Sampling

CHAPTER

McGraw-Hill

2003 The McGraw-Hill Companies, Inc. All rights reserved.

Sampling

2
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

What Is a Sample?
One of the most important steps in the research process is to select the sample of individuals who will participate (be observed or questioned) as a part of the study. Sampling refers to the process of selecting these individuals. A sample in a research study refers to any group on which information is obtained. The larger group to which one hopes to apply the results is called the population. Sample (statistics) Population (parameters) m

s
3
McGraw-Hill

Fraenkel p.96
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Samples and Populations When it is possible, researchers would prefer to study the entire population in which they are interested. Usually, however, this is difficult to do. Most populations of interest are large, diverse, and scattered over a large geographic area. Finding, let alone contacting, all members can be timeconsuming and expensive. For that reason, of necessity, researchers often select a sample to study. A population can be any size and it will have at least one (and sometimes several) characteristic(s) that sets it off from any other population. Notice that a population is always all of the individuals who possess a certain characteristic (or set of characteristics). Fraenkel p. 97
4
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Samples and Populations (contd) In educational research, the population of interest is usually a group of persons (students, teachers, or other individuals) who possess certain characteristics. In some cases, however, the population may be defined as a group of classrooms, schools, or even facilities.

Target population the actual population to which a researcher would like to generalize rarely available ideal choice
Accessible population the population to which a researcher is able to generalize realistic choice Fraenkel p. 97
5
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Random versus Nonrandom Sampling

Random sampling its intent is to select a representative sample from the population

Fraenkel p. 98
6
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Representative vs. Non-representative Samples

7
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Random Sampling Methods

Simple Random Sample a sample in which each and every member of the population has an equal and independent chance of being selected
Fraenkel p. 99

8
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Part of a Table of Random Numbers (Table 6.1)

011723 912334 086401 059397 666278 051965 063045 560132 727009 000037 667899 042397 987650 091126
9
McGraw-Hill

223456 379156 016265 022334 106590 004571 786326 345678 344870 121191 234345 045645 568799 021557

222167 233989 411148 080675 879809 036900 098000 356789 889567 258700 076567 030032 070070 102322

032762 109238 251287 454555 899030 037700 510379 033460 324588 088909 090076 657112 143188 209312

062281 934128 602345 011563 909876 500098 024358 050521 400567 015460 345121 675897 198789 909036

565451 987678 659080 237873 198905 046660 145678 342021 989657 223350 121348 079326 097451 342045

2003 The McGraw-Hill Companies, Inc. All rights reserved.

Random Sampling Methods

Stratified Random Sample a process in which certain subgroups, or strata, are selected for the sample in the same proportion as they exist in the population This requires the researcher to know some information about the people in the population
Fraenkel p. 100
10
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Selecting a Stratified Sample (Figure 6.2)

11
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Random Sampling Methods (contd)

Cluster Random Sampling the selection of groups, or clusters, of subjects rather than individuals Two-Stage Random Sampling a combination of cluster random sampling with individual random sampling
Fraenkel p. 100-101

12
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Random Sampling Methods (Figure 6.3)

13
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Nonrandom Sampling Methods Systematic Sampling every nth individual in the population list is selected for inclusion in the sample

If the initial individual is selected randomly then this method is technically known as systematic sampling with a random start.
The sampling interval is the distance in the list between each of the individuals selected. It is determined by dividing the population size by the desired sample size. The sampling ratio is the proportion of the population that is selected for the sample. It is determined by dividing the sample size by the population size. Fraenkel p. 102-103
14
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Nonrandom Sampling Methods (contd) Convenience Sampling a group of individuals who (conveniently) are available for study. In general, these samples cannot be considered representative of any population and should be avoided if at all possible. If this is the only type of sample the researcher can use, then they should be extremely careful to include demographic information and other characteristics of the sample that was studied. The study should also be replicated with a number of similar samples to decrease the likelihood that the results obtained were simply a one-time occurrence. Fraenkel p. 103
15
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Convenience Sampling (Figure 6.4)

16
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Nonrandom Sampling Methods (contd)

Purposive Sampling a sampling method in which the researcher selects the sample based on previous knowledge of a population. This is different from convenience sampling in that researchers do not simply study whoever is available, but use their judgment to select a sample that they believe, based on prior information, will provide the data they need.

17
McGraw-Hill

Fraenkel p. 103-105
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Nonrandom Sampling Method (Figure 6.5)

18
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Sample Size
A sample should be as large as the researcher can obtain with a reasonable expenditure of time and energy.

Type of Study Descriptive Experimental or Causal/Comparative Correlational

Minimum Sample Size 100 30 in each group 50

Fraenkel p.106-109
19
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

External Validity: Generalizing from a Sample Both the nature of the sample and the environmental conditions the setting - within which a study takes place must be considered in thinking about generalizability. The extent to which the results of a study can be generalized determines the external validity of the study.

20

Population generalizibility refers to the degree to which a sample represents the population of interest. If the results of a study only apply to the group being studied, and if that group is fairly small or is narrowly defined, the usefulness of any findings is seriously limited. This is why trying to find a representative sample is so important. Fraenkel p. 109
2003 The McGraw-Hill Companies, Inc. All rights reserved.

McGraw-Hill

External Validity: Generalizing from a Sample (contd) One aspect of generalizability that is often overlooked in methods or treatment studies is that which pertains to the teachers, counselors, administrators, or others who administer the various treatments. A study that randomly selects students but not teachers is only entitled to generalize the outcomes to the population of students - if taught by the same teachers. To generalize to other teachers, the sample of teachers must also be selected randomly and must be sufficiently large.

The best sampling plan is of no value if information is missing on a sizable portion of the initial sample.
21
McGraw-Hill

Fraenkel p. 110
2003 The McGraw-Hill Companies, Inc. All rights reserved.

External Validity: Generalizing from a Sample (contd) Do researchers always want to generalize? No, not if the results of an investigation are of interest only as applied to a particular group of people at a particular time, and where all of the members of the group are included in the study. Random sampling is sometimes not feasible which makes much of the research in education subject to criticism. Replication of a study can give a researcher more confidence about generalizing their findings despite the lack of a random sample. Fraenkel p. 109
22
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Ecological Generalizability Ecological generalizability refers to the degree to which results of a study can be extended to other settings or conditions. Researchers must make clear the nature of the environmental conditions the setting - under which a study takes place. Only when outcomes have been shown to be similar through replication across specific environmental conditions can we generalize across these conditions.

Fraenkel p. 111
23
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.

Population as Opposed to Ecological Generalizing (Figure 6.6)

24
McGraw-Hill
2003 The McGraw-Hill Companies, Inc. All rights reserved.