Matched Pairs Design

8 Key excerpts on "Matched Pairs Design"

• 

  eBook - PDF

  Real World Research

  • Colin Robson(Author)
  • 2024(Publication Date)
  • Wiley

    (Publisher)

  Can be post-test only or pre-test post-test. 5. Matched Pairs Designs. Establishing pairs of participants with similar scores on a variable known to be related to the dependent variable (DV) of the experiment. Random alloca- tion of members of pairs to different experimental groups (or to an experimental and control group). This approach can be used in several two-group designs. Attractive, but can introduce complexities both in set-up and in interpretation. 6. Repeated measures designs. The same participant tested under two or more experimental treatments or conditions (or in both an experimental and a control condition). Can be thought of as the extreme example of a Matched Pairs Design. 1 6 0 R E A L W O R L D R E S E A R C H be possible to control for age as a variable by, say, working only with people aged between 25 and 30 years. However, creating matched age pairs allows you to carry out a relatively sensitive test without the conclusions being restricted to a particular – and narrow – age range. Designs involving repeated measures The ultimate in matching is achieved when an individual’s performance is compared under two or more conditions. Designs with this feature are known as repeated measures designs. We have come across this already in one sense in the ‘before and after’ design – although the emphasis there is not on the before and after scores per se, but on the relative difference between them in the treatment and comparison groups as a measure of the treatment effect. The website discusses some methodological problems with designs using matching or repeated measures. Choosing among true experimental designs Box 6.6 gives suggestions for the conditions under which particular experimental designs might be used when working outside the laboratory. Cook and Campbell (1979) have discussed some of the real-world situations that are conducive to carrying out randomized experiments.

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Medical Statistics from A to Z

  A Guide for Clinicians and Medical Students

  • Brian S. Everitt(Author)
  • 2021(Publication Date)
  • Cambridge University Press

    (Publisher)

  P Paired availability design: A design that can reduce selection bias in situations where it is not possible to use random allocation of subjects to treatments. The design has three fundamental characteristics:  The intervention is the availability of treatment, not its receipt.  The population from which subjects arise is well defined with little in- or out- migration.  The study involves many pairs of control and experimental groups. In the experimental groups, the new treatment is made available to all subjects, although some may not receive it. In the control groups, the experimental treatment is generally not available to subjects, although some may receive it in special circumstances. [Statistics in Medicine, 1994, 13, 2269–78.] Paired Bernoulli data: Data arising when an investigator records whether a particular characteristic is present or absent at two sites on the same individual, for example the presence or absence of spots on the legs and arms. [Biometrics, 1988, 44, 253–7.] Paired samples: Two samples of observations with the characteristic feature that each observation in one sample has one and only one matching observation in the other sample. There are several ways in which such samples can arise in medical investigations. The first, self-pairing, occurs when each subject serves as his or her own control, as in, for example, therapeutic trials in which each subject receives both treatments, one on each of two separate occasions. Next, natural pairing can arise, particularly, for example, in laboratory experiments involving littermate controls. Lastly, artificial pairing may be used by an investigator to match the two subjects in a pair on important characteristics likely to be related to the response variable. See also matched pairs t-test. [Agresti, A., 2019, An Introduction to Categorical Data Analysis, Wiley-Blackwell, Oxford.] Paired samples t-test: Synonym for matched pairs t-test.

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Teacher-Led Research

  Designing and implementing randomised controlled trials and other forms of experimental research

  • Richard Churches, Eleanor Dommett(Authors)
  • 2016(Publication Date)
  • Crown House Publishing

    (Publisher)

  Figure 3.6 ). In this design, the control condition will be the usual whole class pedagogy and the intervention will be the use of group work, with the two conditions counterbalanced.

  Figure 3.6. Example of a classroom-based within-participant design.

  Figure 3.7. Example of a classroom-based within-participant design with ‘double’ counterbalancing.

  A further refinement to control for order effect (related to having done the cities lesson before the countryside lesson) would be to double up the number of children involved and double counterbalance out these effects (as in Figure 3.7 ).

  Matched Pairs Design

  Earlier on, when we talked about the limitations of a between-participant design, we highlighted the obvious difficulty presented by using different participants in each condition. Namely, that you can end up measuring not the intervention that you exposed people to but the differences between them. We also pointed out that you can’t use a within-participant design if the effects of your treatment are irreversible.

