Common and Specific Common and Specific Approaches Approaches in the Analysis of Q-Sort in the Analysis of Q-Sort Data Data with with PQM PQM ethod ethod Peter Schmolck Peter Schmolck Universität der Bundeswehr München by OfficeOne: AutoDateTime 9:24 (0:02:59 )
70
Embed
Common and Specific Approaches in the Analysis of Q-Sort Data with PQMethod Peter Schmolck Universität der Bundeswehr München by OfficeOne: AutoDateTime.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Common and Specific ApproachesCommon and Specific Approachesin the Analysis of Q-Sort Datain the Analysis of Q-Sort Data
withwith PQMPQMethodethod
Peter SchmolckPeter SchmolckUniversität der Bundeswehr München
by OfficeOne: AutoDateTime9:24 (0:02:59)
2
OutlineOutline
Outline
1. HISTORY
2. TYPICAL STEPS OF ANALYSISJOB’S TRAVEL-STUDY EXAMPLE
2.1 Creating the Study Files, Entering the Data
2.2 Analysis: 1st Overview - Likely Number of Factors
2.3 Analysis: Choice of Solution - Fine Tuning
2.4 Results and Interpretation
9:25 (0:03:52) 20:01
3
Outline contd.
3. MORE SPECIFIC QUESTIONS AND SOLUTIONS WITH PQMETHOD
3.1 Relating Q Factors to External Data (with Travel-Study Example)
3.2 Splitting up a Bipolar Factor into its Two Poles (Mother-in-Law Study by Andrea Kettenbach)
9:25 (0:04:07) 30:03
4
1. History 1. History 1. History 1. History
1. History Intro
9:24 (0:02:52) 0:05
5
20 May 1992 20 May 1992 Steven Brown announced availability of QMETHOD
- Author John Atkinson
- Only for mainframe computer systems (IBM, later also VAX)
- Mainframe era over at that time, replaced by personal computers
1 March 1992 1 March 1992 Windows 3.1
1. History
10:38 (0:00:10) 0:05
6
10 Aug 1994 10 Aug 1994 Peter Schmolck announced PC Version of Atkinson’s QMETHOD
- downloadable from WWW homepage
09 Aug 199609 Aug 1996 “QMethod Page”
“http://www.rz.unibw-muenchen.de/~p41bsmk/qmethod” - still OK (but also, e.g. schmolck.org/qmethod)
1. History contd.
9:28 (0:06:52) 0:06
7
20 May 1997 20 May 1997 Version 2 beta
Selection of features and improvements added
- automatic pre-flagging (greatly facilitated program debugging!)
- “.dat” file format, with rows=sorts (mainframe “.raw” still supported)
- Sort-Ids
- Consensus-statements table, eventually resolved the non-distinguishing vs. consensus-statements riddle
- Principal Components extraction
28 Nov 2002 28 Nov 2002 The current release, 2.11
1. History contd.
F1 F2 F3
0 +1 +4 distinquishing for F3
0 +1 +3 not distinguishing for any F, but not consensus (diff. F1 – F3!)
0 +1 +2 not distinguishing for any F, and consensus
9:34 (0:13:17) 0:07
8
2. Typical Steps of 2. Typical Steps of AnalysisAnalysis
2.2 Analysis: 1st Overview - Likely Number of Factors
2.3 Analysis: Choice of Solution - Fine Tuning
2.4 Results and Interpretation
2. Typical Steps of Analysis Outline
9:35 (0:13:50) 0:10*
9
Job‘s OS Publication on which the following demonstrations are based
2. Typical Steps .. Travel Study Example
9:35 (0:14:25) 0:11
10
C:\PQMETHOD\projects\q-conference>pqmethod travel +---------------------------------------------------+ | PQMethod - 2.11 | | (November 2002) | +---------------------------------------------------| | by [email protected] | | Adapted from Mainframe-Program QMethod | | by John Atkinson at KSU | +---------------------------------------------------| | The QMethod Page: | | http://www.rz.unibw-muenchen.de/~p41bsmk/qmethod/ | +---------------------------------------------------+ Hit ENTER to begin
2.1 Creating the Study Files Intro
9:35 (0:14:35) 0:11
11
Current Project is ... C:\PQMETHOD\projects\q-conference/travel Choose the number of the routine you want to run and enter it. 1 - STATES - Enter (or edit) the file of statements 2 - QENTER - Enter q sorts (new or continued) 3 - QCENT - Perform a Centroid factor analysis 4 - QPCA - Perform a Principal Components factor analysis 5 - QROTATE - Perform a manual rotation of the factors 6 - QVARIMAX - Perform a varimax rotation of the factors 7 - QANALYZE - Perform the final Q analysis of the rotated factors 8 - View project files travel.* X - Exit from PQMethod Last Routine Run Successfully - (Initial)2 Checking old input data file .... Enter the title of your study to a max of 68 characters. ____________________________________________________________________Medium-distance decision making strategies How many q statements are there?42
2.1.1 Creating the Study Files
9:36 (0:15:01)
..
