SPSS Statistics

SPSS Statistics

Your hub for statistical analysis, data management, and data documentation. Connect, learn, and share with your peers! 

 View Only

  • 1.  Regression tests

    Posted Thu May 02, 2024 02:02 PM

    Hello,

    I am using SPSS to analyse my questionnaire data for my dissertation. I want to formulate a table with my own data but I don't know how to do it. Something like this:

    M-type (n = 30) I-type (n = 34) E-type (n = 36) Total p-valuea p-value adjustedb
    Age (years) 30.2 ± 8a 24.8 ± 4.5b 30.4 ± 7.7a 28.4 ± 7.3 .001
    Gender [N (%)] .841
    Male 12 (40) 13 (38) 12 (33) 37 (37)
    Female 18 (60) 21 (62) 24 (67) 63 (63)
    Race [N (%)] .01
    Asian 2 (6.7)a 15 (44.1)b 5 (13.9)a 22(22)
    Black 7 (23.3) 3 (8.8) 8 (22.2) 18 (18)
    White 20 (66.7) 14 (41.2) 22 (61.1) 56 (56)
    More than one 1 (3.3) 2 (5.9) 1 (2.8) 4 (4)
    BMI 25.6 ± 3.5 22.2 ± 3.6 24.3 ± 3.5 24.0 ± 3.8 .001 .119
    Sleep quality 3.2 ± 2.9a 5.7 ± 2.5a,b 5.2 ± 4.0b 4.8 ± 3.4 .007 .004
    Sleep duration (hr) 7.9 ± 0.7 8.2 ± 1.1 8.0 ± 0.9 8.0 ± 0.9 .447 .117
    Physical activity (MET) 3805.3 ± 4515.1a 2094.5 ± 2152.8b 2324.8 ± 2175.0a,b 2690.6 ± 3120.3 .06 .023

    Data expressed as mean ± standard deviation or count (%).

    aPearson's chi-squared test with Bonferroni correction was used for categorical variables. One way ANOVA with Bonferroni correction was used for continuous variables.

    bOne way ANCOVA with Bonferroni correction was used to control for age, race, and gender. Different superscript letters indicate significant differences among chronotypes based on Pearson's chi-squared test (race), one way ANOVA (age), and ANCOVA (BMI, sleep quality, sleep duration, and physical activity) with Bonferroni correction. Abbreviations: M-type, morning-type; I-type, intermediate-type; E-type, evening-type; BMI, body mass index; MET, metabolic equivalent.

    and this:

    E-type versus M-type E-type versus I-type
    Dependent variable E-type,
    mean ± SD    		 M-type, mean ± SD β 95% CI p-value I-type,
    mean ± SD    		 β 95% CI p-value
    Number of snacks per week 13.9 ± 4.5 11.2 ± 4.6 −0.317 −0.514 to −0.120 0.002 10.1 ± 4.1 −0.181 −0.388 to 0.026 .086
    Number of snacks between breakfast and lunch 3.4 ± 2.3 3.1 ± 2.4 −0.115 −0.301 to 0.071 0.223 2.1 ± 1.9 −0.062 −0.258 to 0.134 .533
    Number of snacks between lunch and dinner 5.1 ± 1.6 4.4 ± 1.8 −0.181 −0.411 to 0.050 .123 4.1 ± 1.9 −0.139 −0.381 to 0.103 .258
    Number of snacks after dinner 5.3 ± 1.6 3.7 ± 2.2 −0.422 −0.641 to −0.202 <.001 3.9 ± 2.2 −0.215 −0.445 to 0.016 .067
    Energy dense snacks 14.4 ± 4.4 11.3 ± 4.7 −0.289 −0.489 to −0.089 .005 10.5 ± 4.0 −0.165 −0.375 to 0.045 .123
    Salty snacks 2.6 ± 0.9 2.2 ± 1.1 −0.305 −0.535 to −0.075 .010a 2.2 ± 1.1 −0.137 −0.378 to 0.105 .264
    Soft drinks 2.0 ± 1.1 1.5 ± 1.1 −0.141 −0.357 to 0.075 .197 1.4 ± 1.3 −0.080 −0.306 to 0.147 .488
    Candy 2.0 ± 1.0 1.1 ± 0.9 −0.342 −0.576 to −0.107 .005 1.5 ± 1.2 −0.164 −0.41 to 0.083 .190
    Baked goods 2.1 ± 1.0 1.8 ± 1.1 −0.149 −0.356 to 0.058 .156 1.5 ± 0.8 −0.057 −0.274 to 0.160 .604
    Ice cream 1.7 ± 0.9 1.3 ± 0.8 −0.127 −0.361 to 0.108 .286 1.7 ± 1.1 0.078 −0.169 to 0.324 .533
    Salted meats 1.7 ± 1.2 1.3 ± 1.2 −0.185 −0.374 to 0.005 .056 0.7 ± 0.8 −0.200 −0.399 to −0.001 .049a
    Snack bars 2.2 ± 1.1 2.1 ± 1.1 −0.001 −0.214 to 0.213 .995 1.4 ± 1.3 −0.108 −0.332 to 0.116 .341
    Healthy snacks 12.1 ± 5.1 10.8 ± 5.7 −0.131 −0.303 to 0.041 .133 8.5 ± 3.7 −0.085 −0.266 to 0.095 .351
    Fruits 2.9 ± 1.1 2.9 ± 1.3 −0.037 −0.257 to 0.182 .736 2.5 ± 1.1 −0.018 −0.249 to 0.212 .874
    Dairy 2.8 ± 1.3 2.4 ± 1.5 −0.094 −0.293 to 0.105 .350 1.9 ± 1.4 −0.056 −0.265 to 0.153 .596
    Vegetables 2.6 ± 1.4 2.4 ± 1.4 −0.092 −0.295 to 0.111 .369 1.8 ± 1.2 −0.065 −0.278 to 0.149 .549
    Nuts 1.7 ± 1.2 1.6 ± 1.3 −0.113 −0.305 to 0.079 .246 1.2 ± 1.1 −0.034 −0.236 to 0.168 .739
    Breads/rolls/biscuits 2.1 ± 1.1 1.6 ± 1.2 −0.185 −0.371 to 0.002 .052 1.2 ± 1.1 −0.168 −0.364 to 0.028 .092

