7  Descriptive Statistics of the NRW80+ Dataset

In this chapter, I illustrate the process of importing NRW80+ data (see Zank et al., 2022) into R. Additionally, I present descriptive statistics and graphical visualizations to gain insights into Likert-scaled surveys. The paper adheres to the APA style, implementing the R template provided by the ‘papaja’ package (Aust & Barth, 2023).

Zank, S., Woopen, C., Wagner, M., Rietz, C., & Kaspar, R. (2022). Quality of Life and Well-being of Very Old People in NRW (Representative Survey NRW80+) - Cross-Section Wave 1. GESIS, Cologne. ZA7558 Data file Version 2.0.0. https://doi.org/10.4232/1.13978
Aust, F., & Barth, M. (2023). papaja: Prepare reproducible APA journal articles with R Markdown. https://github.com/crsh/papaja

7.1 Technical Note

In the following, I load (and install) packages that I use later on and I show information about my R session with sessionInfo().

# (Install and) load pacman package 
if (!require(pacman)) install.packages("pacman")

# load packages that are already installed and install packages that are not 
# installed yet and then load them:
pacman::p_load(tinylabels, 
               papaja,
               haven, 
               labelled, 
               janitor,
               skimr, 
               rstatix, 
               HH, 
               likert, 
               expss,
               tidyr, 
               ggstats,
               psych,
               sjlabelled,
               sjmisc,
               tidyverse, 
               MASS,
               dplyr,
               magick,
               tinytable)

# sessionInfo()

7.2 Import Data

I host a R script on my GitHub account (see https://raw.githubusercontent.com/hubchev/courses/main/scr/readin_GESIS.R) that explains how to import the NRW80+ data. I have manually saved the data, gesis.RData, in a subfolder named data.

7.3 How to Use the NRW80+ Data

7.3.1 Load and Subset Data

I load the data and select some variables that are of particular interest to me.

getwd()
[1] "/home/sthu/Dropbox/hsf/courses/ewa"
load("/home/sthu/Dropbox/hsf/23-ws/ewa/data/gesis.RData")
df <- dfdta |>
  select(starts_with("alter"), 
         ALT_agegroup, 
         ALT_sex, 
         famst1, famst7, 
         demtectcorr, 
         kogstat, 
         final, 
         geschlecht)

# Remove the common prefix from all variables
df <- df |> 
  mutate_all(~ set_label(., gsub("^Alternserleben: ", "", get_label(.))))

For simplification, let us focus on the questions that refer to the “Experience of Ageing” and create a new dataset df_alterl that contains only those questions:

df_alterl <- df |> 
  select(alterl1, 
         alterl2, 
         alterl3, 
         alterl4, 
         alterl5, 
         alterl6, 
         alterl7, 
         alterl8, 
         alterl9, 
         alterl10) |> 
  drop_unused_labels() 

# to remove unused labels you can use drop_unused_labels():
df_alterl_un <- df_alterl |>
  drop_unused_labels()

summary(df_alterl)
    alterl1          alterl2          alterl3          alterl4      
 Min.   :-2.000   Min.   :-2.000   Min.   :-2.000   Min.   :-2.000  
 1st Qu.: 1.000   1st Qu.: 2.000   1st Qu.: 1.000   1st Qu.: 2.000  
 Median : 3.000   Median : 4.000   Median : 2.000   Median : 3.000  
 Mean   : 2.656   Mean   : 3.282   Mean   : 2.349   Mean   : 2.763  
 3rd Qu.: 4.000   3rd Qu.: 4.000   3rd Qu.: 3.000   3rd Qu.: 4.000  
 Max.   : 5.000   Max.   : 5.000   Max.   : 5.000   Max.   : 5.000  
    alterl5         alterl6          alterl7          alterl8      
 Min.   :-2.00   Min.   :-2.000   Min.   :-2.000   Min.   :-2.000  
 1st Qu.: 2.00   1st Qu.: 2.000   1st Qu.: 2.000   1st Qu.: 1.000  
 Median : 3.00   Median : 4.000   Median : 3.000   Median : 3.000  
 Mean   : 2.99   Mean   : 3.405   Mean   : 3.237   Mean   : 2.712  
 3rd Qu.: 4.00   3rd Qu.: 5.000   3rd Qu.: 4.000   3rd Qu.: 4.000  
 Max.   : 5.00   Max.   : 5.000   Max.   : 5.000   Max.   : 5.000  
    alterl9          alterl10     
 Min.   :-2.000   Min.   :-2.000  
 1st Qu.: 2.000   1st Qu.: 1.000  
 Median : 3.000   Median : 2.000  
 Mean   : 2.969   Mean   : 2.305  
 3rd Qu.: 4.000   3rd Qu.: 3.000  
 Max.   : 5.000   Max.   : 5.000

