You are on page 1of 13

Cortina, J.M. (1993). What is coefficient alpha? An examination of theory and applications. Journal of Applied Psychology, 78, 98-104.

Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika. 16, 297-334.
Hatcher, L. (1994). A step-by-step approach to using the SAS(R) system for factor analysis and structural equation modeling. Cary, NC: SAS Institute.
Miller, M.B. (1995). Coefficient alpha: A basic introduction from the perspectives of classical test theory and structural equation modeling. Structural Equation Modeling, 2(3), 255-273.
Nunnaly, J. (1978). Psychometric theory. New York: McGraw-Hill.
Reuterberg, S.E., & Gustafsson, J.E. (1992). Confirmatory factor analysis and reliability: Testing measurement model assumptions. Educational and Psychological Measurement, 52, 795-811.
Santos, J.R.A., Lippke, L., and Pope, P. (1998). PROC FACTOR: A tool for extracting hidden gems from a mountain of variables. Proceedings of the 23rd Annual SAS Users Group International Conference. Cary, NC: SAS Institute Inc.
Zimmerman, D.W., Zumbo, B.D., & Lalonde, C. (1993). Coefficient alpha as an estimate of test reliability under violation of two assumptions. Educational and Psychological Measurement, 53, 33-49.
Cortina, J.M. (1993). What is coefficient alpha? An examination of theory and applications. Journal of Applied Psychology, 78, 98-104.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika. 16, 297-334.
Hatcher, L. (1994). A step-by-step approach to using the SAS(R) system for factor analysis and structural equation modeling. Cary, NC: SAS Institute.
Miller, M.B. (1995). Coefficient alpha: A basic introduction from the perspectives of classical test theory and structural equation modeling. Structural Equation Modeling, 2(3), 255-273.
Reuterberg, S.E., & Gustafsson, J.E. (1992). Confirmatory factor analysis and reliability: Testing measurement model assumptions. Educational and Psychological Measurement, 52, 795-811.
Santos, J.R.A., Lippke, L., and Pope, P. (1998). PROC FACTOR: A tool for extracting hidden gems from a mountain of variables. Proceedings of the 23rd Annual SAS Users Group International Conference. Cary, NC: SAS Institute Inc.
Zimmerman, D.W., Zumbo, B.D., & Lalonde, C. (1993). Coefficient alpha as an estimate of test reliability under violation of two assumptions. Educational and Psychological Measurement, 53, 33-49.
Santos, J.R.A., Lippke, L., and Pope, P. (1998). PROC FACTOR: A tool for extracting hidden gems from a mountain of variables. Proceedings of the 23rd Annual SAS Users Group International Conference. Cary, NC: SAS Institute Inc.
http://trochim.human.cornell.edu/kb/reltypes.htm http://www.nsoe.uconn.edu/siegle/research/Instrument/Reliability/Reliability.htm
N= 4 Sum Var= 15.1837 Alpha= 0.995
var 4.2449 3.34694 3.34694 4.2449 59.8367
1 5 7 4 5 21
2 6 8 5 6 25
3 7 8 6 7 28
4 5 7 4 5 21
5 6 7 5 6 24
6 2 4 1 2 9
7 1 3 1 1 6
We judge the reliability of data tables by estimating how well the items that reflect the same construct yield similar results. We are
looking at how consistent the results are for different items for the same construct within the measure. N is equal to the number
of items and r-bar is the average inter-item correlation among the items. One can see from this formula that if you increase the
number of items, you increase Cronbach's alpha. Additionally, if the average inter-item correlation is low, alpha will be low. As the
average inter-item correlation increases, Cronbach's alpha increases as well. This makes sense intuitively - if the inter-item
