Sampling Flashcards

Question

Quota sampling

Answer 1

Quota Sampling Research/study excerpt In April 2019, an IRB [institutional review board] approved cross-sectional study was conducted with a convenience sample of college students attending on-campus meetings. Quota sampling was used to ensure roughly equal numbers of ENDS users/non- users and males/females, with modest compensa- tion provided to all participants. (Ashford et al., 2020, Methods section) Critical appraisal Ashford et al. (2020) used convenience sampling to address their study purpose. The University of Kentucky college campus provided easy access to an adequate number of study participants. Quota sampling was an effective way to ensure that equal numbers of ENDS users and nonusers and females and males were included in the sample, which in-creased its representativeness (see Table 9.2). The researchers reported that all ENDS users identified JUUL as their primary product. In addition, the ENDS users were younger, used cigarettes and ma- rijuana, reported a persistent cough, and had al- tered salivary biomarkers.

Answer 2

Effect size Factors that influence the adequacy of sample size (because they affect power) include effect size, type of quantitative study, sensitivity of the measure- ment methods, and data analysis techniques. The effect is the presence of the phenomenon examined in a study. The effect size is the extent to which the null or statistical hypothesis is false or, stated an- other way, the strength of the expected relationship between two variables or differences between two groups. In a study in which the researchers are com- paring two populations, the null hypothesis states that the difference between the two populations is zero. However, if the null hypothesis is false, an iden- tifiable effect is present—a difference between the two groups does exist. If the null hypothesis is false, it is false to some degree; this is the effect size (Aber- son, 2019). The statistical test tells you whether there is a significant difference between groups, or whether variables are significantly related. The effect size tells you the size of the difference between gthe groups or the strength of the relationship be- tween two variables (Grove & Cipher, 2020). When the effect size is large (e.g., considerable difference between groups or a strong relationship between two variables), detecting it is easier and can be done with a smaller sample. When the effect size is small (e.g., only a small difference between groups or a weak relationship between two vari- ables), detecting it is more difficult and requires a larger sample. There are different types of effect size measures, and each corresponds to the type of stat- istic computed (see Chapter 11). Often the effect size is smaller with a small sample, so effects are more difficult to detect. Increasing the sample size also increases the effect size, making it more likely that the effect will be detected, and the study find- ings will be significant. When critically appraising a study, determine whether the study sample size was adequate by noting whether a power analysis was conducted and what power was achieved. Also, did the researchers calculate the power level of study re- sults that were not significant

Answer 3

Sample size for different types of quantitative studies Descriptive (particularly those using survey ques- tionnaires), correlational, and outcomes studies often require large samples with more than 100 par- ticipants. In these studies, researchers may exam- ine multiple variables, and extraneous variables are likely to affect participant responses to the variables studied. Researchers often make statistical compari- sons on multiple subgroups in a sample, such as groups formed by gender, age, or severity of ill- ness, requiring that an adequate sample be available for each subgroup being analyzed (Grove & Cipher, 2020). When variables are included in the data ana- lyses to answer the research questions or test the hypotheses, the sample size must be increased to detect differences between groups or relationships among variables. However, analyzing demographic variables to describe the sample does not require an increase in sample size. Quasi-experimental and ex- perimental studies often have smaller samples than descriptive and correlational studies because fewer variables are studied and a controlled intervention is implemented in a structured setting (Kazdin, 2017).