7.3.2 Get an Overview by Counting

7.3.2.1 table() of R base

With the table() function, you can count how many observations of each unique value a variable contains:

table(df_alterl$alterl1)

Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        80          6        390        266        451        511        159

To do that for each variable of a dataset is easy using ~, the pipe operator, and map() of the package purrr (Wickham & Henry, 2023):

Wickham, H., & Henry, L. (2023). purrr: Functional Programming Tools. https://CRAN.R-project.org/package=purrr 
df_alterl |> 
  map(~ table(.))
$alterl1
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        80          6        390        266        451        511        159 

$alterl2
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        36          4        196        245        379        648        355 

$alterl3
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        20          3        500        577        403        244        116 

$alterl4
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
       122          8        222        260        527        543        181 

$alterl5
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
       101          4        199        211        452        680        216 

$alterl6
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        19          3        149        324        358        537        473 

$alterl7
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        20          2        145        362        471        525        338 

$alterl8
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        20          3        516        350        325        340        309 

$alterl9
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        83         10        261        228        425        564        292 

$alterl10
.
Weiß nicht Verweigert  Gar nicht  Ein wenig      Mäßig      Stark Sehr stark 
        44          7        537        433        486        251        105

Using proportions() returns the conditional proportions:

df_alterl |> 
  map(~ proportions(table(.)))
$alterl1
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.042941492 0.003220612 0.209339775 0.142780462 0.242082662 0.274288782 
 Sehr stark 
0.085346216 

$alterl2
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.019323671 0.002147075 0.105206656 0.131508320 0.203435319 0.347826087 
 Sehr stark 
0.190552872 

$alterl3
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.010735373 0.001610306 0.268384326 0.309715513 0.216317767 0.130971551 
 Sehr stark 
0.062265164 

$alterl4
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.065485776 0.004294149 0.119162641 0.139559850 0.282877080 0.291465378 
 Sehr stark 
0.097155126 

$alterl5
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.054213634 0.002147075 0.106816962 0.113258186 0.242619431 0.365002684 
 Sehr stark 
0.115942029 

$alterl6
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.010198604 0.001610306 0.079978529 0.173913043 0.192163178 0.288244767 
 Sehr stark 
0.253891573 

$alterl7
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.010735373 0.001073537 0.077831455 0.194310252 0.252818035 0.281803543 
 Sehr stark 
0.181427805 

$alterl8
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.010735373 0.001610306 0.276972625 0.187869028 0.174449812 0.182501342 
 Sehr stark 
0.165861514 

$alterl9
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.044551798 0.005367687 0.140096618 0.122383253 0.228126677 0.302737520 
 Sehr stark 
0.156736447 

$alterl10
.
 Weiß nicht  Verweigert   Gar nicht   Ein wenig       Mäßig       Stark 
0.023617821 0.003757381 0.288244767 0.232420827 0.260869565 0.134728932 
 Sehr stark 
0.056360709

7.3.2.2 tabyl() of janitor

With tabyl() which is part of janitor (Firke, 2023), we can get both nicely:

Firke, S. (2023). janitor: Simple Tools for Examining and Cleaning Dirty Data. https://CRAN.R-project.org/package=janitor 
df_alterl |> 
  tabyl(alterl1) 
 alterl1   n     percent
      -2  80 0.042941492
      -1   6 0.003220612
       1 390 0.209339775
       2 266 0.142780462
       3 451 0.242082662
       4 511 0.274288782
       5 159 0.085346216
df_alterl |> 
  map(~ tabyl(.))
$alterl1
  .   n     percent
 -2  80 0.042941492
 -1   6 0.003220612
  1 390 0.209339775
  2 266 0.142780462
  3 451 0.242082662
  4 511 0.274288782
  5 159 0.085346216

$alterl2
  .   n     percent
 -2  36 0.019323671
 -1   4 0.002147075
  1 196 0.105206656
  2 245 0.131508320
  3 379 0.203435319
  4 648 0.347826087
  5 355 0.190552872

