Chronbachs Alpha calculation

General
1

I am trying to add the Chronbachs Alpha calculation to every line.It should be a rolling calculation based on every 2 sales values.
so ,one calc for item 1 and item 2 ,then a separate one for item 2 and 3 ,for 3 and 4 ,etc.etc
so ,the Alpha calc should be instead of value ' z = 40/2.'
how do i put the Chronbachs Alpha code in that column Z ?

I have the following code:
input

df <- tibble::tribble(
~seq, ~date, ~sales,
1, "3/01/2017", 40,
2, "4/01/2017", 2,
3, "5/01/2017", 2,
4, "6/01/2017", 2,
5, "7/01/2017", 30,
6, "8/01/2017", 2,
7, "1/02/2017", 9,
8, "2/02/2017", 5,
9, "3/02/2017", 65,
10, "4/02/2017", 3,
11, "5/02/2017", 65
)
library(tibble)
add_column(df, z = 40/2, w = 1)

output

# A tibble: 11 x 5
 seq date      sales     z     w
1 1 3/01/2017 40 20 1
2 2 4/01/2017 2 20 1
3 3 5/01/2017 2 20 1
4 4 6/01/2017 2 20 1
5 5 7/01/2017 30 20 1
6 6 8/01/2017 2 20 1
7 7 1/02/2017 9 20 1
8 8 2/02/2017 5 20 1
9 9 3/02/2017 65 20 1
10 10 4/02/2017 3 20 1
11 11 5/02/2017 65 20 1

library(psy)
data(expsy)
cronbach(cbind(expsy[,c(1,3:10)],-1*expsy[,2]))
2

The cronbach alpha for any two values, is meaningless, how could it not be ?
but here is how that would be demonstrated...

df <- tibble::tribble(
  ~seq, ~date, ~sales,
  1, "3/01/2017", 40,
  2, "4/01/2017", 2,
  3, "5/01/2017", 2,
  4, "6/01/2017", 2,
  5, "7/01/2017", 30,
  6, "8/01/2017", 2,
  7, "1/02/2017", 9,
  8, "2/02/2017", 5,
  9, "3/02/2017", 65,
  10, "4/02/2017", 3,
  11, "5/02/2017", 65
)
library(tidyverse)
library(magrittr) 
library(psy)

df %<>% mutate(lagsales = lag(sales))
df2 <- rowwise(df) %>%  mutate(z = cronbach(cbind(sales,lagsales))$alpha) %>% ungroup

df2
# # A tibble: 11 x 5
# seq date      sales lagsales     z
# <dbl> <chr>     <dbl>    <dbl> <dbl>
#  1 3/01/2017    40       NA    NA
#  2 4/01/2017     2       40    NA
#  3 5/01/2017     2        2    NA
#  4 6/01/2017     2        2    NA
#  5 7/01/2017    30        2    NA
#  6 8/01/2017     2       30    NA
#  7 1/02/2017     9        2    NA
#  8 2/02/2017     5        9    NA
#  9 3/02/2017    65        5    NA
#  10 4/02/2017     3       65    NA
#  11 5/02/2017    65        3    NA

here the cronbach of the pair 40,2 is

> cronbach(cbind(40,2))
$sample.size
[1] 1

$number.of.items
[1] 2

$alpha
[1] NA

i.e. alpha not-applicable

4

great support.! i will try this out and see if i can get that working
cheers

5

I tried it and i got the following message:
Warning in install.packages :
package ‘library(magrittr)’ is not available (for R version 3.6.0)
should i 3.6 out and install 3.5 again?

6

Hi, I had to amend the coding slightly as your package magritter does not exist in version 3.6.

it looks like this

df <- tibble::tribble(
~seq, ~date, ~sales,
1, "3/01/2017", 40,
2, "4/01/2017", 2,
3, "5/01/2017", 2,
4, "6/01/2017", 2,
5, "7/01/2017", 30,
6, "8/01/2017", 2,
7, "1/02/2017", 9,
8, "2/02/2017", 5,
9, "3/02/2017", 65,
10, "4/02/2017", 3,
11, "5/02/2017", 65
)
library(tidyverse)
library(plyr)
library(psy)

df %<>% mutate(sales = lag(sales))
df2 <- rowwise(df) %>% mutate(z = cronbach(cbind(sales,lagsales))$alpha) %>% ungroup

i get the following errorcode:

df %<>% mutate(sales = lag(sales))
Error in df %<>% mutate(sales = lag(sales)) :
could not find function "%<>%"
df2 <- rowwise(df) %>% mutate(z = cronbach(cbind(sales,lagsales))$alpha) %>% ungroup
Error in cbind(sales, lagsales) : object 'lagsales' not found

any ideas how to amend the code?

cheers

7

magrittr does exist for R 3.6 (hard to know what the real problem is, because of lack of information) but this part of the code

Is equivalent to this, without using magrittr (kind of, since part of it is already imported by dplyr)

df <- df %>% mutate(lagsales = lag(sales))
8

great work ! Andrés Castro Socolich , you are a great support!
It provided the Alpha calc.

9

Hello Andre
Quick question

I noticed the Alpha is calculated for the whole line .
How do i calculate a separate rolling Alpha for every two items,so item 2 and 3 , then item 3 and 4 .etc.etc.?

#code:

df <- tibble::tribble(
~seq, ~date, ~sales,
1, "3/01/2017", 40,
2, "4/01/2017", 2,
3, "5/01/2017", 2,
4, "6/01/2017", 2,
5, "7/01/2017", 30,
6, "8/01/2017", 2,
7, "1/02/2017", 9,
8, "2/02/2017", 5,
9, "3/02/2017", 65,
10, "4/02/2017", 3,
11, "5/02/2017", 65
)
library(tidyverse)
library(magrittr)
library(psy)

A tibble: 11 x 5
seq date sales lagsales z


  • 1 1 3/01/2017 40 NA -1.23
    2 2 4/01/2017 2 40 -1.23
    3 3 5/01/2017 2 2 -1.23
    4 4 6/01/2017 2 2 -1.23
    5 5 7/01/2017 30 2 -1.23
    6 6 8/01/2017 2 30 -1.23
    7 7 1/02/2017 9 2 -1.23
    8 8 2/02/2017 5 9 -1.23
    9 9 3/02/2017 65 5 -1.23
    10 10 4/02/2017 3 65 -1.23
    11 11 5/02/2017 65 3 -1.23
10

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.