Added calculations for the one sample t test

diff --git a/src/ChangeLog b/src/ChangeLog
index 7b5fb3d71351dcf4c914f1fd01659e79d5f78653..862ffdaac663dfd078b30ad3a65d9c338f919232 100644 (file)
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,7 @@
+Wed Feb  4 15:34:11 WST 2004 John Darrington <john@darrington.wattle.id.au>
+
+       * t-test.q: Added calculations for the one sample variant of the T-TEST
+
 Tue Feb  3 20:09:54 2004  Ben Pfaff  <blp@gnu.org>
 
        * tab.c: (render_strip) Fix bug that sometimes caused joined text

diff --git a/src/crosstabs.q b/src/crosstabs.q
index 4bd5e68dd46de0ae7331ec046f1e2f536d3d65e2..bf9adc3de0abdc5589ea38d30f7a8f1606f5ad22 100644 (file)
--- a/src/crosstabs.q
+++ b/src/crosstabs.q
@@ -38,7 +38,6 @@
 #include "alloc.h"
 #include "hash.h"
 #include "pool.h"
-#include "dcdflib/cdflib.h"
 #include "command.h"
 #include "lexer.h"
 #include "error.h"

diff --git a/src/lexer.c b/src/lexer.c
index 94793bfb9557d4d0191b411c93405bc09fcd98e2..a3b43a96f63b33d0151c8efa32df3569ba64f018 100644 (file)
--- a/src/lexer.c
+++ b/src/lexer.c
@@ -422,7 +422,7 @@ lex_double_p (void)
 
 /* Returns the value of the current token, which must be a
    floating point number. */
-long
+double
 lex_double (void)
 {
   assert (lex_double_p ());

diff --git a/src/lexer.h b/src/lexer.h
index 83bfa825c3073fb2eaa084c81e70b6dae24c9827..cf7451582106aa3ef066d0bf4db216e637b4c3e6 100644 (file)
--- a/src/lexer.h
+++ b/src/lexer.h
@@ -88,7 +88,7 @@ int lex_end_of_command (void);
 int lex_integer_p (void);
 long lex_integer (void);
 int lex_double_p (void);
-long lex_double (void);
+double lex_double (void);
 
 /* Token matching functions. */
 int lex_match (int);

diff --git a/src/t-test.q b/src/t-test.q
index 589e64a8abc4b57c65ae7f85ebb2a8bbb797cccb..34c3f07842cc88e4f8a89b98b42c527772e4d84b 100644 (file)
--- a/src/t-test.q
+++ b/src/t-test.q
@@ -55,6 +55,7 @@
 static struct cmd_t_test cmd;
 
 
+
 static struct pool *t_test_pool ;
 
 /* Variable for the GROUPS subcommand, if given. */
@@ -129,6 +130,16 @@ enum {
   T_PAIRED
 };
 
+
+static int common_calc (struct ccase *);
+static void common_precalc (void);
+static void common_postcalc (void);
+
+static int one_sample_calc (struct ccase *);
+static void one_sample_precalc (void);
+static void one_sample_postcalc (void);
+
+
 int
 cmd_t_test(void)
 {
@@ -170,6 +181,11 @@ cmd_t_test(void)
       return CMD_FAILURE;
     }
 
+  procedure(common_precalc,common_calc,common_postcalc);
+
+  if (mode == T_1_SAMPLE)
+    procedure(one_sample_precalc,one_sample_calc,one_sample_postcalc);
+
   t_test_pool = pool_create ();
 
   ssbox_create(&stat_summary_box,&cmd,mode);
@@ -250,11 +266,15 @@ tts_custom_groups (struct cmd_t_test *cmd unused)
   return 1;
 }
 