$alterl3
  .   n     percent
 -2  20 0.010735373
 -1   3 0.001610306
  1 500 0.268384326
  2 577 0.309715513
  3 403 0.216317767
  4 244 0.130971551
  5 116 0.062265164

$alterl4
  .   n     percent
 -2 122 0.065485776
 -1   8 0.004294149
  1 222 0.119162641
  2 260 0.139559850
  3 527 0.282877080
  4 543 0.291465378
  5 181 0.097155126

$alterl5
  .   n     percent
 -2 101 0.054213634
 -1   4 0.002147075
  1 199 0.106816962
  2 211 0.113258186
  3 452 0.242619431
  4 680 0.365002684
  5 216 0.115942029

$alterl6
  .   n     percent
 -2  19 0.010198604
 -1   3 0.001610306
  1 149 0.079978529
  2 324 0.173913043
  3 358 0.192163178
  4 537 0.288244767
  5 473 0.253891573

$alterl7
  .   n     percent
 -2  20 0.010735373
 -1   2 0.001073537
  1 145 0.077831455
  2 362 0.194310252
  3 471 0.252818035
  4 525 0.281803543
  5 338 0.181427805

$alterl8
  .   n     percent
 -2  20 0.010735373
 -1   3 0.001610306
  1 516 0.276972625
  2 350 0.187869028
  3 325 0.174449812
  4 340 0.182501342
  5 309 0.165861514

$alterl9
  .   n     percent
 -2  83 0.044551798
 -1  10 0.005367687
  1 261 0.140096618
  2 228 0.122383253
  3 425 0.228126677
  4 564 0.302737520
  5 292 0.156736447

$alterl10
  .   n     percent
 -2  44 0.023617821
 -1   7 0.003757381
  1 537 0.288244767
  2 433 0.232420827
  3 486 0.260869565
  4 251 0.134728932
  5 105 0.056360709

7.3.2.3 frq() of sjmisc

As the variables df_alterl1 are factors. Thus, we can use the sjmisc package, see Lüdecke (2018) and the cheatsheet of sjmisc http://strengejacke.de/sjmisc-cheatsheet.pdf. Also worth a reading is browseVignettes("sjmisc").

Lüdecke, D. (2018). sjmisc: Data and Variable Transformation Functions. Journal of Open Source Software, 3(26), 754. https://doi.org/10.21105/joss.00754

For example, we can use frq() for nice frequency tables:

df_alterl |> 
  map(~ frq(. , show.na = T))
$alterl1
Beziehungen und andere Menschen mehr schätzen (x) <numeric> 
# total N=1863 valid N=1863 mean=2.66 sd=1.61

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  80 |  4.29 |    4.29 |   4.29
    2 |  Ein wenig |   6 |  0.32 |    0.32 |   4.62
    3 |      Mäßig | 390 | 20.93 |   20.93 |  25.55
    4 |      Stark | 266 | 14.28 |   14.28 |  39.83
    5 | Sehr stark | 451 | 24.21 |   24.21 |  64.04
    6 |       <NA> | 511 | 27.43 |   27.43 |  91.47
    7 |       <NA> | 159 |  8.53 |    8.53 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl2
Gesundheit mehr Aufmerksamkeit widmen (x) <numeric> 
# total N=1863 valid N=1863 mean=3.28 sd=1.45

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  36 |  1.93 |    1.93 |   1.93
    2 |  Ein wenig |   4 |  0.21 |    0.21 |   2.15
    3 |      Mäßig | 196 | 10.52 |   10.52 |  12.67
    4 |      Stark | 245 | 13.15 |   13.15 |  25.82
    5 | Sehr stark | 379 | 20.34 |   20.34 |  46.16
    6 |       <NA> | 648 | 34.78 |   34.78 |  80.94
    7 |       <NA> | 355 | 19.06 |   19.06 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl3
geistige Leistungsfähigkeit nimmt ab (x) <numeric> 
# total N=1863 valid N=1863 mean=2.35 sd=1.28

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  20 |  1.07 |    1.07 |   1.07
    2 |  Ein wenig |   3 |  0.16 |    0.16 |   1.23
    3 |      Mäßig | 500 | 26.84 |   26.84 |  28.07
    4 |      Stark | 577 | 30.97 |   30.97 |  59.04
    5 | Sehr stark | 403 | 21.63 |   21.63 |  80.68
    6 |       <NA> | 244 | 13.10 |   13.10 |  93.77
    7 |       <NA> | 116 |  6.23 |    6.23 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl4
mehr Erfahrung, um Dinge und Menschen einzuschätzen (x) <numeric> 
# total N=1863 valid N=1863 mean=2.76 sd=1.72