  There is, however, another option that is an extension of the between-participant design. This is called a matched pairs (or case-matched) design. This type of design aims to produce similar benefits to the within-participant design with regard to reducing the individual differences between people in the study. To conduct a Matched Pairs Design you begin by ‘matching’ up participants in pairs according to individual characteristics that you think might confound your research if you did not control for them. For example, the two highest attaining girls might be paired and the two highest attaining boys and so on. Once you have all your participants paired in such a way, you then randomly allocate each member of a pair to the control or intervention group. In this way, you force a balance of characteristics between your participant groups while also including randomisation. In Chapter 4 we will teach you how to do this in practice. Figure 3.8

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Experimental Design and Statistics for Psychology

  A First Course

  • Fabio Sani, John Todman(Authors)
  • 2008(Publication Date)
  • Wiley-Blackwell

    (Publisher)

  The two occupying the first and second ranks would then be assigned, one to each condition, using a random procedure (e.g., a coin toss) to decide which one went into each con-dition. This would be repeated for the pair occupying the third and fourth ranks and so on down to the lowest scoring pair. This design is known as a matched subjects design , and constitutes an attempt to approach the control of participant variation achieved in the repeated measures design. However, this design is not always prac-ticable. While matching people on variables such as sex and age is straightforward, matching them on variables such as personality, intelligence, background and so on may be complicated and very time consuming. In addition, if there are several vari-ables on which it would be desirable to match participants, it can be difficult to find pairs who are a reasonable match on all of those variables. When it would be inappropriate or impractical to use a repeated measures or matched subjects design, the independent groups design is always an option. Recall that this is the type of design upon which our example experiment is based. You will have noticed that we assigned different participants to the two different conditions on the basis of a strictly random procedure . We did not specify at the time what sort of random procedure we used in our experiment. Well, a possible strategy would be to put 40 cards with the participants’ names written on them into a box, and then pick 20 cards out without looking in the box. A coin might then be tossed to decide whether those 20 participants would be allocated to the experimental or control condition. By allocating participants at random we expect the groups to be fairly well matched on all possible participant variables. Obviously, randomization does not ensure that the two groups will be perfectly matched on any particular participant variable. On the other hand, we can be confident that a systematic error produced by participant

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Statistics

  Principles and Methods

  • Richard A. Johnson, Gouri K. Bhattacharyya(Authors)
  • 2019(Publication Date)
  • Wiley

    (Publisher)

  Here, the time to cure is the response of interest. Figure 1a portrays a design of independent samples where the 8 patients are randomly split into groups of 4, one group is treated with drug 1, and the other with drug 2. The observations for drug 1 will have no relation to those for drug 2 because the selection of patients in the two groups is left completely to chance. Split at random Drug 2 Drug 1 (a) Figure 1a Independent samples, each of size 4. 306 CHAPTER 10/COMPARING TWO TREATMENTS To conduct a Matched Pairs Design, one would first select the patients in pairs. The two patients in each pair should be as alike as possible in regard to their physiological conditions; for instance, they should be of the same gender and age group and have about the same severity of the disease. These preexisting conditions may be different from one pair to another. Having paired the subjects, we randomly select one member from each pair to be treated with drug 1 and the other with drug 2. Figure 1b shows this Matched Pairs Design. Drug 2 Drug 1 (b) Pair 1 Pair 2 Pair 3 Pair 4 Figure 1b Matched Pairs Design with four pairs of sub- jects. Separate random assignment of Drug 1 each pair. In contrast with the situation of Figure 1a, we would expect the responses of each pair to be dependent for the reason that they are governed by the same preexisting conditions of the subjects. In summary, a carefully planned experimental design is crucial to a successful comparative study. The design determines the structure of the data. In turn, the design provides the key to selecting an appropriate analysis. Exercises 10.1 Grades for first semester are compared to those for second semester. The five one-semester courses biology, chemistry, English, history, and psychology must be taken next year.