+
+
+
 static int
 tts_custom_pairs (struct cmd_t_test *cmd unused)
 {
   struct variable **vars;
   int n_vars;
+
   int n_before_WITH ;
   int n_after_WITH = -1;
   int paired ; /* Was the PAIRED keyword given ? */
@@ -615,7 +635,14 @@ ssbox_one_sample_populate(struct ssbox *ssb, struct cmd_t_test *cmd)
 
   for (i=0; i < cmd->n_variables; ++i)
     {
+      struct t_test_proc *ttp;
+      ttp= &cmd->v_variables[i]->p.t_t;
+
       tab_text (ssb->t, 0, i+1, TAB_LEFT, cmd->v_variables[i]->name);
+      tab_float (ssb->t,1, i+1, TAB_RIGHT, ttp->n, 2, 0);
+      tab_float (ssb->t,2, i+1, TAB_RIGHT, ttp->mean, 8, 2);
+      tab_float (ssb->t,3, i+1, TAB_RIGHT, ttp->std_dev, 8, 2);
+      tab_float (ssb->t,4, i+1, TAB_RIGHT, ttp->se_mean, 8, 3);
     }
   
 }
@@ -844,6 +871,7 @@ trbox_one_sample_init(struct trbox *self, struct cmd_t_test *cmd )
   tab_text (self->t, 4, 2, TAB_CENTER | TAT_TITLE, _("Mean Difference"));
   tab_text (self->t, 5, 2, TAB_CENTER | TAT_TITLE, _("Lower"));
   tab_text (self->t, 6, 2, TAB_CENTER | TAT_TITLE, _("Upper"));
+
 }
 
 
@@ -857,7 +885,48 @@ trbox_one_sample_populate(struct trbox *trb, struct cmd_t_test *cmd)
 
   for (i=0; i < cmd->n_variables; ++i)
     {
+      int which =1;
+      double t;
+      double p,q;
+      double df;
+      int status;
+      double bound;
+      struct t_test_proc *ttp;
+      ttp= &cmd->v_variables[i]->p.t_t;
+
+
       tab_text (trb->t, 0, i+3, TAB_LEFT, cmd->v_variables[i]->name);
+
+      t = (ttp->mean - cmd->n_testval ) * sqrt(ttp->n) / ttp->std_dev ;
+
+      tab_float (trb->t, 1, i+3, TAB_RIGHT, t, 8,3);
+
+      /* degrees of freedom */
+      df = ttp->n - 1;
+
+      tab_float (trb->t, 2, i+3, TAB_RIGHT, df, 8,0);
+
+      cdft(&which, &p, &q, &t, &df, &status, &bound);
+
+      assert(status == 0 ); /* FIXME: use proper error message */
+
+      /* Multiply by 2 to get 2-tailed significance */
+      tab_float (trb->t, 3, i+3, TAB_RIGHT, q*2.0, 8,3);
+
+      tab_float (trb->t, 4, i+3, TAB_RIGHT, ttp->mean_diff, 8,3);
+
+
+      q = (1 - cmd->criteria)/2.0;  /* 2-tailed test */
+      p = 1 - q ;
+      which=2; /* Calc T from p,q and df */
+      cdft(&which, &p, &q, &t, &df, &status, &bound);
+      assert(status == 0 ); /* FIXME: proper error message */
+
+      tab_float (trb->t, 5, i+3, TAB_RIGHT,
+                ttp->mean_diff - t * ttp->se_mean, 8,4);
+
+      tab_float (trb->t, 6, i+3, TAB_RIGHT,
+                ttp->mean_diff + t * ttp->se_mean, 8,4);
     }
 }
 