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht | 122 |  6.55 |    6.55 |   6.55
    2 |  Ein wenig |   8 |  0.43 |    0.43 |   6.98
    3 |      Mäßig | 222 | 11.92 |   11.92 |  18.89
    4 |      Stark | 260 | 13.96 |   13.96 |  32.85
    5 | Sehr stark | 527 | 28.29 |   28.29 |  61.14
    6 |       <NA> | 543 | 29.15 |   29.15 |  90.28
    7 |       <NA> | 181 |  9.72 |    9.72 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl5
besseres Gespür, was wichtig ist (x) <numeric> 
# total N=1863 valid N=1863 mean=2.99 sd=1.66

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht | 101 |  5.42 |    5.42 |   5.42
    2 |  Ein wenig |   4 |  0.21 |    0.21 |   5.64
    3 |      Mäßig | 199 | 10.68 |   10.68 |  16.32
    4 |      Stark | 211 | 11.33 |   11.33 |  27.64
    5 | Sehr stark | 452 | 24.26 |   24.26 |  51.91
    6 |       <NA> | 680 | 36.50 |   36.50 |  88.41
    7 |       <NA> | 216 | 11.59 |   11.59 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl6
Einschränkung der Aktivitäten (x) <numeric> 
# total N=1863 valid N=1863 mean=3.40 sd=1.38

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  19 |  1.02 |    1.02 |   1.02
    2 |  Ein wenig |   3 |  0.16 |    0.16 |   1.18
    3 |      Mäßig | 149 |  8.00 |    8.00 |   9.18
    4 |      Stark | 324 | 17.39 |   17.39 |  26.57
    5 | Sehr stark | 358 | 19.22 |   19.22 |  45.79
    6 |       <NA> | 537 | 28.82 |   28.82 |  74.61
    7 |       <NA> | 473 | 25.39 |   25.39 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl7
weniger Energie (x) <numeric> 
# total N=1863 valid N=1863 mean=3.24 sd=1.32

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  20 |  1.07 |    1.07 |   1.07
    2 |  Ein wenig |   2 |  0.11 |    0.11 |   1.18
    3 |      Mäßig | 145 |  7.78 |    7.78 |   8.96
    4 |      Stark | 362 | 19.43 |   19.43 |  28.40
    5 | Sehr stark | 471 | 25.28 |   25.28 |  53.68
    6 |       <NA> | 525 | 28.18 |   28.18 |  81.86
    7 |       <NA> | 338 | 18.14 |   18.14 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl8
Abhängigkeit von der Hilfe Anderer (x) <numeric> 
# total N=1863 valid N=1863 mean=2.71 sd=1.53

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  20 |  1.07 |    1.07 |   1.07
    2 |  Ein wenig |   3 |  0.16 |    0.16 |   1.23
    3 |      Mäßig | 516 | 27.70 |   27.70 |  28.93
    4 |      Stark | 350 | 18.79 |   18.79 |  47.72
    5 | Sehr stark | 325 | 17.44 |   17.44 |  65.16
    6 |       <NA> | 340 | 18.25 |   18.25 |  83.41
    7 |       <NA> | 309 | 16.59 |   16.59 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl9
Freiheit, Tage nach eigenem Willen zu verleben (x) <numeric> 
# total N=1863 valid N=1863 mean=2.97 sd=1.68

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  83 |  4.46 |    4.46 |   4.46
    2 |  Ein wenig |  10 |  0.54 |    0.54 |   4.99
    3 |      Mäßig | 261 | 14.01 |   14.01 |  19.00
    4 |      Stark | 228 | 12.24 |   12.24 |  31.24
    5 | Sehr stark | 425 | 22.81 |   22.81 |  54.05
    6 |       <NA> | 564 | 30.27 |   30.27 |  84.33
    7 |       <NA> | 292 | 15.67 |   15.67 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

$alterl10
Motivation fällt schwerer (x) <numeric> 
# total N=1863 valid N=1863 mean=2.31 sd=1.38