correlations are high, then there is evidence that the items are measuring the same underlying construct. This is really what is
meant when someone says they have "high" or "good" reliability. They are refering to how well their items measure a single
unidimensional latent construct. Note, that a reliability coefficient of 0.80 or higher is considered as "acceptable" in most Social
Science applications
http://www.nsoe.uconn.edu/siegle/research/Instrument/Reliability/Reliability.htm
We judge the reliability of data tables by estimating how well the items that reflect the same construct yield similar results. We are
looking at how consistent the results are for different items for the same construct within the measure. N is equal to the number
of items and r-bar is the average inter-item correlation among the items. One can see from this formula that if you increase the
number of items, you increase Cronbach's alpha. Additionally, if the average inter-item correlation is low, alpha will be low. As the
average inter-item correlation increases, Cronbach's alpha increases as well. This makes sense intuitively - if the inter-item
correlations are high, then there is evidence that the items are measuring the same underlying construct. This is really what is
meant when someone says they have "high" or "good" reliability. They are refering to how well their items measure a single
unidimensional latent construct. Note, that a reliability coefficient of 0.80 or higher is considered as "acceptable" in most Social
Science applications
based upon p75 Heyes application of Pearson
x1 x2 x1-mean sqrd x2-mean sqrd DxDy bestfit.htm
1 20 21 5.9 34.81 5.3 28.09 31.27
2 9 7 -5.1 26.01 -8.7 75.69 44.37
3 14 19 -0.1 0.01 3.3 10.89 -0.33
4 13 10 -1.1 1.21 -5.7 32.49 6.27
5 10 13 -4.1 16.81 -2.7 7.29 11.07
6 14 17 -0.1 0.01 1.3 1.69 -0.13
7 18 22 3.9 15.21 6.3 39.69 24.57
8 15 19 0.9 0.81 3.3 10.89 2.97
9 11 14 -3.1 9.61 -1.7 2.89 5.27
10 17 15 2.9 8.41 -0.7 0.49 -2.03
112.9 210.1 123.3
mean 14.1 15.7
r= 0.80058
r= 0.80058
reliability or prediction
-0.934 Pearson Product Moment Correlation (takes into account the mean value)
-0.903 Spearman Rank Correlation Coefficient (considers only the rank order)
test assess test assess
15 5 10 1 Is there a correlation?
79 2 1 8
20 4 8 4 Correlation does not prove causal effect.
18 5 9 1
23 4 6 4 Non-correlation indicates no connection.
21 5 7 1
60 3 3 6
62 2 2 8
55 3 5 6
59 2 4 8
based upon p75 Heyes application of Pearson
x1 x2 x1-mean sqrd x2-mean sqrd DxDy
1 20 21
2 9 7
3 14 19
4 13 10
5 10 13
6 14 17
7 18 22
8 15 19
9 11 14
10 17 15
mean 14.1 15.7
r=
r= 0.8006
0
1
2
3
4
5
6
0
x x-m
(x-m)
2
29 -11 121
31 -9 81
40 0 0
41 1 1
49 9 81
50 10 100
384 > 64.00 > 8
6 n
40 mean
#N/A mode
40.5 median
8 sd 68%
mean
6 12 15 20 25 30 35 20.4
6 14 15 16 17 18 35 17.3
60 120 150 200 250 300 350 156.0
6 6.0
6 6.0
sd
6.5
3.9
65.3
0.0
0.0
age
copy and
sort
use
frequency
function
alt data
23 23 0 23
45 45 2 45
65 65 7 46
45 45 12 52
34 34 16 41
34 34 19 34
56 56 20 45
65 65 56
45 45 52
34 34 49
35 35 48
56 56 52
54 54 47
24 24 48
49 49 0 61
45 45 29 48
78 78 39 45
65 65 49 54
39 39 59 39
55 55 65 next 55
14.6
0
5
10
15
20
25
1 2 3 4 5 6 7
A B C D
1 1 20
2 0 45
3 1 20
4 1 40
5 1 45
6 0 56
7 0 54
8 1 20
9 1 45
10 0 23
0.6 36.8 mean
b 4 42.5 median
g 6 20 mode
8.2 not experimental design
8.14 common mistake
0
10
20
30
40
50
60
1 2 3
40 > 25
8.2 not experimental design
8.14 common mistake
3 4 5 6 7 8 9 10
Principles: significance; reliability; descriptive; inferential;
Cohen & Manion p38-44
Number: nominal; ordinal; interval; ratio
7.12 worked example of poor tables of data
Codable data (nursery) mmm

You might also like