U4A1 - Your Research Question....Please follow all instructions and read the attachments. DUE SUNDAY BY 9PM CST. My Field is Public Service Leadership

HOW TO INTERPRET SPSS DATA

Instructions Review the two analyses presented here: reliability data for a 15-item scale that measures a construct of interest, and factor analysis of 15-item scale measuring some construct (answers the construct validity "question"--is this measuring one construct or a construct with multiple dimensions?).

Notation will be provided to help you interpret the data.

Reliability A measure is considered to be reliable if it gives consistent scores across administrations. Reliability is neccessary for validity. The different types of reliability include:

• Test-retest • Equivalent-forms • Split-halves • Internal consistency - Chronbach's alpha looks at consistency of responses • Inter-rater

Split-Half Reliability and Internal consistency (Chronbach's alpha) assess reliability differently.

• Split-half coefficients are obtained by computing scores for the separate halves of the scale (according to how you list them in SPSS).

• Consistency with coefficient alpha is assessed among the items in the instrument. The greater the consistency in responses among items, the higher the coefficient alpha.

Important Points

• Reliability (a) is measured on the same scale as correlation, therefore ±1.0. The closer to 1.0, the better the internal consistency of the responses.

• The Larger the number of items, the larger the alpha. • Negative reliability is possible, and indicates a problem with your data. Most often this results when

you have some positive-direction questions and some negative-direction questions. Question response values should be coded so that the most desirable response is represented by the highest value. An example of questions that would result in negative reliability:

If 0=never, 1=sometimes and 2=always

(a) My friends support me: 0 1 2 (b) My parents fail to support me: 0 1 2

• Reliability is a property that applies to scores. Therefore the same instrument used 100 times may yield 100 different reliability coefficients.

• So, even if the author of an instrument provides a reliability coefficient for the instrument, that value was based on a normed group and does not apply to your administration of the instrument to your sample of participants. You must analyze the reliability of the instrument with your sample.

Assumption of Equivalency

• If looking at split-half reliability, the two halves should measure the same dimension. • For Internal Consistency (Chronbach's alpha) it is assumed that every item is equivalent to every

other item. Therefore, all items should measure the same underlying dimension. If tenable, the only differences in participant responses should be due to measurement error.

Validity The degree to which a test measures what it is intended to measure. Validity Types include:

• Face validity - Is the instrument measuring what it's supposed to? Does a factor have face validity? • Criterion validty - Compare measurement from two instruments that should be measuring the

same thing. • Content validity - Does insturment measure content of interest? • Construct validity - Degree that a construct is measured.

Case Processing Summary Cases N Percent Valid: 223 55.8 Excludeda: 177 44.3 Total: 400 100.0

a. Listwise deletion based on all variables in the procedure.

Note: 400 is the total number of participants.

Reliability Statistics Cronbach's Alpha Cronbach's Alpha Based on

Standardized Items N of Items

.852 .861 15

Note: .852 is the reliability and 15 is the number of items.

Item Statistics Mean Standard Deviation N

var1: 4.906 1.5812 223 var2: 4.211 1.7796 223 var3: 5.269 1.3587 223 var4: 4.879 1.5708 223 var5: 3.924 2.0835 223

Mean Standard Deviation N var6: 5.166 1.4471 223 var7: 5.251 1.5096 223 var8: 4.507 1.9793 223 var9: 3.516 2.1050 223 var10: 5.283 1.5381 223 var11: 4.915 1.6099 223 var12: 4.291 1.8035 223 var13: 4.991 1.5797 223 var14: 4.094 1.9487 223 var15: 3.798 2.2560 223

Note: Descriptive statistics for each question and for entire scale.

Summary Item Statistics Mean Minimum Maximum Range Maximum/MinimumVariance N of Items

Item Means: 4.600 3.516 5.283 1.767 1.503 .348 15 Item Variances:

3.109 1.846 5.090 3.244 2.757 .974 15

Note: Descriptive statistics for each question and for entire scale.

Item-Total Statistics Scale Mean if Item Deleted

Scale Variance if Item Deleted

Corrected Item- Total Correlation

Squared Multiple Correlation

Cronbach's Alpha if Item Deleted

var1: 64.094 198.320 .600 .482 .837 var2: 64.789 201.230 .458 .390 .844 var3: 63.731 205.828 .509 .390 .843 var4: 64.121 204.774 .451 .356 .845 var5: 65.076 196.755 .452 .397 .845 var6: 63.834 202.355 .561 .454 .840 var7: 63.749 200.657 .575 .425 .839 var8: 64.493 194.584 .526 .345 .840 var9: 65.484 202.773 .339 .284 .853 var10: 63.717 202.285 .523 .495 .841 var11: 64.085 200.258 .542 .411 .840 var12: 64.709 200.469 .466 .293 .844 var13: 64.009 202.288 .506 .508 .842 var14: 64.906 196.608 .496 .343 .842 var15: 65.202 195.360 .429 .242 .848

Note: Cronbach's Alpha if Item Deleted is a very important column to consider when interpreting reliability of scale.