Value |      Label |   N | Raw % | Valid % | Cum. %
---------------------------------------------------
   -2 | Weiß nicht |   0 |  0.00 |    0.00 |   0.00
   -1 | Verweigert |   0 |  0.00 |    0.00 |   0.00
    1 |  Gar nicht |  44 |  2.36 |    2.36 |   2.36
    2 |  Ein wenig |   7 |  0.38 |    0.38 |   2.74
    3 |      Mäßig | 537 | 28.82 |   28.82 |  31.56
    4 |      Stark | 433 | 23.24 |   23.24 |  54.80
    5 | Sehr stark | 486 | 26.09 |   26.09 |  80.89
    6 |       <NA> | 251 | 13.47 |   13.47 |  94.36
    7 |       <NA> | 105 |  5.64 |    5.64 | 100.00
 <NA> |       <NA> |   0 |  0.00 |    <NA> |   <NA>

7.3.3 First Summary Statistics

7.3.3.1 Using summary() and get_summary_stats()

First, I am interested in the class of the data and some very basic summary statistics.

summary(df)
    alterl1          alterl2          alterl3          alterl4      
 Min.   :-2.000   Min.   :-2.000   Min.   :-2.000   Min.   :-2.000  
 1st Qu.: 1.000   1st Qu.: 2.000   1st Qu.: 1.000   1st Qu.: 2.000  
 Median : 3.000   Median : 4.000   Median : 2.000   Median : 3.000  
 Mean   : 2.656   Mean   : 3.282   Mean   : 2.349   Mean   : 2.763  
 3rd Qu.: 4.000   3rd Qu.: 4.000   3rd Qu.: 3.000   3rd Qu.: 4.000  
 Max.   : 5.000   Max.   : 5.000   Max.   : 5.000   Max.   : 5.000  
                                                                    
    alterl5         alterl6          alterl7          alterl8      
 Min.   :-2.00   Min.   :-2.000   Min.   :-2.000   Min.   :-2.000  
 1st Qu.: 2.00   1st Qu.: 2.000   1st Qu.: 2.000   1st Qu.: 1.000  
 Median : 3.00   Median : 4.000   Median : 3.000   Median : 3.000  
 Mean   : 2.99   Mean   : 3.405   Mean   : 3.237   Mean   : 2.712  
 3rd Qu.: 4.00   3rd Qu.: 5.000   3rd Qu.: 4.000   3rd Qu.: 4.000  
 Max.   : 5.00   Max.   : 5.000   Max.   : 5.000   Max.   : 5.000  
                                                                   
    alterl9          alterl10        alter_int        alter_cont    
 Min.   :-2.000   Min.   :-2.000   Min.   : 80.00   Min.   : 80.11  
 1st Qu.: 2.000   1st Qu.: 1.000   1st Qu.: 82.00   1st Qu.: 82.99  
 Median : 3.000   Median : 2.000   Median : 86.00   Median : 86.59  
 Mean   : 2.969   Mean   : 2.305   Mean   : 86.48   Mean   : 86.98  
 3rd Qu.: 4.000   3rd Qu.: 3.000   3rd Qu.: 90.00   3rd Qu.: 90.56  
 Max.   : 5.000   Max.   : 5.000   Max.   :102.00   Max.   :102.92  
                                   NA's   :6        NA's   :6       
   alterl_m1       alterl_m2         alterp        ALT_agegroup  
 Min.   :1.000   Min.   :1.000   Min.   :-4.000   Min.   :1.000  
 1st Qu.:2.600   1st Qu.:2.200   1st Qu.:-4.000   1st Qu.:1.000  
 Median :3.200   Median :2.800   Median :-4.000   Median :2.000  
 Mean   :3.168   Mean   :2.877   Mean   : 2.632   Mean   :1.883  
 3rd Qu.:3.800   3rd Qu.:3.600   3rd Qu.:-4.000   3rd Qu.:3.000  
 Max.   :5.000   Max.   :5.000   Max.   :99.000   Max.   :3.000  
 NA's   :16      NA's   :14                                      
    ALT_sex          famst1           famst7        demtectcorr     
 Min.   :1.000   Min.   :-1.000   Min.   :-3.000   Min.   :-11.000  
 1st Qu.:1.000   1st Qu.: 1.000   1st Qu.:-3.000   1st Qu.: -1.000  
 Median :2.000   Median : 4.000   Median : 0.000   Median :  0.000  
 Mean   :1.502   Mean   : 2.765   Mean   :-1.179   Mean   : -1.742  
 3rd Qu.:2.000   3rd Qu.: 4.000   3rd Qu.: 0.000   3rd Qu.:  0.000  
 Max.   :2.000   Max.   : 5.000   Max.   : 1.000   Max.   :  2.000  
                                                                    
