{"id":75,"date":"2012-04-25T22:33:47","date_gmt":"2012-04-26T02:33:47","guid":{"rendered":"http:\/\/r4stats.wordpress.com\/?page_id=75"},"modified":"2022-08-25T07:45:08","modified_gmt":"2022-08-25T11:45:08","slug":"statistics","status":"publish","type":"page","link":"https:\/\/r4stats.com\/examples\/statistics\/","title":{"rendered":"Statistics"},"content":{"rendered":"

Below is a comparison of the commands used to perform various statistical analyses in R, SAS, SPSS, and Stata. For R functions that are not included in base R, the library() function loads the package that contains the function right before it is used. The variables gender and workshop are categorical factors, and q1 to q4, pretest and posttest are considered continuous and normally distributed.<\/p>\n

The practice data set is shown here<\/a>. The programs and the data they use are also available for download here<\/a>. Detailed step-by-step explanations are in the books, along with the output of each analysis.<\/p>\n

Analysis of Variance<\/strong><\/p>\n

R<\/strong><\/p>\n

\nmyModel &lt;- aov(posttest ~ workshop,\n  data = mydata100)\nsummary(myModel)\npairwise.t.test(posttest, workshop)\nTukeyHSD(myModel, "workshop")\nplot(TukeyHSD(myModel, "workshop"))\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC GLM;\n  CLASS workshop;\n  MODEL posttest = workshop;\n  MEANS workshop \/ TUKEY;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nUNIANOVA posttest  BY workshop\n  \/POSTHOC = workshop ( TUKEY )\n  \/PRINT = ETASQ HOMOGENEITY\n  \/DESIGN = workshop.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nanova posttest workshop\n<\/pre>\n
Correlate, Pearson<\/strong><\/p>\n
\ncor( mydata[3:6],\n  method = "pearson",\n     use = "pairwise")\ncor.test(mydata$q1,\n  mydata$q2, use = "pairwise")\n\n# Again, adjusting p-values for multiple testing.\nlibrary("Rcmdr")\nrcorr.adjust( mydata[3:6] )\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC CORR;\n  VAR q1-q4;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nCORRELATIONS\n  \/VARIABLES=q1 TO q4.\n<\/pre>\n
Stata<\/strong><\/p>\n
\ncorrelate q*\n<\/pre>\n
Correlate, Spearman<\/strong><\/p>\n
R<\/strong><\/p>\n
\ncor( mydata[3:6],\n  method = "spearman",\n  use = "pairwise")\ncor.test(mydata$q1,\n  mydata$q2,\n  use = "pairwise")\n\n# Again, adjusting p-values for multiple testing.\nlibrary("Rcmdr")\nrcorr.adjust(mydata[3:6], type = "spearman")\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC CORR SPEARMAN;\n  VAR q1-q4;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nNONPAR CORR\n  \/VARIABLES=q1 to q4\n  \/PRINT=SPEARMAN.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nspearman q*\n<\/pre>\n
Crosstabulation & Chi-squared <\/strong><\/p>\n
R<\/strong><\/p>\n
\nmyWG &lt;- table(workshop, gender)\nchisq.test(myWG)\n\nlibrary("gmodels")\nCrossTable(workshop, gender,\n  chisq = TRUE,\n  format = "SAS")\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC FREQ;\n  TABLES workshop*gender \/ CHISQ;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nCROSSTABS\n  \/TABLES=workshop BY gender\n  \/FORMAT= AVALUE TABLES\n  \/STATISTIC=CHISQ\n  \/CELLS= COUNT ROW\n  \/COUNT ROUND CELL\n<\/pre>\n
Stata<\/strong><\/p>\n
\ntab gender workshop, row col exact\n<\/pre>\n
Descriptive Statistics<\/strong><\/p>\n
R<\/strong><\/p>\n
\nsummary(mydata)\n\nlibrary("Hmisc")\ndescribe(mydata)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC MEANS;\nVAR q1--posttest;\n\nPROC UNIVARIATE; VAR q1--posttest;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nDESCRIPTIVES VARIABLES=q1 to posttest\n  \/STATISTICS=MEAN STDDEV VARIANCE\n   MIN MAX SEMEAN.\n\nEXAMINE VARIABLES=q1 to posttest\n  \/PLOT BOXPLOT STEMLEAF NPPLOT\n  \/COMPARE GROUP\n  \/STATISTICS DESCRIPTIVES EXTREME\n  \/MISSING PAIRWISE.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nsummary q*\n\nsummary q*, detail\n<\/pre>\n
Frequencies<\/strong><\/p>\n
R<\/strong><\/p>\n
\nsummary(mydata)\n\nlibrary("Deducer")\nfrequencies(mydata)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC FREQ;\n  TABLES workshop--q4;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nFREQUENCIES VARIABLES= workshop TO q4.\n<\/pre>\n
Stata<\/strong><\/p>\n
\ntab1 workshop gender q*\n<\/pre>\n
Kruskal-Wallis<\/strong><\/p>\n
R<\/strong><\/p>\n
\nkruskal.test(posttest ~\n  workshop)\n\npairwise.wilcox.test(posttest,\n  workshop)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC npar1way;\n  CLASS workshop;\n  VAR posttest;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nNPAR TESTS\n  \/K-W=posttest BY\n   workshop(1 3).\n<\/pre>\n
