Sampling designs and techniques: types, applications, and limitations

Sampling Designs and Evaluability Assessments

Identify and discuss 2 types sampling designs, their benefits and limitations

In essence, non-probability sampling and probability sampling are the only two types of sampling designs.

Non-probability sampling:

A researcher selects a small proportion of the units to represent the whole. The selection is deliberate without any defined procedure. For example, if the objective of the researcher is to study the economic conditions of a particular population residing in a defined area. The researcher deliberately chooses a few towns or cities to represent the entire area under study (Levy & Lemeshow, 2008). Based on the objectives and specificity of the study, the following are some of the sampling techniques under the non-probability sampling.

Purposive or Judgmental Sample

In purposive or judgmental sampling, the researcher selects the sample based on the familiarity of the population and the objectives of the study. The researcher prefers using a purposive sample because the sample under interview fits a specific objective or description (Levy & Lemeshow, 2008).

a) Quota Sample

In quota sampling, the researcher selects units into a sample based on pre-specified characteristics. This ascertains that the total sample selected shares similar distribution of characteristics with the whole population under study (Thompson, 2002). For instance, a researcher conducting a national quota sample might want to know the proportions that represent the male and female. As well, the researcher might want to know the proportions that represent, gender, age, race or ethnicity and so forth. The researcher then collects a sample with similar proportions as the national population (Schaeffer, Mendenhall & Ott, 2006).

B. Probability sampling:

Probability sampling also share others names, such as random sampling or chance sampling. In the sampling procedure, the researcher gives each unit in the population an equal probability of selection for the sample. Every selection is independent of each other. Probability measures the accuracy of the obtained results (Thompson, 2002). This means that the researcher is capable of measuring the errors of estimation from a random sample. This fact makes probability-sampling design preferred than probability sampling design.

Probability sampling maintains the principle of Statistical Regularity, which confirms that if the selection of the sample is random, the sample bears the components and features as the whole. Therefore, probability sampling remains the best methodology to choose a representative sample (Schaeffer, Mendenhall & Ott, 2006).

b) Simple Random Sample

This is an essential sampling technique applied in statistical and computational methods. In the process of collecting a simple random sample, each proportion of the sample has a number. The researcher generates a set of random numbers and the units assigned those numbers are part of the sample. For example, suppose the researcher selects a population with 200 and decides to use a random sample of 40 people. he/she assigns them numbers 1 to 200, and then generates a list of 40 random numbers (Thompson, 2002). The persons assigned those numbers are the ones used in the sample.

c) Systematic Sample

In a systematic sample, the researcher lists down the elements of the population, and then selects every nth element to include in the sample. For example, in a population, a study that consists of 50 units, and the researcher wants 25 units, the researcher will list down the units and then choose every fifth unit to include in the sample. In avoiding any business, the researcher must choose the first unit at random (Thompson, 2002).

Benefits of probability sampling

I. The sampling requires simple information as opposed to non-probability sampling.

II. The selection mode is dependent and any unit selected has a probability of being used in the research study.

Limitations of probability sampling

I. It requires a sampling frame of elements, and it may be difficult to obtain suitable sampling frame. It is impractical to construct one that is appropriate.

II. Contain larger sampling errors that may obstruct accurate results.

Benefit of non-probability sampling

I. The statistical procedures used in analyzing and computing data are easier

II. The procedure does not require any advanced auxiliary information in computing data

Limitations of non-probability sampling

I. The results obtained may contain larger errors because a small amount of sample represents the whole

II. In a widely dispersed population, the costs incurred in data collection are higher than in probability sampling.

Part II: Secondary data analysis on Department of Justice Data relevant to traffic stops and racial profiling:

Describe the feasibility study questions that would be relevant for a racial profiling evaluation utilizing secondary data.

What questions would you ask her relevant to the nature of the data in order for you to determine if the study is feasible; consequently, allowing you to choose an appropriate research methodology.

The major problem in conducting research pertaining to racial profiling in traffic stops is approximating the risk set against which to compare the race distribution of stopped drivers. Recently, researchers prefer using residential population data or conduct traffic studies in which they compute the race distribution of drivers at certain locations. However, resident population data give inaccurate estimates of the population at risk of a traffic stop, while traffic surveys are expensive to conduct. Most research studies on racial profiling seek to establish whether the patrol officers are likely to prevent minority drivers than white drivers from the at-risk population (Ridgeway, 2006).

In estimating the at-risk population, numerous studies use the secondary data. Some of them use the census-based estimates of the racial distribution populations. This approach contains serious limitations recognized by researchers and the law courts. Some of the issues addressed with evaluability assessments are the difference in car ownership and the travel patterns. For instance, if there are any racial differences in the driving behavior, then the racial distribution of the population at-risk may vary from the racial distribution of the driving population.

The U.S. Supreme Court upholds the legality of the traffic stops pursuant to trivial violations of the law. Another evaluability assessment addressed that determines the validity of the feasibility study is the exposure of the police in controlling traffic stops and racial profiling (Ridgeway, 2006).

In addition, the following questions are relevant to the nature of the data collected that would help determine if the study is feasible. Moreover, they will allow in choosing a suitable researcher's methodology.

(1) I will ask the traffic officer how they choose the drivers they choose for traffic violations.

(2) Do the traffic officers have an official police about profiling?

(3) Is there any training that police gets about racial profiling in their academy?

(4) Is the training in progress for the veteran officers?

(5) Are there any circumstances that traffic officers underscored the relevance of profiling?

(6) What times does profiling worked? And when it failed.

(7) Are there unsolved or lawsuits waiting in your local court system about racial profiling? Do traffic officers use windshield-mounted cameras? Are the tapes produced in evidence of lawsuits?