    kogstat          final         geschlecht   
 Min.   :-4.00   Min.   :81.00   Min.   :1.000  
 1st Qu.:-4.00   1st Qu.:81.00   1st Qu.:1.000  
 Median :-4.00   Median :81.00   Median :2.000  
 Mean   :-3.21   Mean   :81.09   Mean   :1.502  
 3rd Qu.:-4.00   3rd Qu.:81.00   3rd Qu.:2.000  
 Max.   : 7.00   Max.   :82.00   Max.   :2.000  
                                                
sumstat_alter <- df |> 
  get_summary_stats(
    alterl1, 
    alterl2, 
    alterl3, 
    alterl4, 
    alterl5, 
    alterl6, 
    alterl7, 
    alterl8, 
    alterl9, 
    alterl10,  
    type = "five_number")  
Warning: attributes are not identical across measure variables; they will be
dropped
sumstat_alter
# A tibble: 10 × 7
   variable     n   min   max    q1 median    q3
   <fct>    <dbl> <dbl> <dbl> <dbl>  <dbl> <dbl>
 1 alterl1   1863    -2     5     1      3     4
 2 alterl2   1863    -2     5     2      4     4
 3 alterl3   1863    -2     5     1      2     3
 4 alterl4   1863    -2     5     2      3     4
 5 alterl5   1863    -2     5     2      3     4
 6 alterl6   1863    -2     5     2      4     5
 7 alterl7   1863    -2     5     2      3     4
 8 alterl8   1863    -2     5     1      3     4
 9 alterl9   1863    -2     5     2      3     4
10 alterl10  1863    -2     5     1      2     3

7.3.3.2 Using psych::describe()

A powerful alternative for descriptive summary statistics is provided by the function describe() of the psych package (William Revelle, 2023).

sumstat_alter_psych <- df |>
  select(starts_with("alterl")) |> 
  select(-ends_with("m1"), -ends_with("m2")) |> 
  psych::describe() |> 
  as_tibble(rownames="Question")  |> 
  select(-skew, -kurtosis, -range, -vars) 

sumstat_alter_psych
# A tibble: 10 × 10
   Question     n  mean    sd median trimmed   mad   min   max     se
   <chr>    <dbl> <dbl> <dbl>  <dbl>   <dbl> <dbl> <dbl> <dbl>  <dbl>
 1 alterl1   1863  2.66  1.61      3    2.76  1.48    -2     5 0.0374
 2 alterl2   1863  3.28  1.45      4    3.43  1.48    -2     5 0.0336
 3 alterl3   1863  2.35  1.28      2    2.28  1.48    -2     5 0.0296
 4 alterl4   1863  2.76  1.72      3    2.96  1.48    -2     5 0.0398
 5 alterl5   1863  2.99  1.66      3    3.20  1.48    -2     5 0.0385
 6 alterl6   1863  3.40  1.38      4    3.54  1.48    -2     5 0.0321
 7 alterl7   1863  3.24  1.32      3    3.33  1.48    -2     5 0.0306
 8 alterl8   1863  2.71  1.53      3    2.68  1.48    -2     5 0.0355
 9 alterl9   1863  2.97  1.68      3    3.14  1.48    -2     5 0.0389
10 alterl10  1863  2.31  1.38      2    2.28  1.48    -2     5 0.0321

7.3.3.3 Using summarize() and the tidyverse

As you may be aware, the tidyverse package provides powerful and flexible functions such as filter, select, group_by, and summarize. Here is an example demonstrating how these functions can be utilized to create descriptive statistic tables:

descriptives <- dfdta |>  
  # filter(alterl1 > 0) |> 
  group_by(geschlecht)  |> 
  summarize(
    Mean = mean(alterl1)
    , Count = n()
    , SD = sd(alterl1)
    , Min = min(alterl1)
    , Max = max(alterl1)
  )

descriptives
# A tibble: 2 × 6
  geschlecht    Mean Count    SD Min             Max           
  <dbl+lbl>    <dbl> <int> <dbl> <dbl+lbl>       <dbl+lbl>     
1 1 [Männlich]  2.71   927  1.50 -2 [Weiß nicht] 5 [Sehr stark]
2 2 [Weiblich]  2.60   936  1.72 -2 [Weiß nicht] 5 [Sehr stark]