Answer 4

Data analysis techniques Data analysis techniques vary in their capability to detect differences in the data. Statisticians refer to this as the "power of the statistical analysis." An interaction also occurs between the measurement sensitivity and power of the data analysis technique (Leedy & Ormod, 2019). The power of the analysis technique increases as precision in measurement increases. Because of this, techniques for analyzing variables measured at interval and ratio levels are more powerful in detecting relationships and differ- ences than those used to analyze variables measured at nominal and ordinal levels (see Chapter 10 for de- tails on levels of measurement). Larger samples are needed when the power of the planned statistical analysis is weak (Grove & Cipher, 2020). When groups are being compared related to study variables, you may use statistical procedures, such as the t test and analysis of variance. Equal group sizes maximize the effect size, which improves statistical power. The more unbalanced the group sizes, the smaller the effect size, which means a larger sample is needed to detect significant differences (Kraemer & Blasey, 2016). The chi-square test is the weakest of the statistical tests and requires large sample sizes to achieve acceptable levels of power

Answer 5

Power analysis is a planning tool used to determine the sample size needed to achieve adequate statistical power, while statistical power refers to the ability of a statistical test to detect a true effect if it exists.

Answer 6

UNDERSTANDING NURSING RE... CRITICAL APPRAISAL GUIDE- LINES Adequacy of the Sample Size and Sampling Method in Quantitative and Outcomes Stud- ies The initial critical appraisal guidelines for the sam- pling process in quantitative and outcomes studies were introduced earlier. This section focuses on questions about the sampling method and the ad- equacy of sample size, which influence the repre- sentativeness of the sample. 1. Is the sampling method probability or nonprob- ability? Is the specific sampling method(s) used in a study identified and appropriate (see Tables 9.1 and 9.2)? 2. Is the sample size identified? Is a power ana- lysis conducted and accurately reported (Grove & Cipher, 2020; Tam et al., 2020)? 3. Was the sample size adequate, as indicated by the power analysis (Aberson, 2019)? 4. If groups were included in the study, is the sam- ple size for each group approximately equal and adequate (Grove & Cipher, 2020)? 5. Is the sample representative of the accessible and target populations?

Answer 7

Sampling methods in qualitative and mixed methods research Qualitative research is conducted to gain insights, discover meaning, and promote understanding about a particular phenomenon, situation, or cul- tural element (Bryant & Charmaz, 2019; Creswell & Poth, 2018). The intent is an in-depth understand- ing of a situation or problem from the perspective of selected study participants. The sampling in quali-tative and parts of mixed method research focuses on obtaining in-depth data to address the study purpose. In ethnographic studies, researchers often select the setting and site and then the population and topic of interest (Marshall & Rossman, 2016). In a phenomenological study, researchers often se- lect the phenomenon of interest and then iden- tify potential participants. The participants selected need to have experience and be knowledgeable in the area of study and willing to share rich, in- depth information about the phenomenon or situ- ation being studied. For example, if the goal of the study is to describe the phenomenon of living with chronic pain, the researcher will select individuals who are articulate and reflective, have a history of chronic pain, and are willing to share their chronic pain experience (Creswell & Poth, 2018). Common sampling methods used in qualitative and mixed methods studies are purposeful, network, theoret- ical, and convenience (covered earlier) sampling (see Table 9.2). Researchers need to describe the sampling methods used in their study in enough detail to pro- mote the readers' confidence in the findings.

Answer 8

Network sampling Network sampling, sometimes referred to as snow- ball or chain sampling, involves finding a few par- ticipants who meet the sampling criteria and asking them for their assistance in finding others with simi- lar characteristics. Network sampling holds prom- ise for locating participants who would be difficult or impossible to obtain in other ways or who had not been previously identified for study (Creswell & Poth, 2018; Marshall & Rossman, 2016). Network sampling takes advantage of social networks and the fact that friends tend to have characteristics in common. This strategy is also particularly useful for finding participants in socially devalued popula- tions, such as persons who are dependent on alcohol, addicted to drugs, abuse children, or commit sexual offenses or criminal acts. These persons rarely make themselves known for study. Other groups, such as widows, grieving siblings, or persons successful at lifestyle changes, also may be located using network sampling. They are typic- ally outside the existing healthcare system and are otherwise difficult to find. Network sampling is an effective strategy for identifying participants who can provide great insight and essential information about a phenomenon or situation being studied. Researchers often obtain the first few study par- ticipants through a purposeful or convenience sam- gpling method and expand the sample size using network sampling. In qualitative research, sampling continues until saturation occurs. Saturation occurs when newly collected data begin to be like previously collected data and yield no new insights or infor- mation. With saturation, concepts are understand- able and well described, details of a phenomenon are available, or patterns or themes of a theory emerge (Bryant & Charmaz, 2019). As discussed earlier in Research Example 9.9, Mulholland et al. (2020) used purposeful and snow- ball sampling methods to identify military families with children who experienced parental deploy- ment. Research Example 9.10 provides addition in- formation about the snowball sampling and sample size in this study.

