engage in meaningful dialogue with classmates
  • Express opinions clearly and logically, in a professional manner
• Use the rubric on this page as you compose your answers.

Access the following information. You may read the PDF online or download it.

American Nurses Association. (2014).Fast facts: The nursing workforce 2014: Growth, salaries, education, demographics & trends. ANA.Retrieved fromhttps://www.nursingworld.org/globalassets/practiceandpolicy/workforce/fastfacts_nsgjobgrowth-salaries_updated XXXXXXXXXXpdf(Links to an external site.)

• Review thedatapresented in theANA Fast Factsand describe some of the key attributes/characteristics of this sample of the nursing workforce.
• Discuss some of the data that you found interesting; include what you believe the purpose (intent) of ANA sharing these results.
• Theinstrumentsandtoolsthat we use to collect data need to bereliableandvalid. Define these terms and explain the importance of each. Shareoneway that can be used to collect data that you were not aware of or familiar with.