Calculate Sums and Means - Cghlewis/data-wrangling-functions GitHub Wiki

Row Sums and Means

When calculating row sums or means, we are often creating a new variable (containing a new value for each row) to be added to our existing data set.

In education, there are many reasons why we may want to calculate row sums or means.

One example is that we often create sum scores or means scores for different measures, and when your data is structured in a wide format (one row per participant), you will want those scores to be calculated row-wise for every individual in your data.

stu_id	item1	item2	item3	sum_score
234	4	3	2	9
255	3	5	2	10
276	1	2	4	7

You may also want to calculate a row-wise total for something like total number of students in a school when your data is disaggregated by a category such as grade level per school.

sch_id	grade_6	grade_7	grade_8	total
4578	120	113	142	375
5900	55	48	61	164
5787	180	175	154	509

There are several different ways to achieve this.

In the tidyverse the two most common methods of calculating row sums, means, etc. is to use:

base::rowSums() or base::rowMeans()

dplyr::rowwise()

There has been some debate around which function is more efficient. Some people have said that rowwise is less efficient and will slow you down because it essentially is a dplyr::group_by() for every row in your data. However, others have said that rowwise has been reinvigorated and is the preferred method now. Both work well so I say take your pick.

I also want to note, that as someone who works with a lot of .sav files and users who use SPSS, I have also included an example of how to calculate row values when your variables contain labelled NA values (as used often in SPSS). R does not recognize labelled NA (like -999 as user-defined missing value) as an NA value in rowwise calculations. Therefore I have included an example, in Calculate row sums or means, of how you may want to work with these types of variables.

Column Sums and Means

We may also need to calculate sums and means for columns (i.e., mean test score for the sample, sum of students across all classrooms, and so forth). Here we are usually summarizing our data, rather than adding new columns to our existing data sets.

Here is an example where we want to calculate a mean math score across all students in our sample.

stu_id	math1	math2	math3	math_score
234	4	3	2	9
255	3	5	2	10
276	1	2	4	7

mean_math_score
8.67

Calculate row values

Calculate column values

Main functions used in examples

Package	Functions
base	rowSums(); rowMeans()
dplyr	rowwise(); summarise()
janitor	adorn_totals()

Other functions used in examples

Package	Functions
dplyr	mutate(); across(); c_across(); select(); case_when(); summarise(); group_by(); if_all()
tidyselect	starts_with(); contains()
base	sum(); mean(); round(); ifelse(); diff()
labelled	na_values()
stringr	str_detect()
tidyr	pivot_longer()
haven	zap_labels()

Resources