12
Enter the leftmost column value (e.g. -5):-4 Enter the rightmost column value (e.g. 5):4 Enter the Number of Rows for each Column from -4 to 4. For Example: 2 3 3 4 4 4 3 3 2 :2 3 5 7 8 7 5 3 2 Ready to process another sort. Enter one of the following codes: A - to add a new sort C - to change a previous sort D - to delete a sort S - to show a previous sort Q - to query status of this study X - to exit QENTER (stop entering/changing sorts)
a
2.1.1 Creating the Study Files contd.
9:36 (0:15:07) 0:12
13
Enter identification code for subject no. 1 (A case label consisting of max. 8 characters)Anita NN
Enter the Sort Values for Subject 1 Anita NN Enter the Statement Numbers, Separated by Spaces, for Column -4:20 23 Enter the Statement Numbers, Separated by Spaces, for Column -3:14 17 35 Enter the Statement Numbers, Separated by Spaces, for Column -2:4 8 38 6 41 Enter the Statement Numbers, Separated by Spaces, for Column -1:12 9 16 24 25 37 39 Enter the Statement Numbers, Separated by Spaces, for Column 0:5 10 28 31 32 34 36 42 Enter the Statement Numbers, Separated by Spaces, for Column 1:2 3 15 19 26 29 33
SubjNo: 1 ID: Anita NN The following statements have been entered more than once. 22 The following statements have not been entered 27 The sort must be re-entered. Look at the problems above and decide what column you want to modify first. Give the value of the column you want to change:4 The current values for column 4 are: 11 22 Enter all of the new values, even ones that were good:11 27
SubjNo: 1 ID: Anita NN The Sum is 0.00, and the Mean is 0.00, for Subject 1 Anita NN The Sort is OK, Do You Want to Change It Anyway? (y/N):n Do you want to enter another sort? (Y/n):n Ready to process another sort. Enter one of the following codes: A - to add a new sort C - to change a previous sort D - to delete a sort S - to show a previous sort Q - to query status of this study X - to exit QENTER (stop entering/changing sorts)x Current Project is ... C:\PQMETHOD\projects\q-conference/travel Choose the number of the routine you want to run and enter it. 1 - STATES - Enter (or edit) the file of statements 2 - QENTER - Enter q sorts (new or continued) 3 - QCENT - Perform a Centroid factor analysis 4 - QPCA - Perform a Principal Components factor analysis 5 - QROTATE - Perform a manual rotation of the factors 6 - QVARIMAX - Perform a varimax rotation of the factors 7 - QANALYZE - Perform the final Q analysis of the rotated factors 8 - View project files travel.* X - Exit from PQMethod Last Routine Run Successfully - QENTERx Thank you for using PQMethod Press <ENTER> to exit
