Discussion: Planning for Data Collection

Tips of Learning: Week 8
Polit and Beck (2012), Chapter 12
The aggregate of cases in which a researcher is interested is called a population. A sample is selection of a portion of the population to represent the entire population.
Population:  The aggregate of cases in which a researcher is interested
Sampling:  Selection of a portion of the population (a sample) to represent the entire population 
Element: Basic population unit about which information is collected
Establish population characteristics 
Determine participation in study
o	Maximize construct validity
o	Inclusion
o	Exclusion
Costs, practical constraints, people’ ability to participate, and design considerations
Researchers usually sample from the accessible population but should identify the target population to which they want to generalize their results.
Representative sample:  A sample whose key characteristics closely approximate those of the population
Sampling bias:  The systematic over- or under-representation of segments of the population on key variables
Probability sampling:  Involves random selection of elements
o	Simple random sampling
Example – a sample of 250 members randomly selected from a roster of ANA members
o	Stratified random sampling
o	Cluster sampling
o	Multistage sampling
Example – 100 inmates randomly sampled from a random selection of five federal penitentiaries
o	Systematic sampling
Example – every 20th patient admitted to the ER between January and June
Nonprobability sampling:  Does not involve selection of elements at random
o	Convenience sampling 
Example – all the oncology nurses participating in a continuing education seminar
o	Snowball sampling 
o	Quota sampling
Example – the first 20 male and first 30 female patients admitted to the hospital with hypothermia
o	Consecutive sampling
o	Purposive sampling
Example – 25 critical care experts
Power Analysis	
o	The probability that a statistical test will detect a significant difference that exists - the risk of a Type I error can be calculated using power analysis.
o	Level of significance (usually p=0.05)
o	Sample size
o	Power - acceptable level is .80
o	Effect size - the degree to which the null hypothesis is false