@@ -882,3 +951,127 @@ trbox_base_finalize(struct trbox *trb)
 {
   tab_submit(trb->t);
 }
+
+
+/* Calculation Implementation */
+
+/* Per case calculations common to all variants of the T test */
+static int 
+common_calc (struct ccase *c)
+{
+  int i;
+
+  double weight = dict_get_case_weight(default_dict,c);
+
+  for(i=0; i< cmd.n_variables ; ++i) 
+    {
+      struct t_test_proc *ttp;
+      struct variable *v = cmd.v_variables[i];
+      union value *val = &c->data[v->fv];
+
+      ttp= &cmd.v_variables[i]->p.t_t;
+
+      if (val->f != SYSMIS) 
+       {
+         ttp->n+=weight;
+         ttp->sum+=weight * val->f;
+         ttp->ssq+=weight * val->f * val->f;
+       }
+    }
+  return 0;
+}
+
+/* Pre calculations common to all variants of the T test */
+static void 
+common_precalc (void)
+{
+  int i=0;
+  
+  for(i=0; i< cmd.n_variables ; ++i) 
+    {
+      struct t_test_proc *ttp;
+      ttp= &cmd.v_variables[i]->p.t_t;
+      
+      ttp->sum=0;
+      ttp->n=0;
+      ttp->ssq=0;
+      ttp->sum_diff=0;
+    }
+}
+
+/* Post calculations common to all variants of the T test */
+void 
+common_postcalc (void)
+{
+  int i=0;
+
+  for(i=0; i< cmd.n_variables ; ++i) 
+    {
+      struct t_test_proc *ttp;
+      ttp= &cmd.v_variables[i]->p.t_t;
+      
+      ttp->mean=ttp->sum / ttp->n;
+      ttp->std_dev= sqrt(
+                        ttp->n/(ttp->n-1) *
+                        ( (ttp->ssq / ttp->n ) - ttp->mean * ttp->mean )
+                        ) ;
+
+      ttp->se_mean = ttp->std_dev / sqrt(ttp->n);
+
+      ttp->mean_diff= ttp->sum_diff / ttp->n;
+    }
+}
+
+/* Per case calculations for one sample t test  */
+static int 
+one_sample_calc (struct ccase *c)
+{
+  int i;
+
+  double weight = dict_get_case_weight(default_dict,c);
+
+  for(i=0; i< cmd.n_variables ; ++i) 
+    {
+      struct t_test_proc *ttp;
+      struct variable *v = cmd.v_variables[i];
+      union value *val = &c->data[v->fv];
+
+      ttp= &cmd.v_variables[i]->p.t_t;
+      
+      if (val->f != SYSMIS) 
+       ttp->sum_diff += weight * fabs(val->f - cmd.n_testval);
+    }
+
+  return 0;
+}
+
+/* Pre calculations for one sample t test */
+static void 
+one_sample_precalc (void)
+{
+  int i=0;
+  
+  for(i=0; i< cmd.n_variables ; ++i) 
+    {
+      struct t_test_proc *ttp;
+      ttp= &cmd.v_variables[i]->p.t_t;
+      
+      ttp->sum_diff=0;
+    }
+}
+
+/* Post calculations for one sample t test */
+static void 
+one_sample_postcalc (void)
+{
+  int i=0;
+  
+  for(i=0; i< cmd.n_variables ; ++i) 
+    {
+      struct t_test_proc *ttp;
+      ttp= &cmd.v_variables[i]->p.t_t;
+
+      
+      ttp->mean_diff = ttp->sum_diff / ttp->n ;
+    }
+}

diff --git a/src/tab.c b/src/tab.c
index 3166740f2002645b4f3db15deb088a675ba077ad..e017418d2955394af7f1d6fbdc8fa9e7aef3b313 100644 (file)
--- a/src/tab.c
+++ b/src/tab.c
@@ -612,6 +612,11 @@ tab_float (struct tab_table *table, int c, int r, unsigned char opt,
 
   assert (table != NULL && w <= 40);
   
+  assert (c >= 0);
+  assert (c < table->nc);
+  assert (r >= 0);
+  assert (r < table->nr);
+
   f.type = FMT_F;
   f.w = w;
   f.d = d;

diff --git a/src/var.h b/src/var.h
index cc384153b7c28fb469e5de121515fa77e401ebad..71c645387b0c8bc8aa45f3211f265722b741d912 100644 (file)
--- a/src/var.h
+++ b/src/var.h
@@ -128,6 +128,33 @@ struct crosstab_proc
     int count;                 /* max - min. */
   };
 
+
+/* T-TEST private data */
+struct t_test_proc
+  {
+    double mean;
+
+    double std_dev;
+    
+    /* count */
+    double n;
+
+    double sum;
+
+    /* Sum of squares */
+    double ssq;
+
+    /* Std Err of Mean */
+    double se_mean;
+
+    /* Sum of differnces */
+    double sum_diff;
+
+    /* Mean of differences */
+    double mean_diff ;
+  };
+
+
 /* FREQUENCIES private data. */
 enum
   {
@@ -312,6 +339,7 @@ struct variable
        struct sort_cases_proc srt;
        struct matrix_data_proc mxd;
        struct match_files_proc mtf;
+       struct t_test_proc t_t;
       }
     p;
   };