2.1.2 Entering the Data finished
9:37 (0:16:14) 0:14
17
2.1.3 The PQMethod Study Files
9:38 (0:17:06) 0:14*
18
1st header record: n sorts, n statements, and study title
2.2.3 Up to How Many Varimax Factors With at Least 2 or 3 Representatives (“Flags”)?
2.2.4 Intercorrelations Between Provisional Factor Scores
2.2.5 Considerations Related to Theoretical Expectations and Interpretability
2.2 Analysis: 1st Overview Outline
9:41 (0:20:06) 0:17*
21
Current Project is ... C:\PQMETHOD\projects\q-conference/travel Choose the number of the routine you want to run and enter it. 1 - STATES - Enter (or edit) the file of statements 2 - QENTER - Enter q sorts (new or continued) 3 - QCENT - Perform a Centroid factor analysis 4 - QPCA - Perform a Principal Components factor analysis 5 - QROTATE - Perform a manual rotation of the factors 6 - QVARIMAX - Perform a varimax rotation of the factors 7 - QANALYZE - Perform the final Q analysis of the rotated factors 8 - View project files travel.* X - Exit from PQMethod Last Routine Run Successfully - (Initial)4 Eigenvalues As Percentages Cumul. Percentages ----------- -------------- ------------------ 1 13.5325 34.6986 34.6986 2 6.4105 16.4371 51.1357 3 2.3483 6.0212 57.1570
3 and 4 Varimax-Rotated Components with „auto-flags“
3.2.2 MiL Study: Determining the Factor Solution 2
Three is OK! Four is too many
11:00 (0:02:52) 0:46
44
Steps to accomplish splitting up the factor #1:
3.2.3 MiL Study: Splitting up the Bipolar Factor Intro
1) Start with QVARIMAX, 3 factors, loaded in PQROT, and flag factors
2) Save factor numbers: 1 1 2 3 3) Reload saved four factors
4) Invert new factor #2 (copy of previous #1)
5) Remove all „-“ flags for factors 1 und 2
6) Save factors and run QANALYZE
For these final analyses, the .dat file was reordered according to MiL’s “grade”
11:02 (0:04:10) 0:47
45
3.2.2 MiL Study: Splitting up the Bipolar Factor 1)
1) Start with QVARIMAX, 3 factors, loaded in PQROT, and flag factors
Sorted by „Grades“
11:02 (0:04:51) 0:48
46
3.2.2 MiL Study: Splitting up the Bipolar Factor 2)
2) Save factor numbers: 1 1 2 3
11:02 (0:04:55) 0:49
47
3.2.2 MiL Study: Splitting up the Bipolar Factor 3)
3) Reload saved four factors
11:03 (0:05:00) 0:49
48
3.2.2 MiL Study: Splitting up the Bipolar Factor 4)
4) Invert new factor #2 (copy of previous #1)
11:03 (0:05:17) 0:50
49
3.2.2 MiL Study: Splitting up the Bipolar Factor 5)
5) Remove all „-“ flags for factors 1 und 2
11:03 (0:05:57) 0:51
50
3.2.2 MiL Study: Splitting up the Bipolar Factor 6)
….. and run QANALYZE
6) Save factors …
11:04 (0:06:04) 0:51
51
3.2.4 QANALYZE 3.2.4 QANALYZE the Results the Results
(MiL.LIS)(MiL.LIS)
3.2.4 QANALYZE 3.2.4 QANALYZE the Results the Results
(MiL.LIS)(MiL.LIS)
3.2.4.1 Contents of the <study>.lis File
3.2.4.3 Factor Q-Sort Values (in MS-Word)
3.2.4.2 The Four Factor-Prototype Sorts
3.2.4.4 Some additional bits of clarification from the <study>.lis
3.2.4 QANALYZE the Results Intro
11:04 (0:06:54) 0:51*
52
Overview of Tables:
• Correlation matrix between sorts
• Unrotated factor loading matrix
• Cumulative communalities matrix
• Rotated factor loading matrix
• Statistics of individual sorts (Mean, Sd)
• Factor score matrix (Z-scores and ranks)
• Correlations between factor scores
• For every factor in turn: Statements sorted by Z-score
• For every pair of factors: Statements sorted by difference
3.2.4.1 Contents of the <study>.lis file
Quite voluminous.
Correlations Between Factor Scores
1 2 3 4
1 1.0000 -0.8601 0.1911 -0.2098
2 -0.8601 1.0000 0.0870 0.2939
3 0.1911 0.0870 1.0000 -0.0392
4 -0.2098 0.2939 -0.0392 1.0000
11:08 (0:02:38) 0:52
53
Overview of Tables (contd.)