    Each linear regression model was adjusted for race, age, body mass index, and sleep quality. Standardized and adjusted beta estimates and 95% confidence intervals are presented. p-values for the adjusted model are noted.

    aThe p-value of salty snacks and salted meats was not significant after correcting for FDR. Evening chronotype was used as the reference category. Abbreviations: M-type, morning-type; I-type, intermediate-type; E-type, evening-type; SD, standard deviation; CI, confidence interval.

    I realise it's a lot to ask but I don't know where to start. I would appreciate any suggestions please. Thank you.



    ------------------------------
    Anna
    ------------------------------


  • 2.  RE: Regression tests

    Posted Thu May 02, 2024 05:38 PM
    Edited by David Dwyer Thu May 02, 2024 05:38 PM

    Hi @Anna Nolan
    Are the tables above from the IBM SPSS Statistics Output Viewer?  Were these once SPSS Statistics pivot table objects? If so, then I suggest using a TableLook.

    First, consult the style manual appropriate for your discipline. Then create a TableLook that corresponds to that manual.  For instance, if your preferred style manual is the APA Manual of Style, IBM SPSS Statistics already has a TableLook for that:



    ------------------------------
    David Dwyer
    SPSS Technical Support
    IBM Software
    ------------------------------

    Original Message


  • 3.  RE: Regression tests

    Posted Fri May 03, 2024 09:55 AM
    SPSS will not produce exactly the table structure you want, but for the first table, you can come pretty close using the Custom Tables procedure.  Be sure to set the measurement levels for the variables correctly, and then put the variables in the rows and the statistics in the columns,  You can choose the statistics you want.  Then when you export to Excel or wherever, you will be close to the desired table,

    For the second table, you would need to copy the regression coefficients manually into the spreadsheet

    --



    Original Message


  • 4.  RE: Regression tests

    Posted Fri May 03, 2024 12:48 PM

    Do you know how I would get the adjusted p-values after I run Chi tests / one way ANOVA / linear regression please? I have some tests that appear to be significant but when I use SPSS it is showing that they are insignificant!

    Eg the average BMIs are 24.83, 26.6 and 28.16 which I thought were significantly different across the three groups but when I run one-way ANOVA with Bonferroni correction the p-value was 0.191 (I think!).

    Thank you



    ------------------------------
    Anna Nolan
    ------------------------------

    Original Message


  • 5.  RE: Regression tests

    Posted Fri May 03, 2024 01:07 PM
    The Bonferroni correction is one way of controlling the problem with multiple tests where doing enough tests is likely to eventually produce a significant result by chance. You can read about it here

    Bonferroni is a conservative test.  There are a number of less conservative tests.  In CTABLES, you can choose no correction, Bonferroni, or Benjamini-Hochberg.  Some of the post hoc tests in ANOVA or ONEWAY include multiple testing corrections and some don't.  The dialog help explains this.

    If you have a number of tests from any procedure and want to control the family-wise error rate, the STATS PADJUST extension command provides six methods that differ in their assumptions and method of control, including Bonferroni..  If you don't already have it, you can install it via Extensions > Extension Hub.  It will appear on the menus as Analyze > Descriptive Statistics > Calculate adjusted p values.

    --



    Original Message


  • 6.  RE: Regression tests

    Posted Tue May 07, 2024 09:24 AM

    Thanks very much I appreciate your help



    ------------------------------
    Anna Nolan
    ------------------------------

    Original Message


  • 7.  RE: Regression tests

    Posted Fri May 03, 2024 12:43 PM

    Thank you 



    ------------------------------
    Anna Nolan
    ------------------------------

    Original Message