home *** CD-ROM | disk | FTP | other *** search
/ Explore the World of Soft…e: Engineering & Science / Explore_the_World_of_Software_Engineering_and_Science_HRS_Software_1998.iso / programs / statistc / kstat41.exe / XXKS2.LZH / KS4WHAT.HLP < prev    next >
Text File  |  1995-08-16  |  8KB  |  164 lines
  1. What Analysis Should You Use?
  2. @1,You want to get descriptive statistics of SINGLE variable(s)
  3. @2,You want to get DESCRIPTIVE statistics of two RELATED variables
  4. @3,You want to COMPARE two variables, Independent or Paired
  5. @4,You want to COMPARE more than two variables, Independent or Related
  6. @5,You want to examine ASSOCIATION between two variables
  7. @6,You want to examine ASSOCIATION between more than two variables
  8. @7,Definitions of Terms Used
  9.  
  10. ##1
  11. DESCRIPTIVE STATISTICS & GRAPHS      PROCEDURES TO USE
  12. ═══════════════════════════════      ════════════════════
  13.                ┌─ Data is ─────>     Mean, S.D., Box Plot, 5 number summary
  14.                │  Normal             Histogram, Conf. Interval
  15.                │                     (Stat Module)
  16.                │
  17.                │─ Data not ────>     Median, Box Plot
  18.                │  Normal             Histogram, 5 number summary
  19. One Sample  ───│                     (Stat Module)
  20.                │─ Data is
  21.                │  Categorical──>     Frequencies, Pictogram
  22.                │                     (Crosstabs Module)
  23.                │
  24.                └─ Observations─>     Time Series Plot
  25.                   Over Time          (Stat Module)
  26.  
  27. ##2
  28. DESCRIPTIVE STATISTICS & GRAPHS      PROCEDURES TO USE
  29. ═══════════════════════════════      ════════════════════
  30.                ┌─ Data are─────>     Pearson's Corr. Coeff. &
  31.                │  Normal             X─Y Scatterplot
  32.                │                     (Stat Module &
  33.                │                     Regression Module)
  34.                │
  35. Two Samples────│─ Data not─────>     Spearmans Corr. Coeff. &
  36. (Related)      │  Normal             X─Y Scatterplot
  37.                │                     (Stat Module &
  38.                │                     Regression Module)
  39.                │
  40.                └─ Data are─────>     Crosstabulations and
  41.                   Qualitative        3─D Bar Chart
  42.                                      (Crosstabs Module)
  43.  
  44. ##3
  45. COMPARISON TESTS ─ TWO SAMPLES                    TEST TO USE
  46. ════════════════════════════════                  ═════════════════════
  47.                             ┌─ Data are─────>     Paired t─test
  48.                             │  Normal             (t─test & ANOVA Module)
  49.                             │                      
  50.               │───Samples───│─ Data not ────>     Freidmans Test
  51.               │   Related   │  Normal             (Non-Parametrics Module)
  52.               │             │                      
  53.               │             └─ Data are
  54.               │                Dichotomous──>     McNemar's test
  55. Two Samples ──│                                   (Crosstabs Module)
  56.               │                                    
  57.               │             ┌─ Data are─────>     Ind. Group t─test
  58.               │             │  Normal             (t─test, ANOVA Module)
  59.               │             │                      
  60.               │             │
  61.               │──Samples────│─ Data not─────>     Mann─Whitney U test
  62.                  Independent│  Normal             (Non-Parametrics Module)
  63.                             │                      
  64.                             │
  65.                             └─ Data are─────>     Chi─Square (Homogeniety)
  66.                                Qualitative        (Crosstabs Module)
  67.                                                    
  68. ##4
  69. COMPARING MORE THAN TWO SAMPLES               TEST TO USE
  70. ═════════════════════════════════════════     ═════════════════════
  71.                           ┌─ Data are─────>   Repeated Measures ANOVA
  72.                           │  Normal           (t─test & ANOVA Module)
  73.                           │
  74.               ┌─Samples───│─ Data not ────>   Friedman ANOVA
  75.               │ Related   │  Normal           (Non-Parametrics Module)
  76.               │           │                    
  77.               │           └─ Data are
  78.               │              Dichotomous──>   Cochran's Q test
  79.               │                               (Non-Parametrics Module)
  80. More than     │                                
  81. Two Samples ──│           ┌─ Data are─────>   Independent Group ANOVA
  82.               │           │  Normal           (t─test & ANOVA Module)
  83.               │           │                    
  84.               │           │
  85.               └─Samples───│─ Data not─────>   Kruskal─Wallis
  86.                Independent│  Normal           (Non-Parametrics Module)
  87.                           │                    
  88.                           │
  89.                           └─ Data are─────>   Chi─Square Test
  90.                              Qualitative       (Crosstabs Module)
  91.                                                 
  92. ##5
  93.  
  94. TESTING ASSOCIATION BETWEEN TWO VARIABLES           PROCEDURE TO USE
  95. ═════════════════════════════════════════           ═════════════════
  96.  
  97.                               ┌─ Data are─────>     Pearson Correlation
  98.                               │  Normal             Simple Linear Regression
  99.                               │                     (Regression Module)
  100.                               │                      
  101.                               │
  102.        Two Samples Related────│─ Data not ────>     Spearman Correlation
  103.                               │  Normal             (Regression Module)
  104.                               │                      
  105.                               │─ Data are
  106.                               │  Qualitative──>     Chi-Square (Independence)
  107.                               │                      (Crosstabs Module)
  108.                               │                       
  109.                               └─ Data mixed────>    Spearman Correlation
  110.                                  Normal, Not        (Regression Module)
  111.                                  Normal              
  112. ##6
  113.  
  114.  
  115. MORE THAN TWO ASSOCIATED VARIABLES           PROCEDURE TO USE
  116. ═════════════════════════════════════        ═════════════════
  117.  
  118.                        ┌─ Data are─────>     Multiple Regression
  119.                        │  Normal             (Regression Module)
  120.                        │                      
  121.                        │
  122. More than 2 Samples  ──│─ Data not─────>     Kendall partial rank─
  123. Related                │  Normal             correlation
  124.                        │                     (N.A.)
  125.                        │
  126.                        └─ Data are─────>     Discriminant Analysis
  127.                           Qualitative        (N.A.)
  128.  
  129.  
  130.  
  131. ##7
  132.                                   DEFINITIONS
  133.  
  134. NORMAL refers to data that are well approximated by a normal (Gaussian)
  135. distribution.
  136.  
  137. NOT NORMAL refers to quantative data that are not normally distributed.
  138.  
  139. CATEGORICAL refers to nominal data, such as male/female or brown/blue/black.
  140.  
  141. QUANTITATIVE refers to data that are numeric such as height, batting average,
  142. number of people per household, etc.
  143.  
  144. QUALITATIVE refers to data that describe attributes such as hair color,
  145. socioeconomic class, sex, etc.
  146.  
  147. ASSOCIATED refers to variables where knowledge of one helps predict the
  148. other.
  149.  
  150. INDEPENDENT refers to variables where knowledge of one does not help predict
  151. others. Usually, samples from unrelated populations.
  152. (continued)
  153. ##8
  154.                                   DEFINITIONS
  155.                                   (Continued)
  156.  
  157. RELATED refers to samples where multiple measures are taken on the same or
  158. related entities. For example, before after weights for a diet, or heights of
  159. twins.
  160.  
  161. DICHOTOMOUS refers to data that are categorical and can take on only one of
  162. two possible states. For example, yes,/no or on/off. VARIABLE refers to the
  163. observed measure, such as height, hair color, etc.
  164.