Answer 9

Network Sampling Research/study excerpt A link to the online survey was posted to the re- searcher's personal Facebook page in which viewers could share the post or tag others to view the post, thus initiating the snowball effect... These indi- viduals were purposively selected based on known military experience that made them potential can- didates... or they had known connections to others that would qualify for the study. Sampling was completed once data saturation was achieved with 15 participants. This was determined by the re- searchers when the data revealed to be consistent gwithout new or emerging themes. The sample size was appropriate due to the qualitative nature of the study. (Mulholland et al., 2020, p. 36) Critical appraisal Mulholland et al. (2020) alternated the use of snow- ball and purposeful sampling methods to identify quality participants to address the study purpose. Facebook was an effective way to initiate snow- ball sampling with the tagging of other military families. Multiple types of social media were used to identify participants who had connections with other military families who might qualify for the study. Data saturation was achieved indicating an adequate sample for this study

Answer 10

Theoretical sampling is mainly implemented in grounded theory research. This study method is de- signed and implemented to develop a model, frame- work, or theory. The researcher gathers data from any person or group who is able to provide rele- vant, varied, and rich information for model or theory generation. The data are considered relevant and rich if they include information that generates, clarifies, and saturates the theoretical codes identi- fied through analysis. The theoretical codes form the basis for framework or theory generation (Bryant & Charmaz, 2019). A code is saturated if it is com- plete and the researcher can see how it fits in the emerging model or theory. When a code or concept is unclear, the researcher continues to seek partici- pants and gather data, especially participants with different characteristics. The process continues until the codes are saturated, and the theory evolves from the codes and data. Diversity in the sample is en- couraged so that the theory developed covers a wide range of behaviors in varied situations and settings. Crooks and colleagues (2020, Abstract section) conducted a grounded theory study to investigate "the sociocultural conditions and processes of be- coming a sexual Black woman in order to under- stand the sociocultural drivers of STI/HIV [sexually transmitted infections/human immunodeficiency virus] rates among this group." The sample included 20 Black women aged 19 to 62 years in a Midwest- ern community. The sampling methods used to re- cruit study participants are described in Research Example 9.11.

Answer 11

Theoretical Sampling Research/study excerpt Purposive and theoretical sampling techniques were used for recruitment to provide conceptual clarity and further exploration of data... We ini- tially recruited a purposive sample of women that had STI, however, after six interviews we made a theoretical sampling decision to include women without STI to allow for more variation in partici- pant experience. (Crooks et al., 2020, Data collec- tion section) Critical appraisal Crooks et al. (2020) detailed their sampling methods. Using purposive sampling, they were able to identify women with varied histories and experiences. Initially the study included Black women with STIs. The researchers realized they needed to include women without STIs to expand their understanding of the relationships of these women. Theoretical sampling continued until data saturation was reached after 20 interviews. Crooks et al. (2020) found that "Black men, silencing Black girls and women, cultural norms and messaging about sexuality, and gendered societal expectations and sexual stereotypes contribute to STI/HIV risk in Black girls and women... Our findings demon- strate how the intersection of social and systemic structures (i.e., history, incarceration, unemploy- ment) shape the context of Black heterosexual rela- tionships." (Crooks et al., 2020, Abstract section)

Sampling Flashcards

(35 cards)