• Factor Q-Sort Values (in „pile scaling“)
• Same table sorted by Consensus vs. Disagreement
• Factor characteristics, like „No. of defining variables“
• Distinguishing statements per factor
• Consensus statements
3.2.4.1 Contents of the <study>.lis File contd.
Bulk of loquacious output are variations of the same basic content:
• Its core message consists in the Factor-Prototype Sorts:
• Factor Q-Sort Values (in „pile scaling“)
11:11 (0:05:51) 0:54
• Disclosing a secret little trick here …
• I changed „pile scaling“ from -2 thru +2 -5 thru +5 in the 2nd header record before running QANALYZE
54Make it a table in MS-Word ….
3.2.4.2 Factor Q-Sort Values (Protoype Sorts)
11:12 (0:06:17) 0:56*
55
3.2.4.2 Factor Q-Sort Values (in MS-Word)
11:12 (0:07:01) 0:57
56
3.2.4.3 Prototype Sorts F1+ Fabulous MiL
11:15 (0:09:35) 0:58
57
3.2.4.3 Prototype Sorts F1- Dreadful MiL
11:15 (0:10:01) 1:00
58
3.2.4.3 F2 Annoying but sweet and helpful MiL
11:19 (0:14:10) 1:01
59
3.2.4.3 F3 Distant, touchy and cold-hearted MiL
11:20 (0:14:45) 1:02
60
3.2.4.4 Additional bits of clarification
Some additional bits of clarification to be gleaned from the MiL.lis file
Distinguishing Statements , e.g.
Z-Scores
More precise, and preserve form of distribution.
Example:
F2- / Dreadful MiL : Statements on top more extreme than bottommost statements .
Z-Scores for Factor F1 – Dreadful MiL
6 She is annoying. 1.741 4 She interferes. 1.57121 She knows things better. 1.57112 She nags all day long. 1.513 3 She knows everything better. 1.50825 She is guileful and deals in an underhanded manner. 1.50824She gets easily offended. 1.414
Bottommost less descriptive, less unanimity:35 She remains neutral. -1.23133 She has a wide range of interests. -1.23130 She shows large interest in the well-being of the f -1.24947 She is sympathetic. -1.379
11:24 (0:18:36) 1:04*
61
3.2.4.4 Additional bits of clarification contd.
Consensus statements
There is only one consensus statement:
11:25 (0:19:24) 1:06
62
3.3Splitting and Combining
Data from Different P-Samples
3.3Splitting and Combining
Data from Different P-Samples
3.3.1 Types of Research Approaches and Problems to Be Solved
3.3.2 Handling Q-Sort Data Sets in PQMethod
3.3 Splitting and Combining Data … Intro
11:25 (0:19:43) 1:07
63
3.3.1. Types of Approaches and Problems
Splitting the P set set into sub-samples for separateanalyses
Identifying factors with a small (theoretical, structured) sub-sample
Merging data from different studies that use the same Q sample
11:28 (0:22:18) 1:08
64
3.3.1. Types of Approaches and Problemscontd.
Excursus: The meaning of “Secondary Factor Analysis”
Comparing Q-factor structures between P samples
Correlating sets of Q-factor scores (factor prototype sorts)
“Spiking” a data set with factor scores from another sample
Before-after type of designs
Combined analysis (shared factor solution)
Separate analyses (focus on differing factor structures)
cf. Comparing Q-factor structures between P samples
Sets of factor loadings as quantitative measures
11:33 (0:27:57) 1:10
65
3.3.2 Handling Q-Sort Data Sets in PQMethod
As I told you already …….
11:34 (0:28:34) 1:14
66
3.3.2 Handling Q-Sort Data Sets in PQMethodcontd.
…. use the Editor for managing PQMethod data
11:34 (0:28:43) 1:16
67
3.3.2 Handling Q-Sort Data Sets in PQMethodcontd.
How do I get the Factor Prototypes into the .dat file?