Matched Subject Design: How to Create an Effective Learning Environment

Random number table 

	ODD
39. Goldie Lox	F	105	663
35. Raye Dium	F	112	667
33. I. C. Freeze	M	190	771
19. Clara Asabell	F	115	68
31. Rick O’Chet	M	140	664
25. June Blossoms	F	100	660
13. Carol Ofthebells	F	95	555
37. Summer Rayne	F	96	665
21. April Showers	F	110	663
11. Betty Won’t	F	100	663
1. Anna BoFanna	F	110	660
7. Rhoda Dendron	F	120	665
22. Ben Dover	M	190	69
27. Sam Iam	M	195	669
9. Howie Dooit	M	160	770
5. Seymour Clearly	M	190	771
17. Stu potts	M	190	69
15. Andy Walkswithme	M	250	775
23. Hugh Gottabekiddinme	M	200	772
3. Chuck Wagon	M	200	772

(McGuire, Kristman, Martin and Bédard, 2017). 

	EVEN
22. Ben Dover	M	190	69
32. Bill Board	M	180	760
28. Mike Robiology	M	215	772
10. Willy Makeit	M	180	770
14. Harold Bethyname	M	240	776
4. Ken Garoo	M	180	770
16. Rod Andreel	M	220	772
24. May Flowers	F	130	770
20. Allyson Wonderland	F	118	668
2. Sandy Beach	F	140	770
26. Helen Highwater	F	135	668
12. Al Gebra	M	190	772
8. Pete Moss	M	150	668
18. Sharon Sometimetogether	F	108	665
40. Cindy Rella	F	124	668
36. Helena Baskit	F	114	666
34. Barry Um	M	200	772
38. Deniece Denephew	F	122	679
6. Phil O’Dendron	M	210	774
30. Polly Warner-Cracker	F	109	554
29. Maye Beeso	F	134	667

Range matching 

	Range F (Weight)>=95
13. Carol Ofthebells	F	95	555
37. Summer Rayne	F	96	665
11. Betty Won’t	F	100	663
25. June Blossoms	F	100	660
39. Goldie Lox	F	105	663
18. Sharon Sometimetogether	F	108	665
30. Polly Warner-Cracker	F	109	554
1. Anna BoFanna	F	110	660
21. April Showers	F	110	663
35. Raye Dium	F	112	667
36. Helena Baskit	F	114	666
Range F>=115	Range F (Weight)>=155
19. Clara Asabell	F	115	68
20. Allyson Wonderland	F	118	668
7. Rhoda Dendron	F	120	665
38. Deniece Denephew	F	122	669
40. Cindy Rella	F	124	668
24. May Flowers	F	130	770
29. Maye Beeso	F	134	667
29. Maye Beeso	F	134	667
26. Helen Highwater	F	135	668
2. Sandy Beach	F	140	770

	Range>=140
33. I. C. Freeze	M	140	771
31. Rick O’Chet	M	150	664
27. Sam Iam	M	160	669
9. Howie Dooit	M	180	770
5. Seymour Clearly	M	180	771
17. Stu potts	M	180	69
15. Andy Walkswithme	M	190	775
23. Hugh Gottabekiddinme	M	190	772
3. Chuck Wagon	M	190	772
22. Ben Dover	M	190	669
	Range>=190
32. Bill Board	M	190	770
28. Mike Robiology	M	195	772
10. Willy Makeit	M	200	770
14. Harold Bethyname	M	200	776
4. Ken Garoo	M	200	770
16. Rod Andreel	M	210	772
12. Al Gebra	M	215	772
8. Pete Moss	M	220	668
34. Barry Um	M	240	772
6. Phil O’Dendron	M	250	774

(Bland, 2015). 

Rank-ordered matching

Ranked Ordered matching	Experimental (gender and Height)	Control (Gender and Height)
	F95	F96
	F100	F100
	F105	F108
	F109	F110
	F110	F112
	F114	F115
	F118	F120
	F122	F124
	F130	F134
	F135	F140
	M140	M150
	M160	M180
	M180	M180
	M190	M190
	M190	M190
	M190	M195
	M200	M200
	M200	M210
	M215	M220
	M240	M250

Advantages of range matching 

Range matching is an important technique that helps in optimizing the variables that would have been present in a case study. Range matching may help control the confounding functions of other critical factors in a case study. For example, range matching, when applied to gender, may help with the control of other important socio-economic factors that are relevant in the society in which the individuals are matched in. 

Additionally, range matching is important because it permits the use of smaller units or sample sizes by preparing a “priori” with a “posteriori”. Range matching is also useful because it limits a stratified analysis in a case study with too many strata ( Kleinbaum and Klein,). 

Disadvantages of range matching 

An individual is only limited to the factors that would already have been known by the researchers as being risk factors to the outcome of the study 

There is bias in range matching because the matching criteria entails matching in relation to exposure and not the outcome of the study. 

It is very challenging to the researcher in case the study requires matching of several variables such as age, weight, height and the color of individuals under the study. It normally becomes tedious and confusing because the researcher is required to restrict one of the variables such that it does not become a confounding factor in the long run. 

Advantages of a random number assignment 

It is effective and applicable to all cases 

The researcher is not required to clearly list all service units ( Gentle, 2013). 

Disadvantages a random number assignment 

Requires numbering of the service units without leaving any gaps 

Serial numbers assigned to the service units may not be of any significance to the researcher 

Its use makes it difficult for the researcher to make use of serial numbers when large numbers are involved in the case study. 

References  

Bland, M. (2015).  An introduction to medical statistics . Oxford University Press (UK). 

Gentle, J. E. (2013).  Random number generation and Monte Carlo methods . Springer Science & Business Media. 

Kleinbaum, D. G., & Klein, M. (2010). Analysis of matched data using logistic regression. In  Logistic regression  (pp. 389-428). Springer New York. 

McGuire, C., Kristman, V. L., Martin, L., & Bédard, M. (2017). The association between depression and traumatic brain injury in older adults: a nested matched case control study.  Journal of Aging and Health , 0898264317708072.