7.3.4 Make Tables using tt()

```{r, echo=FALSE, eval=TRUE, message=FALSE, warning=FALSE}
#| label: tbl-tabrstatix
#| tbl-cap: "Summary Statistics: Experience of Ageing."

tt(sumstat_alter,  output = "markdown", 
   note = "Note: This table contains all variables of `alterl*`.")
```
Tabelle 7.1: Summary Statistics: Experience of Ageing.
variable n min max q1 median q3
Note: This table contains all variables of `alterl*`.
alterl1 1863 -2 5 1 3 4
alterl2 1863 -2 5 2 4 4
alterl3 1863 -2 5 1 2 3
alterl4 1863 -2 5 2 3 4
alterl5 1863 -2 5 2 3 4
alterl6 1863 -2 5 2 4 5
alterl7 1863 -2 5 2 3 4
alterl8 1863 -2 5 1 3 4
alterl9 1863 -2 5 2 3 4
alterl10 1863 -2 5 1 2 3 
```{r, echo=FALSE, eval=TRUE, message=FALSE, warning=FALSE}
#| label: tbl-tabsumstatalterpsych
#| tbl-cap: "Summary Statistics: Experience of Ageing (psych)"

tt(sumstat_alter,  output = "markdown",
   note = "Note: This table contains all variables of `alterl*`.")
```
Tabelle 7.2: Summary Statistics: Experience of Ageing (psych)
variable n min max q1 median q3
Note: This table contains all variables of `alterl*`.
alterl1 1863 -2 5 1 3 4
alterl2 1863 -2 5 2 4 4
alterl3 1863 -2 5 1 2 3
alterl4 1863 -2 5 2 3 4
alterl5 1863 -2 5 2 3 4
alterl6 1863 -2 5 2 4 5
alterl7 1863 -2 5 2 3 4
alterl8 1863 -2 5 1 3 4
alterl9 1863 -2 5 2 3 4
alterl10 1863 -2 5 1 2 3 
```{r, echo=FALSE, eval=TRUE, message=FALSE, warning=FALSE}
#| label: tbl-tabsumstatalterpsychbal
#| tbl-cap: "Summary Statistics: Experience of Ageing (psych)"

tt(sumstat_alter_psych, output = "markdown",
  note = "This table contains all variables of `alterl*` and only observations where all questions had been answered.")
```
Tabelle 7.3: Summary Statistics: Experience of Ageing (psych)
Question n mean sd median trimmed mad min max se
alterl1 1863 2.655931 1.613659 3 2.757210 1.4826 -2 5 0.03738568
alterl2 1863 3.281804 1.449666 4 3.429913 1.4826 -2 5 0.03358626
alterl3 1863 2.348900 1.278429 2 2.277666 1.4826 -2 5 0.02961898
alterl4 1863 2.763285 1.716885 3 2.963783 1.4826 -2 5 0.03977726
alterl5 1863 2.990338 1.661439 3 3.196512 1.4826 -2 5 0.03849266
alterl6 1863 3.404724 1.384050 4 3.537894 1.4826 -2 5 0.03206605
alterl7 1863 3.236715 1.320460 3 3.325956 1.4826 -2 5 0.03059276
alterl8 1863 2.712292 1.534387 3 2.684775 1.4826 -2 5 0.03554909
alterl9 1863 2.969404 1.677112 3 3.142186 1.4826 -2 5 0.03885578
alterl10 1863 2.305421 1.383735 2 2.284373 1.4826 -2 5 0.03205875 
```{r, echo=FALSE, eval=TRUE, message=FALSE, warning=FALSE}
#| label: tbl-tabdescriptives
#| tbl-cap: "Experience of Ageing: Valuing Relationships and Other People More (By Gender)"

tt(descriptives, output = "markdown")
```
Tabelle 7.4: Experience of Ageing: Valuing Relationships and Other People More (By Gender)
geschlecht Mean Count SD Min Max
1 2.713053 927 1.500062 -2 5
2 2.599359 936 1.717715 -2 5

Table Tabelle 7.1 was created with the function get_summary_stats() of the rstatix package (Kassambara, 2023), Tables Tabelle 7.2 and Tabelle 7.3 were created with the function describe() of the psych package (William Revelle, 2023), and Table Tabelle 7.4 was created with the function summarize() of the dplyr package (Wickham et al., 2023).