Stata<\/strong><\/p>\n
\nkwallis q1, by(gender)\n<\/pre>\n
Linear Regression<\/strong><\/p>\n
R<\/strong><\/p>\n
\nmyModel &lt;- lm(q4 ~ q1 + q2 + q3, data = mydata100)\nsummary(myModel)\nplot(myModel)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC REG;\n  MODEL q4 = q1-q3;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nREGRESSION\n  \/MISSING LISTWISE\n  \/STATISTICS COEFF OUTS R ANOVA\n  \/CRITERIA=PIN(.05) POUT(.10)\n  \/NOORIGIN\n  \/DEPENDENT q4\n  \/METHOD=ENTER q1 q2 q3.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nregress q4 q1-q3\nlvr2plot\n<\/pre>\n
Sign Test<\/strong><\/p>\n
R<\/strong><\/p>\n
\nlibrary("PASWR")\nSIGN.test(posttest, pretest,\n  conf.level = .95)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nmyDiff=posttest-pretest;\nPROC UNIVARIATE;\n  VAR myDiff;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nNPTESTS\n  \/RELATED TEST(q1 q2) SIGN\n  \/MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE\n  \/CRITERIA ALPHA=0.05\n   CILEVEL=95.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nbitest posttest &amp;gt;; pretest\n<\/pre>\n
t-Test, Independent<\/strong><\/p>\n
R<\/strong><\/p>\n
\nt.test(q1 ~ gender,\n  data = mydata100)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC TTEST;\n  CLASS gender;\n  VAR q1;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nT-TEST\n  GROUPS = gender('m' 'f')\n  \/VARIABLES = q1.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nttest gender=q1, unpair unequ\n<\/pre>\n
t-Test, Paired<\/strong><\/p>\n
R<\/strong><\/p>\n
\n\n\tt.test(posttest, pretest,\n  paired = TRUE)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC TTEST;\n  PAIRED pretest*posttest;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nT-TEST\n  PAIRS=pretest WITH\n  posttest (PAIRED).\n<\/pre>\n
Stata<\/strong><\/p>\n
\nanova posttest workshop\n<\/pre>\n
Variance Test<\/strong><\/p>\n
R<\/strong><\/p>\n
\nt.test(posttest, pretest,\n  paired = TRUE)\n<\/pre>\n
SAS<\/strong><\/p>\n
It’s built into other procedures, such as GLM.<\/p>\n
SPSS<\/strong><\/p>\n
It’s built into other procedures, such as GLM.<\/p>\n
Stata<\/strong><\/p>\n
\nrobvar posttest, by(gender)\n\n* Or...\nsdtest posttest = gender\n<\/pre>\n
Wilcoxon Rank Sum (Mann-Whitney)<\/strong><\/p>\n
R<\/strong><\/p>\n
\nwilcox.test(q1 ~ gender,\n  data = mydata100)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nPROC NPAR1WAY;\n  CLASS gender;\n  VAR q1;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nNPTESTS\n  \/RELATED TEST(pretest posttest) SIGN WILCOXON.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nranksum posttest, by(gender)\n<\/pre>\n
Wilcoxon Signed Rank (Paired)<\/strong><\/p>\n
R<\/strong><\/p>\n
\nwilcox.test(posttest, pretest, paired = TRUE)\n<\/pre>\n
SAS<\/strong><\/p>\n
\nmyDiff=posttest-pretest;\nPROC UNIVARIATE;\n  VAR myDiff;\n  RUN;\n<\/pre>\n
SPSS<\/strong><\/p>\n
\nNPTESTS\n  \/RELATED TEST(q1 q2) WILCOXON\n  \/MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE\n  \/CRITERIA ALPHA=0.05 CILEVEL=95.\n<\/pre>\n
Stata<\/strong><\/p>\n
\nsignrank q1 = gender\n<\/pre>\n","protected":false},"excerpt":{"rendered":"
Below is a comparison of the commands used to perform various statistical analyses in R, SAS, SPSS, and Stata. For R functions that are not included in base R, the library() function loads the package that contains the function right before it is used. The variables gender and workshop are categorical factors, and q1 to … Continue reading Statistics<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"parent":58,"menu_order":0,"comment_status":"open","ping_status":"closed","template":"","meta":{"jetpack_post_was_ever_published":false,"footnotes":""},"class_list":["post-75","page","type-page","status-publish","hentry"],"aioseo_notices":[],"jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/P6TIhK-1d","jetpack-related-posts":[],"jetpack_likes_enabled":true,"_links":{"self":[{"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/pages\/75","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/comments?post=75"}],"version-history":[{"count":6,"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/pages\/75\/revisions"}],"predecessor-version":[{"id":4146,"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/pages\/75\/revisions\/4146"}],"up":[{"embeddable":true,"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/pages\/58"}],"wp:attachment":[{"href":"https:\/\/r4stats.com\/wp-json\/wp\/v2\/media?parent=75"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}