1. Check item if deleted Cronbach alpha values here releative to overall Cronbach alpha (.852). Any values here greater than the overall value would suggest removal of item from scale (since removing it would result in an improvement to the overall alpha level).

2. Remove items with values greater than .852 and run analysis again. Overall Cronbach alpha should be better with item removed.

Scale Statistics Mean Variance Standard Deviation N of Items 69.000 227.532 15.0841 15

Note: Scale Statistics are descriptive statistics for all items together.

Cronbach's Alpha

Part 1    Value: .792 (Reliability for first 8 items.)

Part 1    N of Items: 8a

Part 2    Value: .721 (Reliability for last 7 items.)

Part 2    N of Items: 7b

Total      N of Items: 15

a. The items are: var1, var2, var3, var4, var5, var6, var7, var8.

b. The items are: var9, var10, var11, var12, var13, var14, var15.

Correlation Between Forms

Spearman-Brown Coefficient: .699 (Relationship between forms.)

Equal Length: .802 (Reliablility of equal length scales.)

Unequal Length: .802 (Reliability of unequal length scales.)

Guttman Split-Half Coefficient: .800 (Lower bound estimate of reliability.)

Summary Item Statistics Mean Minimum Maximum Range Maximum/MinimumVariance N of Items

Item Means, Part 1:

4.764 3.924 5.269 1.345 1.343 .252 8a

Item Means, Part 2:

4.413 3.516 5.283 1.767 1.503 .440 7b

Item Means, Both Parts:

4.600 3.516 5.283 1.767 1.503 .348 15

Item Variances, Part 1:

2.827 1.846 4.341 2.495 2.351 .805 8a

3.432 2.366 5.090 2.724 2.151 1.105 7b

Mean Minimum Maximum Range Maximum/MinimumVariance N of Items Item Variances, Part 2: Item Variances, Both Parts:

3.109 1.846 5.090 3.244 2.757 .974 15

a. The items are: var1, var2, var3, var4, var5, var6, var7, var8.

b. The items are: var9, var10, var11, var12, var13, var14, var15.

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy: .859 (The closer to one, the better the factor analysis.)

Bartlett's Test of Sphericity Approximate Chi-Square: 1057.115

Bartlett's Test of Sphericity df: 105

Bartlett's Test of Sphericity Sig: .000 (One wants a value less than .05.)

Note: KMO and Bartlett's Test both test assumptions for running a factor analysis.

Total Variance Explained

Component Total Initial Eigenvalues, Percent of Variance

Initial Eigenvalues, Cumulative Percent

var1: 5.162 34.412 34.412 var2: 1.767 11.777 46.189 var3: 1.044 6.957 53.146 var4: .984 6.559 59.705 var5: .893 5.954 65.660 var6: .750 5.002 70.662 var7: .655 4.365 75.026 var8: .622 4.148 79.175 var9: .588 3.922 83.096 var10: .532 3.548 86.644 var11: .502 3.348 89.992 var12: .437 2.915 92.907 var13: .404 2.694 95.601 var14: .384 2.562 98.162 var15: .276 1.838 100.000

Extraction Method: Principle Component Analysis.

Note: The number of dimensions being measured is found by counting Eigenvalues greater than one (in this case, there are three dimensions). Total Variance Explained indicates the variability being explained by each dimension.

Rotated Component Matrixa Component 1 Component 2 Component 3

var1: .628 .347 .121 var2: .139 .744 .041 var3: .694 .267 -.082 var4: .651 -.004 .190 var5: .119 .761 .042 var6: .466 .415 .216 var7: .542 .468 .046 var8: .425 .474 .094 var9: -.142 .663 .340 var10: .729 -.041 .273 var11: .595 .165 .277 var12: .125 .448 .494 var13: .698 -.090 .367 var14: .312 .073 .722 var15: .194 .144 .688

Extraction Method: Principle Component Analysis.

Rotation Method: Varimax with Kaiser Normalization.

a. Rotation converged in 7 iterations.

Note: There are three dimensions/components. To determine which questions "hang" together one must consider where each loads. The largest component values for each question will indicate which component or dimension a question loads on. Var1 loads on the first component, var2 on the second, var3 on the first, and so on.

CREDITS Subject Matter Expert: Interactive Design:

Instructional Design:

Project Manager:

Shawn Fitzgerald

Jay Austin

Kristine Holmgren

Anne Echtenkamp

L i c e n s e d u n d e r a C r e a t i v e C o m m o n s A t t r i b u t i o n 3 . 0 L i c e n s e .