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Behavioral Research and Analysis

  An Introduction to Statistics within the Context of Experimental Design, Fourth Edition

  • Max Vercruyssen, Hal W. Hendrick(Authors)
  • 2011(Publication Date)
  • CRC Press

    (Publisher)

  As a result, the remaining sample will not be representative of the total population of interest. With precise matching, often the extreme scores cannot be used. Thus the range of subjects is restricted and the results of the experiment may be generalized only to a restricted portion of the total population. When more than one or two variables are used for matching, large numbers of subjects are likely to be lost for lack of a partner who matches up on all the variables. Because of this factor, matching on the basis of a number of variables rarely is practical. R EPEATED M EASURES (W ITHIN S UBJECTS ) D ESIGN In the two matched groups design, the purpose of matching was to treat each pair of subjects as one subject with respect to at least one of the variables influencing the DM. Because this increases the precision of an experiment, it follows that using the same subject under both (or all) experimental conditions should increase the precision still further. This is quite true. The individual differences that are sources of random error, are controlled by using the same subject repeatedly. This design is called repeated measures, or within subjects, because we are obtaining measures (DM scores) on each subject for more than one treatment condition. Advantages and Uses of Repeated Measures Designs Because the variability is reduced in a repeated measures design, the experiment is more sensitive to treatment effects. Repeated measures also is a very efficient design; the need for several groups of subjects may be reduced to merely several subjects. The repeated measures design is useful if you expect large individual differences, for example, to study the effect of a drug on pain tolerance because pain tolerance (DM value) varies widely among individuals. This technique is also used to examine changes in behavior over time, such as the effects of learning on task performance.

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Statistical Reasoning in Law and Public Policy

  Tort Law, Evidence and Health

  • Joseph L. Gastwirth(Author)
  • 1988(Publication Date)
  • Academic Press

    (Publisher)

  Comparison of the Characteristics of Two Populations Using Chapter 11 Matched or Paired Data 1. Introduction Many of the methods described in the previous chapters were originally designed to compare the central values (averages) or success rates (pro-portions) of two distinct populations. They were designed for the analysis of independent random samples from each population or for studying the entire population using the concept of randomization. If other factors, e.g., age, education, in addition to group membership, may affect the characteristics studied, as in comparing wage data or health status, then one can stratify the data into comparable subgroups and use one of the combination methods discussed previously to obtain a summary statistic, or one can develop an appropriate regression model. This chapter dis-cusses an approach based on matching each member of one group, usu-ally the treatment or experimental one, with one (or more) members of the other group (often called the control group) possessing similar character-istics with respect to major relevant factors. For example, if one matched each newly hired female to a newly hired male, with similar prior experi-ence and educational background, the difference between their salaries is an estimate of the effect of sex on initial salary, that controls for any difference with regard to the factors of experience and education which 587 588 Statistical Reasoning in Law and Public Policy legitimately affect salary. The average of these differences in a sample of such matched pairs enables us to draw an inference about the employer's treatment of women relative to men. Similarly, in health research one should control for age, smoking habits and other factors which affect one's chances of contracting the disease being investigated. Although finding a good match may be difficult, the technique is quite useful when one population is small relative to the other, so matches are usually available.

  Sign up to read

  Learn more about book
• 

  No longer available |Learn more

  Research Methods For The Behavioural Sciences

  • Frederick J Gravetter; Lori-Ann B. Forzano; Tim Rakow, Frederick Gravetter, Frederick Gravetter, Lori-Ann Forzano, Tim Rakow(Authors)
  • 2021(Publication Date)
  • Cengage Learning EMEA

    (Publisher)

  A similar matching process can be used to equate groups in terms of proportions. If a sample consists of 60 per cent older adults (age 40 or more) and 40 per cent younger adults (age less than 40), restricted random assignment could be used to distribute the older adults equally among the different groups. The same process is then used to distribute the younger adults equally among the groups. The result is that the groups are matched in terms of age, with each group containing exactly 60 per cent older and 40 per cent younger participants. Notice that the matching process requires three steps. 1. Identification of the variable (or variables) to be matched across groups. 2. Measurement of the matching variable for each participant. 3. Assignment of participants to groups by means of a restricted random assignment that ensures a balance between groups. DEFINITION Matching involves assigning individuals to groups so that a specific participant variable is balanced, or matched, across the groups. The intent is to create groups that are equivalent (or nearly equivalent) with respect to the variable matched. This section discusses methods of creating matched groups. An alternative process is one in which each participant in one group is matched one-to-one with an ‘equivalent’ participant in another group. The process is called matching subjects (as opposed to matching groups). Technically, a matched-subjects design is not classified as a between-subjects design and is discussed separately in Chapter 9. Matching groups of participants provides researchers with a relatively easy way to ensure that specific participant variables do not become confounding variables. However, there is a price to pay for match- ing, and there are limitations that restrict the usefulness of this process. To match groups with respect to a specific participant variable, the researcher first must measure the variable.

  Sign up to read

  Learn more about book