Kassambara, A. (2023). rstatix: Pipe-Friendly Framework for Basic Statistical Tests. https://CRAN.R-project.org/package=rstatix
William Revelle. (2023). psych: Procedures for Psychological, Psychometric, and Personality Research. Northwestern University. https://CRAN.R-project.org/package=psych
Wickham, H., François, R., Henry, L., Müller, K., & Vaughan, D. (2023). dplyr: A Grammar of Data Manipulation. https://CRAN.R-project.org/package=dplyr

7.3.5 Use the Likert Scale using gglikert()

We have seen that the data contain not only the five different (Likert scaled) answers. Thus, let us remove all values that have, in one or multiple questions, no answer of the Likert scale. The cleaned dataset is named df_alterl_balance.

df_alterl_balance <- df_alterl %>%
  rowwise() %>%
  mutate(has_negative = ifelse(any(c(across(alterl1:alterl10)) < 0), 1, 0)) |> 
  filter(has_negative == 0) |> 
  select(starts_with("alter")) |> 
  as_tibble()

Using the gglikert() of the ggstats package (Larmarange, 2023) allows us to draw nice graphs. I highly recommend reading the vignette of the package in the R documentation which you get with vignette("gglikert").

Larmarange, J. (2023). ggstats: Extension to ’ggplot2’ for Plotting Stats. https://CRAN.R-project.org/package=ggstats

Figures Abbildung 7.1 and Abbildung 7.3 shows the proportions of answers using df_alterl data and Figures Abbildung 7.2 and Abbildung 7.4 does so using the df_alterl_balance data whereby the latter to show the proportions stacked. Do you see any difference and can you explain the differences?

Abbildung 7.1: Experience of Ageing: Proportions of Answers (df_alterl)
Abbildung 7.2: Experience of Ageing: Proportions of Answers (df_alterl_balance)
Abbildung 7.3: Experience of Ageing: Proportions of Answers - Stacked (df_alter)
Abbildung 7.4: Experience of Ageing: Proportions of Answers - Stacked (df_alterl_balance)

As we are interested in the differences of the two samples, it makes sense to look as the summary statistics for the df_alter_balance sample. This is shown in Table Tabelle 7.3.

7.4 Cross-Referencing in R Markdown

In adherence to the APA style guidelines (Association et al., 2022), it is imperative to reference all figures and tables by their respective numbers within the text. Avoid using generic phrases like “the table above” or “the figure below.” Additionally, refrain from hard-coding the numbers for a more dynamic and standardized approach. Xie et al. (2023) explains concisely how to do that with R Markdown, see: https://bookdown.org/yihui/rmarkdown-cookbook/cross-ref.html.

Association, A. P. et al. (2022). Publication manual of the American psychological association. : American Psychological Association.
Xie, Y., Dervieux, C., & Riederer, E. (2023). R Markdown Cookbook. online. https://bookdown.org/yihui/rmarkdown-cookbook/

For example, I can refer to Table Tabelle 7.1 with @tbl-tabrstatix because I have specified the corresponding label in the R code-chunk, see:

7.5 Exercises

  1. With knitr::purl("desc_NRW80.Rmd") you can extract the whole R code from the R Markdown file and write it into the R script desc_NRW80.R. Try it.

  2. The dataset gesis.RData comes with two different tibbles: dfsav and dfdta. Is there a difference between these two when it comes to the statistics that are shown in this paper? To check that, rename the pdf file desc_NRW80.pdf, change the code in Section @ref(sec-load) so that you are using the other data (df <- dfdta |> ... vs. df <- dfsav |> ...), knit the Rmd again, and compare the stats.

  3. Check possible differences in the gglikert plots when using df_alterl_un instead of df_alterl.

  4. The stats above show that dealing with missing or non-standard answers is a crucial thing. Please read chapter Missing Values of Wickham & Grolemund (2023), see: https://r4ds.hadley.nz/missing-values.

  5. The labels of the variables alterl1:alterl10 have “Alternserleben:” at the beginning. This is not necessary and overloads the graphs. Please change the labels for all graphs using the following code in the respective place in the rmd and then knit it again.

Wickham, H., & Grolemund, G. (2023). R for Data Science (2e). https://r4ds.hadley.nz/ 
# Remove the common prefix from all variables
df <- df |> 
  mutate_all(~ set_label(., gsub("^Alternserleben: ", "", get_label(.))))