Polit and Beck (2012), Chapter 13
For structured data, not unstructured data, researchers use formal data collection instruments that place constraints on those collecting data and those providing them.
Identification prioritization of data needs – what kind of data do you need?
When selecting existing instruments consider:
o	Conceptual stability
o	Data quality
o	Cost 
o	Population appropriateness
o	Reputation
Open-ended questions permit respondents to reply in narrative fashion, whereas closed-ended (or fixed-alternative) questions offer response alternatives from which respondents must choose.
Increase in web based surveys
Structured self-reported instruments
o	Interview schedules
Higher response rate
Wider variety of people
Richer data
Data quality:
Interpersonal skills
Ease and build rapport
o	Questionnaires
Less costly
Less time-consuming
No risk of interviewer bias
Group administration:
Most economical
Mail: low response rate
Web-based survey
o	Open-ended questions – people respond in their own words (example- what is it like to be a cancer survivor?)
o	Closed-ended question types:
Multiple choice
Rank order
Forced choice
Visual analog scales
Forced-choice questions require respondents to choose between two competing positions, rating questions ask respondents to make judgments along a bipolar dimension, checklists have several questions with the same response format, visual analog scales (VASs), which are continua used to measure subjective experiences.
If including both positive and negative items on a scale, you would need to reverse score one type or the other before summing the items
Likert type scales (example: 0- strongly disagree, 1 – disagree, 2- neutral, 3- agree, and 4- strongly agree)
o	Summated rating scale
o	Series of statements about a phenomenon
o	Indicate degree of agreement or disagreement
o	Total score is computed by summing item scores, each of which is scored for the intensity and direction of favorability
Q sort
o	Set of card statements into piles
o	Usually 50-100 cards into 9-11 piles
o	Specified criteria
o	Can be used to study Attitudes, Personality, Psychological traits
o	Sorting statements into different piles along a continuum
o	Brief descriptions of event
o	Asked to react to events
o	Assess respondents: Perceptions, Hypothetical behaviors, Decisions
Sampling for structured observations
o	Time sampling involves the specification of the duration and frequency of observational periods and intersession intervals. 
o	Event sampling selects integral behaviors or events of a special type for observation.
Data collection protocols – step by step procedures
o	When collecting questionnaires via the mail, send follow-up reminders
Biophysiologic Measures
o	Tie into relevance for nursing, consequences of nursing interventions; assess clinical procedures; study the correlation between physiologic functioning in patients 
o	In vivo measurement – directly in or on living organisms; vitro – outside of the body
Train your data collectors!
Polit and Beck (2012), Chapter 21 (22 in new text)
Qualitative samples tend to be small, nonrandom, and intensively studied. 
o	No explicit, formal criteria
o	Sample size determined by informational needs
o	Decisions to stop sampling guided by data saturation
o	Data quality can affect sample size
Methods of sampling in qualitative research
o	Convenience (volunteer) sampling
o	Snowball sampling
Ask participants to refer other study participants
o	Theoretical sampling
Theory-based sampling selecting cases on the basis of their representation of important constructs and opportunistic sampling, adding new cases based on changes in research circumstances or in response to new leads that develop in the field.
Need to mention saturation…and then stop enrolling subjects
o	Purposive sampling
Select cases that will most benefit the study
Sampling for representativeness or comparative value
Sampling special or unique cases
Sampling sequentially
Homogeneous sampling (deliberately reducing variation or diversity), typical case sampling (selecting cases that illustrate what is typical), extreme case sampling (selecting the most unusual or extreme cases), intensity sampling (selecting cases that are intense but not extreme).
o	Maximum variation sampling
o	Homogeneous sampling
o	Typical case sampling
o	Extreme case sampling
o	Intensity sampling
o	Stratified purposeful sampling
o	Reputational case sampling
o	Mingling with many members of the culture
o	Informal conversations with 25 to 50 informants
o	Multiple interviews with smaller number of key informants
o	Relies on very small samples (often 10 or fewer)
o	Participants must have experienced phenomenon of interest
Grounded theory:
o	Typically involves samples of 20 to 30 people
o	Selection of participants who can best contribute to emerging theory (usually theoretical sampling)
o	Researchers find findings: Relevant and Meaningful
o	Three models
Extrapolating from a sample to a population
Analytic generalization: researchers strive to generalize from particulars to a broader conceptualizations and theories. 
case to care translation) Transferability involves judgments about whether findings from an inquiry can be extrapolated to a different setting or group of people.
Transferability has close connections to the proximal similarity model that involves a conceptualization about which contexts are more or less like the one in the study in terms of a gradient of similarity for people, settings, times, and contexts. 
Polit and Beck (2012), Chapter 22 (23 in new text)
Qualitative studies typically adopt flexible data collection plans that evolve as the study progresses. Self-reports are the most frequently used type of data in qualitative studies, followed by observation.
Fieldwork Issues
o	Gaining participants’ trust
o	Pacing data collection to avoid being overwhelmed by the intensity of data
o	Avoiding emotional involvement with participants
o	Maintaining reflexivity (researcher’s awareness of themselves as part of the data they are collecting)
Unstructured and loosely structured self-reports
o	Latitude in questions and answers
Broad question: the grand tour question
o	Yield rich narrative data
Types of self-reports
•	Unstructured interviews
•	Semistructured interviews
•	Focus group interviews
•	Joint interviews
•	Life histories
•	Oral histories
•	Critical incidents interviews
o	Use of audio taping over time – decisions about problems
•	Diaries and journals
•	Think aloud method
•	Photo elicitation interviews
•	Self-report narratives on the Internet
Qualitative researchers sometimes collect unstructured observational data, often through participant observation. Participant observers obtain information about the dynamics of social groups or cultures within members’ own frame of reference.
Preparing for the interview
o	Ensure that interviewers and respondents have a common vocabulary.
o	Develop and word questions to be asked; become familiar with questions.
o	Conduct a practice interview.
o	Decide how to present oneself to participants. 
o	Decide on settings for data collection.
o	Obtain needed supplies.
Conducting the interview
o	Take steps to “break the ice” to put participants at ease.
o	Share pertinent information about the study with participants.
o	Develop rapport to gain participants’ trust.
o	Listen intently to guide subsequent questioning.
o	Probe for information as needed.
o	Be ready to handle emotionality.
o	Be prepared to manage crises (e.g., interruptions, equipment problems).
o	Achieve a positive closure.
o	Observations tend to become more focused over time, ranging from descriptive observation (broad observations) to focused observation of more carefully selected events
o	Recording observations:
	Log (field diary)
	Field notes:
	Descriptive (observational) notes
	Reflective notes:
	Methodologic notes
	Theoretical notes (or analytical notes)
	Personal notes
o	Types of observations:
	Descriptive observation
	Focused observation
	Selective observation
Participant observers usually select events to be observed through a combination of single positioning (observing from a fixed location), multiple positioning (moving around the site to observe in different locations), and mobile positioning (following a person around a site).
Enhancing credibility
o	Spent time with participants
o	Member check – participants review the findings
o	World-wide reputation as a researcher
o	Take comprehensive field notes
Keough & Tanabe (2011) – survey research
Research to gather information from a large population
Self-report (low ranking in terms of level of evidence)
Assume participants answer honestly
o	What to look good – social desirabilty
Research question drives expected outcomes and project design; takes into consideration published information on the topic and should address significance to nursing
o	Face-to-face interviews
o	US mail
o	Web-based
o	Email
o	Windshield
After decide on the design…consider the method of data collection
Sample survey – specific population – define the target population
If mail, tailored design method
o	Friendly questionnaire
o	4 contacts with first class stamps with special contact
o	Return envelopes with first class stamps
o	Personal correspondence
o	Token incentive
Decide: create new survey or use existing one (consider validity and reliability)
Consider piloting the survey first
Have a clear plan for data management – clean your data before analysis
Consider costs: web based may cost $4200 whereas mailing may cost $6200
Nurses and other health care professionals are often interested in assessing patient satisfaction with health care services. Imagine that you are a nurse working in a suburban primary care setting that serves 10,000 patients annually. Your organization is very interested in understanding the patient’s point of view to help determine areas of care that can be improved. With this focus in mind, consider how you would create a survey to assess patient satisfaction with the services your organization provides. You may wish to consider variables such as the ease of accessing care, patient wait time, friendliness of the staff, or the likelihood that a patient would recommend your organization to others.
o	Write 5 questions about patient satisfaction
o	How would you collect this data – what methods or instruments would you use (Chapters 13 and 23)
o	What would be your sample size – who is your target population, how would you recruit subjects
o	Provide rationale for your choices and remember to talk about the validity and reliability of the questions and type of method or instrument used – highlight this with suggestions!!!

Get 20% off your first purchase